An Update on the Markarian 501 Search

(A preliminary work: For Collaborators Only!!!)
22 Dec 1997

This memo is an update for the collaboration of the stripping, reconstruction
and analysis of  the Markarian 501 data. I will present here a description
of the stripping method and describe the reconstruction and analysis
performed to date.


The process of removing the subset of raw Milagrito data required for a
Markarian 501 analysis is what I call "stripping". This was done as follows:

1) The Milagrito raw data tapes were loaded into the tape library and
     read in their entirety. Events collected at times in which the zenith
     RA was within +/- 80 deg of the Markarian 501 RA were retained.

2) For runs in which there was no online reconstruction, all retained
     events were reconstructed offline. Events with declinations within
     +/- 15 deg of the Markarian 501 DEC were stripped.

3) For runs where the data was processed online, events with a declination
     within +/-25 deg of Markarian 501 were reconstructed offline. If
     the offline reconstruction yields a declination within +/- 15 deg of
     Markarian 501, then the event is stripped.

4) About 15% of all events are stripped from the raw data with this method.

To date, data tapes 1-124 have been stripped. Three tapes were either partially
read or skipped due to tape read errors. My expectation is that these tapes are
mostly recoverable. Two tapes were written with a data format not compatable
with the current version of DataRead. These tapes should be easy to recover
with modifications to the software.This subset includes data from the turn on
date (Feb 8) to Jun 19. This is about 60% of the data that I was authorized
to analize at the Maryland collaboration meeting.

The stripped data consumes about 250 Gb of data and has been saved to
8 high density DLT tapes.


The stripped data were reconstructed using the standard offline ChiSq
angle fitter. A curvature correction of 4ns/100m was applied in the plane
of the pond with the standard offline function Curve.c. The processed
data was saved to disk and resides at /cosmo/mrk501/proc/proc. Feel
free to analize the data with your favoirite time sloshing program.

As the stripping process was systematic, I restripped 8 of the 10 days
from the previous Markarian data set. I can use these twice stripped days
as a test stripping and reconstruction processes.


I used the same program to analize these data as was presented at the
Maryland collaboration meeting. It is a "standard" time sloshing method
ala the NIM article published by a subset of the Cygnus collaboration.

Presented below are significance plots of the 8 overlaping days for the new
and old processed data sets. Both were reconstructed with a 4ns/100m
curvatrure correction.  Events are considered with theta<45 deg and number
of tubes in the fit greater than 40. The total bin size is 1.5 deg.
The RA dimension has been compacted so that the bins are square in the
local coordinate system. Both plots show a maximum near RA:+1.0 "deg"
and DEC:+0.6 deg. There have been some minor changes to the offline code
since the old data was reconstructed, but qualitatively, the plots are nearly


With the 8 overlapping days subtracted and the hypothosis that there is a souce at
the posistion identified in the above plots, I then analized the independant
data set. The entire data set excluding the 8 overlaping days contains about
70 days of integrated exposure. The significance plot for the 70 days is shown

This result is hardly difinitive or compelling. It is interesting to note that for a
larger bin size (2.3 deg total bin size) the resulting significance map is


The plan is simple. I must complete the stripping process, before any definitive
conclusions should be drawn. It should take about 3 weeks finish the stripping.
Stay tuned.