The first of a new class of objects now known as main belt comets ( MBCs ) or “ activated asteroids ” was identified in 1996 .
The seven known members of this class have orbital characteristics of main belt asteroids yet exhibit dust ejection like comets .
In order to constrain their physical and orbital properties we searched the Thousand Asteroid Light Curve Survey ( TALCS ; Masiero et al. , 2009 ) for additional candidates using two diagnostics : tail and coma detection .
This was the most sensitive MBC survey effort to date , extending the search from MBCs with H \sim 18 ( D \sim 1 km ) to MBCs as small as H \sim 21 ( D \sim 150 m ) .

We fit each of the 924 objects detected by TALCS to a PSF model incorporating both a coma and nuclear component to measure the fractional contribution of the coma to the total surface brightness .
We determined the significance of the coma detection using the same algorithm on a sample of null detections of comparable magnitude and rate of motion .
We did not identify any MBC candidates with this technique to a sensitivity limit on the order of cometary mass loss rate of about 0.1 kg/s .

Our tail detection algorithm relied on identifying statistically significant flux in a segmented annulus around the candidate object .
We show that the technique can detect tail activity throughout the asteroid belt to the level of the currently known MBCs .
Although we did not identify any MBC candidates with this technique , we find a statistically significant detection of faint activity in the entire ensemble of TALCS asteroids .
This suggests that many main belt asteroids are active at very low levels .

Our null detection of MBCs allows us to set 90 % upper confidence limits on the number distribution of MBCs as a function of absolute magnitude , semimajor axis , eccentricity , and inclination .
There are \lesssim 400000 MBCs in the main belt brighter than H _ { V } = 21 ( \sim 150 -m in diameter ) and the MBC : MBA ratio is \lesssim 1:400 .

We further comment on the ability of observations to meaningfully constrain the snow line ’ s location .
Under some reasonable and simple assumptions we claim 85 % confidence that the contemporary snow line lies beyond 2.5 AU .