INFORMATION ABOUT BhamUIB SUBMISSION
------------------------------------

We use a multiple source MCMC algorithm to search for WD binaries in 
the Full Galaxy data set. The search is done by splitting the frequency
band in small segments (~tens of frequency bins (1 yr^{-1})). Since our 
algorithm is still under development, we decided JUST TO ANALYSE THREE 
PARTICULAR FREQUENCY REGIONS:

         a)  0.3 mHz <= f <= 0.4 mHz
         b)  0.9 mHz <= f <= 1.0 mHz
         c)  1.6 mHz <= f <= 1.7 mHz

trying to get from there all the resolvable sources. So, during the
evaluation it's important to notice that we didn't look for any source
with a frequency outside the 3 regions specified above.

We report 493 sources (141 in region (a); 169 in region (b) and 183 in (c))
with an estimation of all their parameters. So, we are submitting files with
493 lines, one per each source. The natural output from an MCMC algorithm
would be the probability density functions, but due to the 
impossibility of submitting them for all sources, we translate their 
information as the mean, mode (or maximum value of the marginalised posteriors)
and 68.2% (1-sigma) probability interval. This information is given in four
different files:

BhamUIB_means.dat   --> mean values
BhamUIB_maxs.dat    --> maxima of the posterior
BhamUIB_lowint.dat  --> Low bound of the 68.2% probability interval
BhamUIB_highint.dat --> High bound of the 68.2% probability interval

Each line corresponds to a source and the 8 columns correspond to the different
parameters: 

frequency (Hz)
fdot (Hz/s)
latitude(rad)
longitude(rad)
log10(A)
iota(rad)
polarisation angle (rad)
initial_phase (rad)

We only searched for low frequency signals and we considered negligible the 
frequency derivative (fdot), i.e.: in our search we fixed fdot = 0.

When mean and max values are different is because of the multimodal structure 
of the posterior distribution, and in those cases THE MAX VALUES give a better
representation of the actual parameters, so THESE ARE THE ONES THAT SHOULD BE 
CONSIDERED IN THE EVALUATION EXERCISE.

---

We are submitting a fifth file with our estimation of the noise PSD at the
different frequency windows we have analysed (BhamUIB_noise.dat). This noise
level would represent the sum of the instrumental and confusion contributions.
We took into account the cyclostationarity of the noise, and that's why we get
an average estimation every 6,5 weeks (2^18*15 sec). The submitted file 
contains the frequency value in the first column and the estimation of the
noise level at the eight epochs that complete a whole year, in the other 
8 columns.