astro-ph0403097/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[manuscript]{aastex}
3: %\usepackage{aastexug}   % User guide style customizations
4:   
5: 
6: %\received{2004 January 27}
7: 
8: \begin{document}
9: 
10: \title{TeV Gamma-Ray Survey of the Northern Hemisphere Sky Using the Milagro Observatory}
11: 
12: \author{R.~Atkins,\altaffilmark{1,10} W.~Benbow,\altaffilmark{2,11} D.~Berley,\altaffilmark{3} E.~Blaufuss,\altaffilmark{3} 
13: J.~Bussons,\altaffilmark{3,12} D.~G.~Coyne,\altaffilmark{2} 
14: %R.~S.~Delay,\altaffilmark{9} 
15: T.~DeYoung,\altaffilmark{2,3} 
16: B.~L.~Dingus,\altaffilmark{4} D.~E.~Dorfan,\altaffilmark{2} R.~W.~Ellsworth,\altaffilmark{5} 
17: %A.~Falcone,\altaffilmark{8, 17} 
18: L.~Fleysher,\altaffilmark{6} R.~Fleysher,\altaffilmark{6} G.~Gisler,\altaffilmark{4} M.~M.~Gonzalez,\altaffilmark{1} 
19: J.~A.~Goodman,\altaffilmark{3} T.~J.~Haines,\altaffilmark{4} E.~Hays,\altaffilmark{3} C.~M.~Hoffman,\altaffilmark{4} 
20: L.~A.~Kelley,\altaffilmark{2} 
21: C.~P.~Lansdell,\altaffilmark{3}
22: J.~T.~Linnemann,\altaffilmark{7}
23: %J.~McCullough,\altaffilmark{2,18} 
24: J.~E.~McEnery,\altaffilmark{1,13} R.~S.~Miller,\altaffilmark{8,14}
25: A.~I.~Mincer,\altaffilmark{6} M.~F.~Morales,\altaffilmark{2,15} P.~Nemethy,\altaffilmark{6} D.~Noyes,\altaffilmark{3} 
26: J.~M.~Ryan,\altaffilmark{8} F.~W.~Samuelson,\altaffilmark{4} 
27: %M.~Schneider,\altaffilmark{2} 
28: %B.~Shen,\altaffilmark{16} 
29: A.~Shoup,\altaffilmark{9} G.~Sinnis,\altaffilmark{4} A.~J.~Smith,\altaffilmark{3} G.~W.~Sullivan,\altaffilmark{3} 
30: %K.~Wang,\altaffilmark{14,19} 
31: %M.~Wascko,\altaffilmark{14,20} 
32: D.~A.~Williams,\altaffilmark{2} S.~Westerhoff,\altaffilmark{2,16} 
33: M.~E.~Wilson,\altaffilmark{1} X.W.~Xu\altaffilmark{4} and G.~B.~Yodh\altaffilmark{9}} 
34: 
35: \altaffiltext{1}{University of Wisconsin, Madison, WI 53706}
36: \altaffiltext{2}{University of California, Santa Cruz, CA 95064}
37: \altaffiltext{3}{University of Maryland, College Park, MD 20742}
38: \altaffiltext{4}{Los Alamos National Laboratory, Los Alamos, NM 87545}
39: \altaffiltext{5}{George Mason University, Fairfax, VA 22030}
40: \altaffiltext{6}{New York University, New York, NY 10003}
41: \altaffiltext{7}{Michigan State University, East Lansing, MI 48824}
42: \altaffiltext{8}{University of New Hampshire, Durham, NH 03824} %-3525}
43: \altaffiltext{9}{University of California, Irvine, CA 92717}
44: \altaffiltext{10}{Now at University of Utah, Salt Lake City, UT 84112}
45: \altaffiltext{11}{Now at Max Planck Institute, Heidelberg, Germany}
46: \altaffiltext{12}{Now at Universite de Montpellier II, Montpellier, France}
47: \altaffiltext{13}{Now at NASA Goddard Space Flight Center, Greenbelt, MD 20771}
48: \altaffiltext{14}{Now at University of Alabama in Hunstville, Huntsville, AL 35899}
49: \altaffiltext{15}{Now at Massachusetts Institute of Technology, Cambridge, MA 02139}
50: \altaffiltext{16}{Now at Columbia University, New York, NY 10027}
51: %\altaffiltext{16}{University of California, Riverside, CA 92521}
52: %\altaffiltext{17}{Now at Purdue University, West Lafayette, IN 47907}
53: %\altaffiltext{18}{Now at Cabrillo College, Aptos, CA 95003}
54: %\altaffiltext{19}{Now at Stanford Linear Accelerator Center, Stanford, CA 94309}
55: %\altaffiltext{20}{Now at Louisiana State University, Baton Rouge, LA 70803}
56: 
57: \begin{abstract}
58: Milagro is a water Cherenkov extensive air shower array that continuously monitors
59: the entire overhead sky in the TeV energy band.  The results from an 
60: analysis of $\sim$3 years of data (December 2000 through November 2003) are presented.  
61: The data has been searched for steady point
62: sources of TeV gamma rays between declinations of 1.1 degrees and 80 degrees.  
63: Two sources are detected, the Crab Nebula and the active galaxy Mrk 421.  
64: For the remainder of the Northern hemisphere we set 95\% C.L. upper limits between 275 and 600 mCrab
65: (4.8-10.5$\times 10^{-12}$ cm$^{-2}$ s$^{-1}$) above 1 TeV for 
66: source declinations between 5 degrees and 70 degrees. Since the sensitivity of
67: Milagro depends upon the spectrum of the source at the top of the atmosphere,
68: the dependence of the limits on the spectrum of a candidate source
69: is presented.   Because high-energy gamma rays from extragalactic sources
70: are absorbed by interactions with the extragalactic
71: background light the dependence of the flux limits on the redshift
72: of a candidate source are given. The upper limits presented here are
73: over an order of magnitude more stringent than previously published limits from TeV gamma-ray all-sky surveys.  
74: \end{abstract}
75: 
76: \keywords{gamma rays: observations --- surveys --- galaxies: active}
77: 
78: %\maketitle
79: 
80: \section{Introduction}
81: \label{sec:Introduction}
82: 
83: Sources of very-high-energy (VHE, $>$100 GeV) gamma rays are observed to be non-thermal in nature and are 
84: typically the sites of
85: particle acceleration.  This acceleration is thought to occur in astrophysical
86: shocks such as those believed to exist in plerions \citep{deJager1992}, supernova remnants \citep{Volk2000}, 
87: active galactic nuclei \citep{Blandford1978}, 
88: and galaxy clusters \citep{Loeb2000} (among other
89: sources).  These shocks may accelerate protons or electrons, both of which lead to the emission
90: of gamma rays.  Since gamma rays are unaffected by the magnetic fields that
91: pervade the Galaxy and the Universe, they can be used to pinpoint the sites of particle acceleration.
92: In addition to these ``classical'' astrophysical sources of VHE gamma rays, other more 
93: exotic objects such as primordial black holes, topological defects, and the decay of relic particles from the
94: big bang may also emit VHE gamma rays.  Perhaps of most interest is the possible existence of
95: a new type of source that has yet to be postulated.  A comprehensive survey of the sky sensitive to emission
96: at all time scales is necessary to detect the many possible sources.  
97: The analysis presented in this paper 
98: is part of an ongoing effort by the Milagro collaboration to search the entire
99: northern hemisphere for such objects.  The search for short bursts of TeV gamma rays has been
100: addressed in a previous paper \citep{Atkins2004}.  The analysis presented here deals specifically with steady 
101: point sources of TeV gamma rays.  
102: 
103: It has been over a decade since the discovery of the first source of VHE gamma rays, the Crab
104: Nebula \citep{Weekes89}.  Since that time there have been seven other confirmed sources of TeV gamma rays \citep{Horan2003}, 
105: six of which lie in the northern hemisphere.  With the exception of the Crab Nebula and the supernova
106: remnant PKS 1706-44, these objects
107: are all active galactic nuclei of the blazar class \citep{Horan2003}.   All of these objects have been discovered by atmospheric
108: Cherenkov telescopes searching for counterparts to sources discovered at lower energies.
109: In contrast, the EGRET instrument detected over 270 objects emitting high-energy gamma rays above 100 MeV \citep{Hartman99}.
110: One hundred and seventy of these objects are not identified at other wavelengths.  The VHE regime is a natural
111: energy band to search for counterparts to these objects.  The small field-of-view and low duty factor
112: of the atmospheric Cherenkov telescopes (ACTs) make comprehensive sky surveys difficult to perform.
113: As a result, very few comprehensive surveys of the VHE sky have been performed to date.  
114: The first VHE survey was performed by a non-imaging ACT
115: \citep{Helmken79, Weekes79} in 1979.  The Milagrito instrument (a prototype to Milagro, with no background rejection
116: capability and a higher energy threshold) also performed a survey of the northern hemisphere \citep{Atkins2001} 
117: and set limits of $\le$3 Crab from any point source in the northern hemisphere.  More recently
118: flux limits of 4-9 times that of the Crab Nebula (above 15 TeV) have been published by the AIROBICC 
119: collaboration \citep{Aharonian2002}.  (1 Crab is equivalent
120: to an integral flux above 1 TeV of $F(>$1 TeV)$=1.75 \times$10$^{-11}$ cm$^{-2}$ s$^{-1}$.) 
121: The limits presented here are over an order of magnitude more stringent than 
122: these previous surveys.  
123: 
124: Since many of the confirmed sources of VHE gamma rays are extragalactic the limits
125: must account for the absorption of TeV 
126: gamma rays by interactions with the extragalactic background light (EBL) \citep{Primack2000, Stecker2002, Kneiske2002}.  
127: The EBL is comprised of 
128: visible radiation emitted by stars and infrared radiation emitted by dust due to reprocessed starlight. 
129: Since direct measurements
130: of the intensity and spectrum of the EBL are problematic due to the foreground light
131: from our galaxy, a model is 
132: used to determine the effect of the EBL on the observed spectrum at earth from a distant source.  
133: 
134: Before employing the results presented here the limitations
135: of this survey need to be understood.  First, the limits presented only apply to point sources, 
136: not to extended objects, such as the galactic plane.  Second, they only apply to the average 
137: VHE emission during the time period over which the data was obtained.  
138: 
139: \section{The Milagro Observatory}
140: \label{sec:Milagro}
141: The Milagro gamma-ray observatory has 723 photomultiplier
142: tubes (PMTs) submerged in a 24-million-liter water reservoir.  The detector is
143: located at the Fenton Hill site of Los Alamos National Laboratory, about 35
144: miles west of Los Alamos, NM, at an altitude of 2630 m asl (750 g/cm$^2$).  The
145: reservoir measures 80m $\times$ 60m $\times$ 8m (depth) and is covered by a light-tight
146: barrier.  The PMTs are arranged in two layers,
147: each on a 2.8m x 2.8m grid.  The top layer of 450 PMTs (under 1.4 meters of
148: water) is used primarily to reconstruct the direction of the air shower.  By
149: measuring the relative arrival time of the air shower across the array the
150: direction of the primary cosmic ray can be reconstructed with an accuracy of
151: roughly $0.75^\circ$.  The bottom layer of 273 PMTs (under 6 meters of water) is used
152: primarily to discriminate between gamma-ray-initiated air showers and hadronic
153: air showers.  
154: 
155: The discrimination of the cosmic ray background is described in detail in \citep{Atkins2003}.
156: The background rejection uses a parameter known as compactness which is equal to the number of PMTs
157: in the bottom layer with more than 2 photoelectrons (PEs) divided by the number of PEs in the PMT with
158: the largest number of PEs (in the bottom layer).
159: A requirement that the compactness be greater than 2.5 retains 51\% of the
160: gamma rays while removing 91.5\% of the cosmic ray background (for events with more than 50 PMTs in the
161: trigger), resulting in an improvement in the gamma-ray
162: flux sensitivity of 1.7.  Compactness greater than 2.5 was required for this analysis.
163: 
164: Milagro began data taking in 1999.  The data set presented here
165: begins on 15 December 2000 and ends on 25 November 2003.  Prior to 25 January 2002 the trigger 
166: was a simple multiplicity trigger requiring 60 PMTs to record a pulse greater than
167: 1/6 of a PE within 180 ns.  After this date a risetime criterion was imposed on the
168: cumulative timing distribution of struck PMTs.
169: If a PMT signal is greater than 1/6 of a PE a trigger signal with an amplitude of 6.25 mV and width of 180 ns is generated.
170: The signals from the 450 PMTs in the top layer are summed and sent to a VME trigger card.  
171: An 80 MHz flash analog-to-digital converter (FADC) digitizes the trigger signal.  
172: For each event that exceeds a ``pre-trigger'' condition the data from the FADC
173: are stored in a FIFO.  The risetime of the trigger signal is defined as the time taken for the trigger signal 
174: to go from 12.5\% of its peak value to 88.5\% of its peak value.
175: For events with more than 75 PMTs no risetime requirement was imposed.  Events with more than 52 PMTs 
176: were required to have a risetime less than 87.5 ns and events with greater than 26 PMTs were 
177: required to have a risetime less than 50 ns.  The requirement on the risetime removes event triggers due to single muons
178: at low multiplicity.  The lower multiplicity requirement increases the effective area to 
179: low-energy events, which improves the sensitivity to gamma-ray bursts.  
180: The risetime requirement has a minimal impact on the sensitivity to Crab-like sources.  
181: 
182: The trigger rate during this time varied from 1500 Hz to 1800 Hz.  A total of 1009 days of data were acquired
183: during this period for a detector ``on-time'' of 92\%.  The data are calibrated and reconstructed (to
184: give the core position, direction, and information used in rejecting the background of the incoming particle)
185: in real time.  Except for selected regions of the sky, only the reconstructed
186: information is saved to disk.  This analysis utilizes the reconstructed
187: data set.  The start date of the analysis is determined by the date that the compactness parameter
188: and an improved shower-core fitter were included in the reconstructed data.
189: 
190: \section{Analysis Strategy}
191: \label{sec:Analysis}
192: 
193: Two maps of the sky are constructed: a signal map (comprised of the actual numbers of
194: events coming from each bin in the sky) and a background map (comprised of an estimate
195: of the cosmic-ray background from each bin in the sky).  The maps are binned in $0.1\times 0.1$ degree bins.
196: To estimate the background a technique called ``direct integration" \citep{Atkins2003} is used.  
197: The method makes use of the fact that the earth rotates and that the
198: detector response is solely a function of local coordinates and time, and that the cosmic rays constitute an
199: isotropic background.
200: The underlying assumption of the method
201: is that the shape of the detector response (in local coordinates) does not vary over the period during which the
202: background is accumulated, which is two hours in this analysis.  This method naturally accounts for rate variations
203: in the detector and makes a high statistics measurement of the background (roughly 12 times as much
204: background as signal is accumulated for each point in the sky).
205: 
206: A bin of size 2.1 degrees in
207: declination ($\delta$) by $2.1/\cos(\delta)$ degrees in right ascension is used
208: to search the skymaps for evidence of a source of TeV gamma rays .  This choice of bin size is
209: based on the angular resolution of the detector as measured by the shadowing by the Moon of cosmic rays 
210: and Monte Carlo simulations of gamma-ray showers.  The $0.1\times0.1$
211: bins in the maps are summed to form these larger bins.  
212: Large bins are formed with centers at the center of each $0.1\times 0.1$
213: degree bin by summing the contents of all the small bins that intersect the large bin.  
214: The summed data from the signal map and background map are then compared.  The declination range of the search
215: is between 1.1 degrees and 80 degrees. The lower limit is determined by the center of the first large bin with
216: a low edge of zero degrees declination.  The upper limit is determined by the rotation of the sky, above
217: 80 degrees the hour angle interval used to generate the background is only 2.5 times the size of the signal region.
218: To calculate the significance of each excess or deficit the prescription of Li and Ma (eq. 17) \citep{LiMa} is used.
219: 
220: 
221: \section{Results of Survey}
222: \label{sec:Results}
223: 
224: Figure 1a shows the distribution of excesses (in standard deviations) for all bins in the
225: map comprised of the entire 3 years of data.  There is a statistically significant surplus of points
226: in the sky with greater than 4 standard deviation excess.  The bulk of this surplus can be
227: attributed to two known sources of TeV gamma rays: the Crab Nebula and the active galaxy Mrk 421.
228: Figure 1b shows the distribution of excesses with 2-degree regions around the Crab Nebula and Mrk 421 removed.
229: This distribution is consistent with expectations from random background fluctuations.  A Gaussian fit to this
230: distribution has a 
231: mean of -2.9$\times$10$^{-3}$ and a sigma of 0.987, consistent with the number of independent entries in the 
232: histogram.
233: 
234: Figure 2 shows the map of the northern hemisphere in TeV gamma rays for the data set.
235: The Crab Nebula (R.A.=83.64, Dec=22.01) and the active galaxy Mrk 421 (R.A.=166.11, Dec=38.21, z=0.03) 
236: are clearly visible in the map.  The significance of the excess at the location of the Crab is
237: 6.3 standard deviations and at the location Mrk 421 the significance of the
238: excess is 4.4 standard deviations. (After correcting for instrumental dead-time and other known effects the results 
239: on the Crab Nebula result in a measured
240: gamma ray rate of $10.0 \pm 1.4$ events/day.  This is slightly different from the result
241: given in Atkins et al.(2003) of $10.7 \pm 1.6$ events/day due to the fact that
242: different calibrations where used in the online reconstruction for some time periods.) 
243: Table 1 gives the location of all regions
244: with an excess of greater than 4 standard deviations. Only the pixel with the largest significance from
245: each independent region is listed in the table.  Near the Crab Nebula the pixel with the greatest
246: significance is at the location of the Crab Nebula, while near Mrk 421 it is 0.5 degrees from the 
247: position of Mrk 421.  
248: The number of these regions is consistent with the expected fluctuations in the background given the large number of trials
249: incurred in examining the entire sky.  Therefore no claim is made that these regions are sources of TeV gamma rays,
250: though they may be interesting regions for followup observations with the more sensitive ACTs.
251: The Whipple Observatory has performed a follow-up observation of the region near R.A.=79.9 Dec=26.8 between November 2002
252: and January 2003 and has reported
253: an upper limit of 90 mCrab \citep{Falcone2003}, below the sensitivity of this survey.  
254: 
255: Both the energy response and the sensitivity to gamma ray sources
256: of Milagro are dependent upon the declination of the source.  
257: Figure 3 shows the median energy of gamma rays that trigger Milagro (determined from Monte Carlo simulation), 
258: satisfy the compactness criterion, and are reconstructed
259: within 1.2 degrees of their true direction, averaged over a complete transit (i.e. 24 hours of observation) 
260: as a function of the declination of the source for several source spectral indices. The requirement that the
261: direction of the particle be reconstructed within 1.2 degrees of its true direction is imposed to account for the
262: bin size used in the analysis (a 2.1 degree wide square bin has the same area as a 1.2 degree circular bin).  
263: 
264: Establishing upper limits to the gamma ray flux from any given 
265: point in the sky is straightforward for galactic (nearby) sources of TeV gamma rays.  
266: The prescription of Helene \citep{Helene83} is used to calculate 
267: the confidence limits on the number of signal events from each region of the sky.
268: Since the response of Milagro is energy dependent, the flux upper limits obtained from these data
269: are dependent upon the energy spectra of the possible sources of TeV gamma rays.  Figure 4 is a 
270: 2-dimensional map of the sky with the 95\% C.L. upper limits to the flux given at each point.  
271: These limits are based on the assumption of source spectra proportional to $E^{-2.59}$ (i.e. similar to the 
272: Crab Nebula \citep{Aharonian2002, Atkins2003} at these energies).  
273: In order to translate the observed upper limits on the number of excess events from a given location into
274: an upper limit on the flux of gamma rays, the detector response is normalized using the results from the Crab Nebula.
275: This procedure accounts for the dead time of the detector, calibration errors, and other systematic effects. 
276: At declinations near 36 degrees (the latitude of Milagro) the 95\% C.L. upper limits are on average 275 mCrab or
277: $F$($>$1 TeV)$<$4.8$\times$ 10$^{-12}$ cm$^{-2}$ s$^{-1}$.  As the declination of the source increases or
278: decreases from this value the source spends less time near zenith (where the atmospheric overburden is least
279: and the response of Milagro is best) and the flux upper limits increase.  At declinations of 5 degrees or 65 degrees 
280: the average 95\% C.L. upper limits are of the order 600 mCrab ($F(>$1 TeV) $<$ 1.05$\times$ 10$^{-11}$ cm$^{-2}$ s$^{-1}$).
281: Figure 5 gives the factor by which these upper limits must be multiplied for potential sources with different
282: spectral indices. The use of this figure is best
283: explained with an example.  From Figure 4 obtain an upper limit from a location, R.A.=180., Dec=40 ($F(>$1 TeV)$< 291$ mCrab).
284: To find the flux upper limit for a source with a differential spectral index of -2.0, find the curve for such a
285: source in Figure 5 (the solid curve).  At the declination of the source the y-axis value (0.41) 
286: is the amount by which the flux upper limit
287: from Figure 4 must be multiplied to give the flux upper limit for this source (119 mCrab).
288: 
289: For an extragalactic source one must also account for the absorption of the gamma rays due to interactions
290: with the EBL \citep{Primack2000, Stecker2002, Hartman99}.  In the absence of a
291: reliable measurement of the EBL, a model of the intensity and energy spectrum is used.  Figure
292: 6 shows the effect of the absorption due to the EBL on the upper limits given in Figure 4.
293: In this figure the ``baseline'' model of Stecker and De Jager \citep{Stecker2002} is used to calculate the
294: effect of the absorption of energetic gamma rays and the source spectra are assumed to be
295: proportional to $E^{-2.59}$.  Using a procedure similar to that
296: described above for Figure 5, one can find the flux upper limit for sources at different redshifts.  
297: For example, using the same location as above (R.A.=180 Dec=40) but a source at a redshift of 0.03 (with a 
298: differential spectral index of -2.59), the curve for z=0.03 in Figure 6 
299: has a value of ~2.4, resulting in a flux upper limit for this location and redshift of 700 mCrab.
300: This upper limit is the normalization of the power law spectrum {\em of the unabsorbed source} at the top
301: of the atmosphere.  Assuming a source with an intrinsic spectrum (before absorption by the EBL)  represented by,
302: \begin{equation}
303: \frac{dN}{dE} = I_0 {E_{TeV}}^{-\alpha}
304: \end{equation}
305: then, for this source $I_0$ would be 2.4 times that of a galactic source with the same spectral index.
306: (Note that to calculate the absolute luminosity of the source
307: one must also multiply by the square of the distance to the source.  This has not been accounted for in Figure 6.)
308: Because the absorption of TeV gamma rays by the EBL distorts the source spectrum before
309: it reaches the earth one can not use Figure 6 in series with Figure 5.  In general the effect on the sensitivity due to 
310: different source spectra is smaller for distant sources, since the EBL tends to make the distant sources look more alike
311: regardless of their spectra.  Figure 7 gives an example of the relative sensitivity to sources at different redshifts with
312: different spectra.  Figure 7a is similar to Figure 6 but the sources here are assumed to have a differential photon spectrum
313: proportional to $E^{-2}$ and in Figure 7b a spectrum proportional to $E^{-3}$.  The figure gives the ratio of the flux upper limit
314: for a source at the given redshift, declination, and spectrum to a local source (z=0.0) with a differential spectral index of
315: -2.59.  A perhaps surprising feature of this figure is that for
316: more distant sources a source with a harder intrinsic spectrum is required to have a larger luminosity than a source with a softer
317: spectrum for Milagro to make a detection.  This is due to the effect of the EBL, where the high energy photons that the source
318: is required to emit (by the model) are absorbed in transit and do not affect the ability of Milagro to observe the source
319: but do count as part of the intrinsic source luminosity. 
320: 
321: 
322: \section{Conclusions}
323: \label{sec:Conclusions}
324: A complete survey of the northern hemisphere (declination $1.1^{\circ}$ to $80^{\circ}$) for point sources
325: of TeV gamma rays has been performed.  These limits apply to the average flux level during the roughly 
326: 3 year period from December 15, 2000 
327: through November 25, 2003.  The average 95\% C.L. upper limits range from 275 mCrab to 600 mCrab depending upon the 
328: declination of the source and are over an order of magnitude more restrictive than previous limits.  
329: A prescription has been given to calculate the corresponding
330: upper limits for sources with different spectra and for extragalactic sources.  For sources with differential spectral
331: indices of -2.0 the upper limits are 57\% lower.  For a source at a redshift of 0.03
332: the flux limits are a factor of 2.4 larger.  While these limits are the best available to date, 
333: Milagro has recently been completed with the construction of an array of 175 water tanks surrounding 
334: the central reservoir.  A comparable dataset, with this now complete Milagro detector, 
335: would improve these limits by a factor $\sim$2.
336: 
337: \acknowledgments
338: 
339: We gratefully acknowledge Scott Delay and Michael Schneider for their dedicated efforts in the 
340: construction and maintenance of the Milagro experiment.  This work has been supported by the 
341: National Science Foundation (under grants 
342: PHY-0070927, %UCSC
343: -0070933, %UCI
344: -0075326, %Milagro Operations
345: -0096256, %UW-Madison
346: -0097315, %LANL via UMD 
347: -0206656, %NYU current; previous is PHY-9901496
348: -0302000, %UMD
349: and
350: ATM-0002744) %UNH
351: the US Department of Energy (Office of High-Energy Physics and 
352: Office of Nuclear Physics), Los Alamos National Laboratory, the University of
353: California, and the Institute of Geophysics and Planetary Physics.  
354: 
355: 
356: \begin{thebibliography}{}
357: 
358: \bibitem[Aharonian et al.(2002)]{Aharonian2002} Aharonian, F., et al., 2002, \aap, 390, 39
359: \bibitem[Atkins et al.(2001)]{Atkins2001} Atkins, R., et al., 2001, \apj, 558, 477
360: \bibitem[Atkins et al.(2003)]{Atkins2003} Atkins, R., et al., 2003, \apj, 595, 803
361: \bibitem[Atkins et al.(2004)]{Atkins2004} Atkins, R., et al., 2004, \apjl, to appear
362: \bibitem[Blandford and Ostriker(1978)]{Blandford1978} Blandford, R.~D. and Ostriker, J.~P., 1978, \apjl, 221, 29
363: \bibitem[De Jager and Harding(1992)]{deJager1992} De Jager, O.~C. and Harding, A.~K., 1992, \apj, 396, 161
364: \bibitem[Falcone et al. (2003)]{Falcone2003} Falcone, A., et al., 2003, in Proc. 28th International Cosmic Ray Conference, ed. T. Kajita, Y. Asaoja, A. Kawachi, Y. Matsubara, and M. Sasaki (Tokyo, Japan), Vol. 5, 2579
365: \bibitem[Li and Ma(1983)]{LiMa} Li, T.~P. and Ma, Y.~Q., 1983, \apj, 272, 317
366: \bibitem[Loeb and Waxman(2000)]{Loeb2000}Loeb, A. and Waxman, E., 2000, Nature, 405, 156
367: \bibitem[Hartman et al.(1999)]{Hartman99} Hartman, R.~C., et al., 1999, \apjs, 123, 79
368: \bibitem[Helene(1983)]{Helene83} Helene, O., 1983, NIM, 212, 319
369: \bibitem[Helmken et al.(1979)]{Helmken79} Helmken, H.~F., Horine, E., and Weekes, T.~C., 1979, Proceedings of the 19th ICRC, 1, 120
370: \bibitem[Horan and Weekes(2003)]{Horan2003} Horan D. and Weekes, T.~C., 2003, astro-ph/030391v1
371: \bibitem[Kneiske, Mannheim, and Hartmann(2002)]{Kneiske2002} Kneiske, T.M., Mannheim, K., and Hartmann, D.H., 2002, \aap, 386, 1
372: \bibitem[Primack et al.(2000)]{Primack2000} Primack, J.~R., Somerville, R.~S., Bullock, J.~S., and Devriendt, J.~E.~G., 2000,AIP Conference Proceedings, 558, 463, AIP: F.~A. Aharonian and H.~J. V{\" o}lk
373: \bibitem[Stecker and De Jager(2002)]{Stecker2002} Stecker, F. and De Jager, O.~C., 2002, \apj, 566, 738
374: \bibitem[Berezhko and V{\" o}lk(2000)]{Volk2000}Berezhko, E.~G. and V{\" o}lk, H.~J., 2000, APh, 14, 201
375: \bibitem[Weekes et al.(1979)]{Weekes79} Weekes, T.~C., Helmken, H.~F., and L'Heureux, J., 1979, Proceedings of the 19th ICRC, 1, 126
376: \bibitem[Weekes et al.(1989)]{Weekes89} Weekes, T.~C., et al., 1989, \apj, 342, 379
377: \end{thebibliography}
378: 
379: 
380: \clearpage
381: 
382: \begin{figure}
383:   \figurenum{1}
384:   \epsscale{1}
385: \plotone{f1.eps}
386:   \caption{(a) The distribution of significances of the excesses and deficits in the analysis of the
387:   D.C. skymap of the northern hemisphere.  (b) The same data with 2 degree regions around
388: the Crab Nebula and Mrk 421 removed.  The dotted curves are the best fit Gaussians to the data.}
389: \end{figure}
390:                                                                                                                              
391: \begin{figure}
392:   \figurenum{2}
393: \plotone{f2.eps}
394:   \caption{The northern hemisphere as seen in TeV gamma rays.  At each point the excess is summed over
395:   a 2.1 degree by $2.1/\cos(\delta)$ bin and the significance of the excess in standard deviations is 
396:   shown by the color scale.}
397: \end{figure}
398: 
399: \begin{figure}
400:   \figurenum{3}
401: \plotone{f3.eps}
402:   \caption{The median energy of gamma-ray events that trigger Milagro, pass the compactness cut, and 
403: are reconstructed with 1.2 degrees of their true direction as a function of source declination (this is 
404: equivalent to the 2.1 degree square bin used in the search).  The response
405: of Milagro is averaged over a complete source transit (i.e. one day's observation)
406: and the source differential spectral index, $\alpha$, for each curve is given in the legend.}
407: \end{figure}
408: 
409: \begin{figure}
410:   \figurenum{4}
411: \plotone{f4.eps}
412:   \caption{The 95\% C.L. upper limits on the integral flux of gamma rays above 1 TeV (assuming an $E^{-2.59}$
413: differential photon spectrum) from each point in the northern hemisphere.  The color scale on the right
414: is in units of the flux from the Crab Nebula.  To enhance the contrast of the figure only declinations below 75 degrees
415: are shown.}
416: \end{figure}
417: 
418: \begin{figure}
419:   \figurenum{5}
420: \plotone{f5.eps}
421:   \caption{The effect of the differential spectral index on the upper limits shown in Figure 4.   
422: The y-axis gives the ratio of the flux upper limit for a source with a differential spectral index as
423: indicated by the curve to a source with a differential spectral index of -2.59.  The use of this figure is described 
424: in the text.}
425: \end{figure}
426: 
427: \begin{figure}
428:   \figurenum{6}
429: \plotone{f6.eps}
430:   \caption{The effect of redshift on the upper limits shown in Figure 4.  These results
431: assume a source spectrum proportional to $E^{-2.59}$.  The y-axis is the ratio of the flux upper limit for a source
432: at the indicated redshift to a source at a redshift of zero.}
433: \end{figure}
434: 
435: \begin{figure}
436:   \figurenum{7}
437: \plotone{f7.eps}
438:   \caption{The effect of redshift on the upper limits shown in Figure 4.  Figure 7a
439: assumes a source spectrum proportional to $E^{-2}$ and 7b a source spectrum proportional to $E^{-3}$.
440: In both cases the y-axis is the ratio of the flux upper limit for the described source 
441: (spectral index, redshift, and declination) and a local source with a differential spectral index of -2.59.}
442: \end{figure}
443: 
444: \clearpage
445: 
446: \begin{deluxetable}{ccccccc}
447: \tabletypesize{\scriptsize}
448: \tablecaption{The locations of all regions with greater than a 4$\sigma$ excess.
449: Only independent regions are entered in the table.  Errors on the location
450: of any possible source are $\sim$0.5 degrees.  The two brightest points on the map
451: are due to the two detected sources: the Crab Nebula$^1$ and Mrk 421$^2$.  Upper
452: limits are not given for these two sources.
453: The units of right ascension and declination are decimal degrees.
454: The last column gives the 95\% C.L. upper limit to the flux in units of the Crab
455: flux.  \label{tbl-1}}
456: \tablewidth{0pt}
457: \tablehead{
458: \colhead{RA} & \colhead{DEC}   & \colhead{ON}   &
459: \colhead{OFF} & \colhead{Excess} & \colhead{Sigma} & \colhead{UL} 
460: }
461: \startdata
462: 0.3       &   34.3 &  3.12308e+06 &    3.11456e+06  &   8623   &  4.7  & 0.84 \\
463: 37.8      &   6.7  &  7.02166e+05 &    6.98667e+05  &   3498   &  4.0  & 1.8 \\
464: 43.6      &   4.8  &  5.85952e+05 &    5.82716e+05  &   3236   &  4.1  & 2.0 \\
465: 49.1      &   22.5 &  2.21431e+06 &    2.20813e+06  &   6175   &  4.0  & 0.87\\
466: 79.9      &   26.8 &  2.57841e+06 &    2.57025e+06  &   8161   &  4.9  & 0.97\\
467: 83.6$^1$  &   22.0 &  2.17188e+06 &    2.16222e+06  &   9665   &  6.3  & NA \\
468: 166.5$^2$ &   38.6 &  3.23552e+06 &    3.22467e+06  &   10850  &  5.8  & NA \\
469: 306.6     &   38.9 &  3.25329e+06 &    3.24531e+06  &   7983   &  4.2  & 0.78\\
470: 313.0     &   32.2 &  3.08380e+06 &    3.07548e+06  &   8320   &  4.5  & 0.85\\
471: 339.1     &   72.5 &  6.63534e+05 &    6.59727e+05  &   3807   &  4.2  & 3.02\\
472: 356.4     &   29.5 &  2.98656e+06 &    2.97910e+06  &   7455   &  4.1  & 0.84\\
473: \enddata
474: \end{deluxetable}
475: 
476: \end{document}
477: