Dynamics of ''small galaxies'' in the Hubble deep field

We have previously found in the Hubble Deep Field (HDF) a significant angular correlation of faint, high-color redshift objects on scales below 1'', or several kiloparsecs in metric size. A correlation at this scale is most likely due to physical associations. We examine the correlation and nearest neighbor statistics to conclude that 38% of these objects in the HDF have a companion within 1'' (or similar to 6 kpc), 3 times the number expected in a random distribution with the same number of objects; the total excess approaches 1.5 objects by separations of 10''. We next examine three possible dynamical scenarios for these object multiplets: (1) the objects are star-forming regions within normal galaxies, whose disks have been relatively dimmed by K-correction and surface brightness dimming; (2) they are fragments merging into large galaxies; (3) they are satellites accreting onto parent normal L* galaxies. We find that hypothesis 1 is most tenable. First, large galaxies in the process of a merger formation would have accumulated too much mass in their centers (5 x 10(12) M. inside 2 kpc) to correspond to any abundant category of present-day objects. Second, accretion by dynamical friction occurs with a predictable slope in density versus radius that is not seen among the faint HDF objects. Since the dynamical friction time is roughly 1 Gyr, a steady state should have been reached by redshift z less than or similar to 5. In the context of these two dynamical scenarios, we consider the possible effects of a gradient in mass-to-light ratio caused by induced star formation during infall. We note that star-forming regions within galaxies clearly present no dynamical problems, but also that large spirals would still appear as such in the HDF, which leads us to favor a scenario in which the faint compact sources in the HDF are giant star-forming regions within small normal galaxies, such as Magellanic irregulars. Last, we note that the ''excess'' number of correlated objects near a given faint source approaches 1.5, suggesting that the previous counts of objects have overestimated the number of galaxies by a factor of 2.5 while underestimating their individual luminosities by the same factor.