The paper presents observational evidence for self-interacting cold dark matter (SIDM). Standard cold dark matter models predict overly dense cores in galaxy centers and an excessive number of halos in the Local Group, conflicting with observations. The authors propose that if dark matter particles self-interact with a large scattering cross-section but negligible annihilation or dissipation, these discrepancies may be resolved. This scenario allows astronomical observations to study dark matter properties inaccessible in the lab. Flat cosmological models with cold dark matter, baryonic matter, and a cosmological constant fit large-scale observations but conflict with galactic and subgalactic scale observations. Observations show that galaxy halos have low density cores, dwarf galaxies have shallow density profiles, and the Milky Way has low density cores. Numerical simulations predict many more dark matter halos in the Local Group than observed. The authors suggest that dark matter particles should have a mean free path of 1 kpc to 1 Mpc at the solar radius. This leads to a scattering cross-section similar to that of ordinary hadrons. The dark matter mass is in the range 1 MeV to 10 GeV. Self-interactions could be due to strong, short-range interactions or weak interactions mediated by light particles. The paper argues that ordinary astrophysical processes cannot resolve the dwarf galaxy problem or the overabundance of halos. If dark matter is not cold but warm, some discrepancies are alleviated, but standard cold dark matter models fit the observed power spectrum of Lyman alpha absorbers, ruling out warm dark matter. The authors propose that dark matter is cold, non-dissipative, but self-interacting. This model satisfies the COBE constraint without self-interactions and avoids the problem of fitting both the IRAS power spectrum and galaxy properties. The self-interactions affect structure on the 1 kpc scale, not the 10 Mpc scale. The mean free path of dark matter particles affects astrophysics. If the mean free path is much longer than 1 Mpc, dark matter behaves as a collisionless gas, forming triaxial halos. If the mean free path is much smaller than 1 kpc, dark matter behaves as a collisional gas, leading to shallower density profiles and more spherical halos. The paper concludes that self-interacting dark matter models are consistent with observations. The estimated range of dark matter cross-section to mass ratio is between 0.45-450 cm²/g. The paper suggests that self-interacting dark matter models predict spherical halo centers, dark matter halos with cores, few dwarf galaxies in groups, and smaller halo radii in dwarf galaxies and galaxy clusters due to collisional stripping. These predictions are consistent with current observations.The paper presents observational evidence for self-interacting cold dark matter (SIDM). Standard cold dark matter models predict overly dense cores in galaxy centers and an excessive number of halos in the Local Group, conflicting with observations. The authors propose that if dark matter particles self-interact with a large scattering cross-section but negligible annihilation or dissipation, these discrepancies may be resolved. This scenario allows astronomical observations to study dark matter properties inaccessible in the lab. Flat cosmological models with cold dark matter, baryonic matter, and a cosmological constant fit large-scale observations but conflict with galactic and subgalactic scale observations. Observations show that galaxy halos have low density cores, dwarf galaxies have shallow density profiles, and the Milky Way has low density cores. Numerical simulations predict many more dark matter halos in the Local Group than observed. The authors suggest that dark matter particles should have a mean free path of 1 kpc to 1 Mpc at the solar radius. This leads to a scattering cross-section similar to that of ordinary hadrons. The dark matter mass is in the range 1 MeV to 10 GeV. Self-interactions could be due to strong, short-range interactions or weak interactions mediated by light particles. The paper argues that ordinary astrophysical processes cannot resolve the dwarf galaxy problem or the overabundance of halos. If dark matter is not cold but warm, some discrepancies are alleviated, but standard cold dark matter models fit the observed power spectrum of Lyman alpha absorbers, ruling out warm dark matter. The authors propose that dark matter is cold, non-dissipative, but self-interacting. This model satisfies the COBE constraint without self-interactions and avoids the problem of fitting both the IRAS power spectrum and galaxy properties. The self-interactions affect structure on the 1 kpc scale, not the 10 Mpc scale. The mean free path of dark matter particles affects astrophysics. If the mean free path is much longer than 1 Mpc, dark matter behaves as a collisionless gas, forming triaxial halos. If the mean free path is much smaller than 1 kpc, dark matter behaves as a collisional gas, leading to shallower density profiles and more spherical halos. The paper concludes that self-interacting dark matter models are consistent with observations. The estimated range of dark matter cross-section to mass ratio is between 0.45-450 cm²/g. The paper suggests that self-interacting dark matter models predict spherical halo centers, dark matter halos with cores, few dwarf galaxies in groups, and smaller halo radii in dwarf galaxies and galaxy clusters due to collisional stripping. These predictions are consistent with current observations.