There is increasing evidence that the central engine of the most common form of gamma-ray burst (GRBs), those that last longer than a few seconds, are afterglows from ultra-relativistic jets produced during the formation of stellar-mass black holes (McFaden & Woosley, 1999). Mirabel & Rodríguez (1999) proposed that ultra-relativistic bulk motion and beaming are needed to explain: 1) the enormous energy requirements of 1054 erg if the emission were isotropic (e.g. Kulkarni et al. 1999; Castro-Tirado et al. 1999); 2) the statistical correlation between time variability and brightness (Ramirez-Ruiz & Fenimore, 2000), and 3) the statistical anti-correlation between brightness and time-lag between hard and soft components (Norris et al. 2000). Beaming reduces the energy release by the beaming factor f = / 4, where is the solid angle of the beamed emission. Additionally, the photon energies can be boosted to higher values. Extreme flows from collapsars with bulk Lorentz factors > 100 have been proposed as sources of -ray bursts (Mészáros & Rees 1997). High collimation (Dado, Dar & de Rújula 2002; Pugliese et al. 1999) can be tested observationally (Rhoads, 1997), since the statistical properties of the bursts will depend on the viewing angle relative to the jet axis.
Recent multi-wavelength studies of gamma-ray afterglows suggest that they are highly collimated jets. The brightness of the optical transient associated to some GRBs show a break (e.g. Kulkarni et al. 1999), and a steepening from a power law in time t proportional to t-1.2, ultimately approaching a slope t-2.5 (e.g. Castro-Tirado et al. 1999). The achromatic steepening of the optical light curve and early radio flux decay of some GRBs are inconsistent with simple spherical expansion, and well fit by jet evolution. It is interesting that the power laws that describe the light curves of the ejecta in microquasars show similar breaks and steepening of the radio flux density (Rodríguez & Mirabel, 1999). In microquasars, these breaks and steepenings have been interpreted (Hjellming & Johnston 1988) as a transition from slow intrinsic expansion followed by free expansion in two dimensions. Besides, linear polarizations of about 2% were recently measured in the optical afterglows (e.g. Covino et al. 1999), providing strong evidence that the afterglow radiation from gamma-ray bursts is, at least in part, produced by synchrotron processes. Linear polarizations in the range of 2-10% have been measured in microquasars at radio (e.g. Rodríguez et al. 1995), and optical (Scaltriti et al. 1997) wavelengths.
The jets in microquasars of our own Galaxy seem to be less extreme local analogs of the super-relativistic jets associated to the more distant gamma-ray bursts. But the latter do not repeat, seem to be related to catastrophic events, and have much larger super-Eddington luminosities. According to the latest models, the same symbiotic disk-jet relationship as in microquasars and quasars powers the GRBs. In fact, it is now believed that the Lorentz factors at the base of the jets inside the collapsing star are 10 as in microquasars and quasars, and they reach values 100 when they break free from the infalling outer layers of the progenitor star. Because of the enormous difference in power, the scaling laws in terms of the black hole mass that are valid for the analogy between microquasars and quasars may not apply in the case of gamma-ray bursts.