Protons with energies up to ~ 10 15 eV are the main component 1 of cosmic rays, but evidence for the specific locations where they could have been accelerated to these energies has been lacking 2 . Electrons are known to be accelerated to cosmic-ray energies in supernova remnants 3,4 , and the shock waves associated with such remnants, when they hit the surrounding interstellar medium, could also provide the energy to accelerate protons. The signature of such a process would be the decay of pions (π 0 ), which are generated when the protons collide with atoms and molecules in an interstellar cloud: pion decay results in γ-rays with a particular spectral-energy distribution 5,6 . Here we report the observation of cascade showers of optical photons resulting fromγ-rays at energies of ~ 10 12 eV hitting Earth's upper atmosphere, in the direction of the supernova remnant RX J1713.7-3946. The spectrum is a good match to that predicted by pion decay, and cannot be explained by other mechanisms.

In this Letter we report the discovery of TeV gamma-ray emission from a supernova remnant made with the CANGAROO 3.8 m telescope. TeV gamma rays were detected at the sky position and extension coincident with the northeast rim of shell-type supernova remnant (SNR) SN 1006 (Type Ia). SN 1006 has been a most likely candidate for an extended TeV gamma-ray source, since the clear synchrotron X-ray emission from the rims was recently observed by ASCA (Koyama et al.), which is strong evidence for the existence of very high energy (up to hundreds of TeV) electrons in the SNR. The observed TeV gamma-ray flux was (2.4 ‫ע‬ 0.0.7 [systematic]) # 10 3.0 ‫ע‬ 0.9 (4.6 ‫ע‬ 0.6 ‫ע‬ 1.4) # 10 1.7 ‫ע‬ 0.5 from the 1996 and 1997 observations, respectively. Also, we set an upper limit on the TeV gamma-ray emission from the southwest rim, which is estimated to be cm Ϫ2 s Ϫ1 (≥ TeV, 95% confidence level) Ϫ12

We have detected sub-TeV gamma-ray emission from the direction of the Galactic center (GC) using the CANGAROO-II Imaging Atmospheric Cerenkov Telescope. We detected a statistically significant excess at energies greater than 250 GeV. The flux was 1 order of magnitude lower than that of the Crab Nebula at 1 TeV with a soft spectrum proportional to . The signal centroid is consistent with the GC direction, and the Ϫ4.6‫5.0ע‬E observed profile is consistent with a pointlike source. Our data suggest that the GeV source 3EG J1746Ϫ2851 is identical to this TeV source, and we study the combined spectra to determine the possible origin of the gammaray emission. We also obtain an upper limit on the cold dark matter density in the Galactic halo.

Temporal changes of X-ray to very high energy gamma-ray emissions from the pulsar-Be-star binary PSR B1259−63/LS 2883 are studied based on three-dimensional smoothed particle hydrodynamic simulations of pulsar wind interaction with Be-disk and wind. We focus on the periastron passage of the binary and calculate the variation of the synchrotron and inverse-Compton emissions using the simulated shock geometry and pressure distribution of the pulsar wind. The characteristic double-peaked X-ray light curve from observations is reproduced by our simulation under a dense Be-disk condition (base density ∼10 −9 g cm −3 ). We interpret the pre-and postperiastron peaks as being due to a significant increase in the conversion efficiency from pulsar spin-down power to the shock-accelerated particle energy at orbital phases when the pulsar crosses the disk before periastron passage, and when the pulsar wind creates a cavity in the disk gas after periastron passage, respectively. On the contrary, in the model TeV light curve, which also shows a double-peak feature, the first peak appears around the periastron phase. The possible effects of cooling processes on the TeV light curve are briefly discussed.

We have been studying the interaction between the Be star and the pulsar in the TeV binary PSR B1259$-$63/LS 2883, using 3-D SPH simulations of the tidal and wind interactions in this Be–pulsar system. We first ran a simulation without pulsar wind nor Be wind, while taking into account only the gravitational effect of the pulsar on the Be disk. In this simulation, the gas particles are ejected at a constant rate from the equatorial surface of the Be star, which is tilted in a direction consistent with multi-waveband observations. We ran the simulation until the Be disk was fully developed and started to repeat a regular tidal interaction with the pulsar. Then, we turned on the pulsar wind and the Be wind. We ran two simulations with different wind mass-loss rates for the Be star, one for a B2 V type and the other for a significantly earlier spectral type. Although the global shape of the interaction surface between the pulsar wind and the Be wind agrees with the analytical solution, the effect of the pulsar wind on the Be disk is profound. The pulsar wind strips off an outer part of the Be disk, truncating the disk at a radius significantly smaller than the pulsar orbit. Our results, therefore, rule out the idea that the pulsar passes through the Be disk around periastron, which has been assumed in previous studies. It also turns out that the location of the contact discontinuity can be significantly different between phases when the pulsar wind directly hits the Be disk and those when the pulsar wind collides with the Be wind. It is thus important to adequately take into account the circumstellar environment of the Be star, in order to construct a satisfactory model for this prototypical TeV binary.