Ultra-compact high velocity clouds ( UCHVCs ) were identified in the Arecibo Legacy Fast ALFA ( ALFALFA ) H i survey as potential gas-bearing dark matter halos .
Here we present higher resolution neutral hydrogen ( H i ) observations of twelve UCHVCS with the Westerbork Synthesis Radio Telescope ( WSRT ) .
The UCHVCs were selected based on a combination of size , isolation , large recessional velocity and high column density as the best candidate dark matter halos .
The WSRT data were tapered to image the UCHVCs at 210 ^ { \prime \prime } ( comparable to the Arecibo resolution ) and 105 ^ { \prime \prime } angular resolution .
In a comparison of the single-dish to interferometer data , we find that the integrated line flux recovered in the WSRT observations is generally comparable to that from the single-dish ALFALFA data .
In addition , any structure seen in the ALFALFA data is reproduced in the WSRT maps at the same angular resolution .
At 210 ^ { \prime \prime } resolution all the sources are generally compact with a smooth H i morphology , as expected from their identification as UCHVCs .
At the higher angular resolution , a majority of the sources break into small clumps contained in a diffuse envelope .
These UCHVCs also have no ordered velocity motion and are most likely Galactic halo clouds .
We identify two UCHVCs , AGC 198606 and AGC 249525 , as excellent galaxy candidates based on maintaining a smooth H i morphology at higher angular resolution and showing ordered velocity motion consistent with rotation .
A third source , AGC 249565 , lies between these two populations in properties and is a possible galaxy candidate .
If interpreted as gas-bearing dark matter halos , the three candidate galaxies have rotation velocities of 8 - 15 km s ^ { -1 } , H i masses of 0.6 - 50 \times 10 ^ { 5 } M _ { \odot } , H i radii of 0.3 - 2 kpc , and dynamical masses of 2 - 20 \times 10 ^ { 7 } M _ { \odot } for a range of plausible distances .
These are the UCHVCs with the highest column density values in the ALFALFA H i data and we suggest this is the best way to identify further candidates .