We measure the cosmic star formation history out to z = 1.3 using a sample of 918 radio-selected star forming galaxies within the 2 square degree COSMOS field. To increase our sample size, we combine 1.4 GHz flux densities from the VLA-COSMOS catalogue with flux densities measured from the VLA-COSMOS radio continuum image at the positions of I < 26.5 galaxies, enabling us to detect 1.4 GHz sources as faint as 40 uJy. We find radio measurements of the cosmic star formation history are highly dependent on sample completeness and models used to extrapolate the faint end of the radio luminosity function. For our preferred model of the luminosity function, we find the star formation rate density increases from 0.019 Solar masses per year per cubic Mpc at z = 0.225 to 0.104 Solar masses per year per cubic Mpc, which agrees to within 33% of recent UV, IR and 3 GHz measurements of the cosmic star formation history.

Periodic dips observed in approx. 20% of low-mass X-ray binaries are thought to arise from obscuration of the neutron star by the outer edge of the accretion disk. We report the detection with the Rossi X-ray Timing Explorer of two dipping episodes in Aql X-1, not previously a known dipper. The X-ray spectrum during the dips exhibited an elevated neutral column density, by a factor between 1 and almost two orders of magnitude. Dips were not observed in every cycle of the 18.95-hr orbit, so that the estimated frequency for these events is 0.10 (+0.07,-0.05)/cycle. This is the first confirmed example of intermittent dipping in such a system. Assuming that the dips in Aql X-1 occur because the system inclination is intermediate between the non-dipping and dipping sources, implies a range of 72-79 deg. for the source. This result lends support for the presence of a massive (> 2 M_sun) neutron star in Aql X-1, and further implies that approx. 30 additional LMXBs may have inclinations within this range, raising the possibility of intermittent dips in those systems also. Thus, we searched for dips from 24 other bursting systems, without success. For the system with the largest number of dip phases covered, 4U 1820-303, the nondetection implies a 95% upper limit to the dip frequency of 1.4E-3/cycle.