1: %
2: %Ap J/ AJ paper on UNID survey
3: \documentclass[12pt,preprint]{aastex}
4: %\newcommand{\gtrsim}{\;_\sim^>\;}
5: %\newcommand{\lessim}{\;_\sim^<\:}
6: \shorttitle{Statistical Analysis of EGRET Blazars}
7: \shortauthors{Bloom}
8: \received{2007 March 14}
9: \begin{document}
10: \title{The Radio and Gamma Ray Connection of EGRET Blazars: Correlation, Regression, \& Monte Carlo Analysis }
11: \author{S. D. Bloom \altaffilmark{1}}
12: \affil{Department of Physics \& Astronomy, Hampden-Sydney College}
13: \affil{Box 821, Hampden-Sydney, Virginia 23943}
14: \altaffiltext{1}{Visiting Scientist, National Radio Astronomy
15: Observatory, Charlottesville, VA}
16: \begin{abstract}
17: A comprehensive statistical analysis of the broadband
18: properties of EGRET blazars is presented. This analysis includes sources identified as
19: blazars in the Sowards-Emmerd publications. Using this sample of 122 sources, we find that there is
20: a relationship $L_\gamma \propto
21: {L_r}^{0.77 \pm 0.03} $ as well as a correlation between $\alpha_{og}$ and $\alpha_{ro}$, and a correlation between radio luminosity and $\alpha_{og}$.
22: Through the use of Monte Carlo simulations, we can replicate the observed luminosity relationship if a synchrotron self-Compton model is assumed. However, this relationship can not be replicated if an external Compton scattering model is assumed. These differences are primarily due to beaming effects.
23:
24: In addition it has been determined that the intrinsic radio luminosity of the parent sample falls in the range $10^{21} < L < 10^{30}\, {\rm Watts\,Hz^{-1}}$ and that the bulk Lorentz factors of the source are in the range $ 1 < \Gamma < 30 $, in a agreement with VLBI observations.
25:
26: Finally, we discuss implications for GLAST, successfully launched in June 2008.
27: \end{abstract}
28:
29: \keywords{gamma rays: observations}
30:
31: \section{Introduction}
32: During the lifetime of the Energetic Gamma-Ray Experiment Telescope
33: (EGRET) instrument on board {\it Compton
34: Gamma-Ray Observatory} (CGRO) 271 sources were detected with 66
35: being confidently identified as blazars in the Third EGRET Catalog \citep{har99}. Many
36: differing statistical analyses of these gamma ray detected blazars
37: have been conducted, among them
38: \citet{fos98};\citet{muc97}. Some of these analyses concentrated on
39: the direct
40: statistical relationship between luminosities in the gamma-ray band and
41: radio bands \citep{sal96,ste93,fan98}. Of these, most have used single
42: dish radio data, but some have used VLBI fluxes \citep{zho97,mat97}, all at various
43: radio frequencies $>$ 1 GHz. In addition, \citet{muc97} and \citet{imp96} use
44: Monte Carlo simulations to aid in interpreting these
45: relationships. Though significant correlations are reported in all
46: of these works, \citet{muc97} show that in some cases these will result from
47: a combination of variability and selection effects. To
48: investigate the variability effects further, \citet{zha01} have compared their
49: statistical results using time averaged data for the entire sample to similar results for
50: a restricted sample for which data were available during high and
51: low states. They report a similar significant correlation in each
52: case and show that using time averaged data tends to under estimate
53: the underlying linear regression slope (as applied to the logarithmic data).
54:
55: Various models have been invoked to explain the origin of the
56: gamma-ray emission of blazars and specifically, the radio gamma-ray
57: correlation. Among these are the synchrotron self Compton (SSC) model
58: \citep{blo96,ghi85} and various models in which the source of seed
59: photons for scattering is from a source external to the jet (henceforth called ECS
60: models for ``external Compton scattering''). \citet{der93} use the
61: accretion disk as the source of soft photons whereas others \citep{ghi96}
62: use broad-line region clouds as the source of soft photons.
63: %Still others have focused on features of the
64: %spectral energy distributions (SEDs), and flux variability.
65:
66: Since the end of the EGRET mission there have
67: also been several reanalyses of the significance of identifications of
68: EGRET sources, particularly those of \citet{mat97,mat01} and
69: \citet{sow03,sow04}.The \citet{sow03,sow04} survey excludes sources
70: with $|b|<10^{\circ}$ (the Mattox papers exclude sources with $|b|<3^{\circ}$),
71: and thus may exclude additional sources that are thought to be blazars \citep{sgu04}.
72: Likewise, some work continues on identifying
73: other individual sources, such as 3EG J0416+3650, possibly identified with
74: 3C 111 \citep{sgu06}. In the eight years since the end of the CGRO
75: mission, there has been substantial
76: modification to the identifications given in the last catalog published by the EGRET
77: team \citep{har99}, so we present a comprehensive statistical
78: analysis including all potential identifications, using homogeneous
79: criteria for inclusion in the sample ({\S 2}). We have also included various
80: statistical techniques of survival analysis for the inclusion of
81: upper and lower limits in the data ({\S 3}). We later discuss whether or not the
82: results of the analyses are dependent on the precise source list.
83: The approach used in understanding the physical implications of our results is to use a Monte Carlo technique
84: \citep{lis97} to generate multiple simulated samples under differing assumptions of overall theoretical or
85: phenomenological model ({\S 4,5}), distribution of physical parameters such as
86: bulk Lorentz factor (e.g., Gaussian, power-law, etc.) and selection
87: effects. Additional physical implications of observed correlations and implications for the Gamma Ray Large Area Space Telescope (GLAST), successfully launched in June 2008, are discussed in {\S 6,7}.
88: \section{Data}
89: To construct a comprehensive list of identified gamma-ray blazars we
90: have started with the Third EGRET Catalog \citep{har99}. We note that
91: we are not re-evaluating the detections themselves. To add or
92: remove sources from this list of identifications using uniform criteria, we
93: have further used the guidance of \citet{sow03,sow04}. Though other
94: standards for inclusion or rejection do exist, we have used these
95: particular surveys to keep inclusion criterion standard. We have taken
96: the best blazar identifications with Figure of Merit (FoM) criterion greater
97: than 0.25 \citep{sow03,sow04}. The FoM used here is a statistic that
98: incorporates both radio and X-ray spectral information in evaluating
99: the probability that a particular radio source is a match to the
100: gamma-ray source. For sources with multiple possible
101: identifications, we include the identification with the
102: highest FoM value. Though the 0.25 cutoff is perhaps arbitrary,
103: it is clear that well below this limit (at about 0.1), the probability
104: for chance
105: associations greatly increases \citep{sow03}.
106: These authors and their current working group have not yet published an analysis for the sources between -40 and -90
107: declination. For these sources, we included all of the original
108: catalog identifications, with the caveat that the more
109: marginal identifications (i.e. counterparts with 8.4 GHz radio flux
110: density $< 0.5$
111: Jy) could possibly not hold up to the analysis using the FoM criterion. The total number of
112: such sources in this declination range is ten, with two below 0.5 Jy (the remainder all have
113: flux density greater than 1 Jy at 8.4 GHz). We also note that we could be excluding candidate identifications heretofore unmentioned
114: in this declination range.
115: A summary of the available data used in our statistical analysis has been compiled in Table 1. Column(1) indicates the
116: EGRET source name, column (2) indicates the best radio identification,
117: followed by the redshift in column(3). Columns (4)-(6) give the
118: monochromatic radio, optical and gamma ray luminosities,
119: respectively (all have units of $ {\rm joules\, sec^{-1} \,Hz^{-1}}$). If the redshift is unknown, we substitute with z=1.
120: There are no significant changes in the results if we use other
121: assumptions, such as z=0.5. To calculate the radio luminosity we use
122: the 8.4 GHz flux densities and spectral indices (Column (7))between 1.4 and 8.4 GHz of \citet{sow03,sow04} except for the
123: lowest declination sources discussed above. For these, we extract the
124: similar radio data from the NASA Extragalactic Database (NED) to calculate the spectral index between
125: 1.4 and 8.4 GHz. Henceforth we use ${F_{\nu}} \propto {\nu^{-\alpha}}$
126: for defining spectral indices. In calculating the optical luminosities we use
127: recent V magnitudes extracted from NED (when available), and we derive an
128: optical flux density and monochromatic luminosity from this magnitude using the conversions of \citet{bes79}.
129: \citet{sow03,sow04} also provide archival R and B magnitudes from the
130: USNO catalog \citep{mon03}, which we have used to interpolate a V value if we did
131: not otherwise have a reference for a V magnitude. In a few cases we
132: only had O magnitudes from the POSS-I survey or 2MASS infrared
133: magnitudes, and thus need to extrapolate to the V band.
134: For sources that
135: have no optical or near infrared magnitudes in the literature, and
136: that were not detected in the POSS-I, SERC-J, or POSS-II we
137: assume magnitude upper limits equal to the appropriate plate limits of
138: the survey (eg., O=21.5,E=20 for POSS-I \citet{mcm02})
139: and
140: derive upper limits to the optical luminosities accordingly.
141: %In
142: %several cases there are also g magnitudes from the Sloan Digital Sky
143: %Survey (SDSS),that only differ very slightly from V magnitudes.In
144: %our sample. The difference is negligible for our purposes.
145: The gamma-ray luminosities are determined from
146: the flux and spectral index (Column (8)) given in \citet{har99}. We
147: use the formula of \citet{tho96} to
148: derive an exact 400 MeV flux. If the spectral index is not
149: known, we assume it is 1.0 (note that we are referring to the energy
150: spectral index and not the photon index). In all cases we use the co-added flux
151: over all viewing periods. In the absence of simultaneous data at all
152: wavelengths, this is a better representation of broad band
153: properties than any single value. Each of the luminosities is
154: calculated using equation (2) of \citet{lis97} and assumes a flat $\Lambda$ cosmology with ${\Omega_m}=0.3$
155: and ${\Omega_{\Lambda}}=0.7$
156: \citep{per97}. Though the luminosity distance can be calculated via numerical integration, it is preferred to find
157: an analytic expression, especially for use in the Monte Carlo
158: simulations where such a calculation will have to be performed many
159: millions of times. \citet{pen99} offers such an option; however
160: this formula can differ from the numerically determined value by over
161: 5\%. Therefore, a polynomial fit to the numerical solutions is
162: determined for the redshift range z=0-5 using the Wolfram Research program
163: {\it Mathematica} to attain a more accurate approximation. When luminosity distances determined via this polynomial
164: fit were compared with the results of a direct numerical integration, the
165: agreement was well within 1 \% for the range $ 0.1 < z <4.0 $; however, at redshifts outside of this range, the disagreement is as large as 7\%.
166: Column(9) shows the broadband
167: spectral index, $\alpha_{ro}$
168: %\equiv{ {log { {F_{\nu o}} \over
169: % {F_{\nu r}}}} \over {{ log{ \nu_o \over\nu_r }}}}$ and Column(10)
170: gives the index
171: $\alpha_{og}$.
172: %\equiv{ { log {{F_{\nu g}} \over
173: %{F_{\nu o}}}} \over {{log{ \nu_g \over \nu_o}}}}$.
174: Column (11) has the gamma-ray variability parameter, $\delta_{var}$,
175: as determined by \citet{nol03}. Column (12) gives the references for
176: the data, where the first reference is for redshift, the second for
177: the radio data, the third for the optical magnitudes, and the fourth for
178: the gamma ray data.
179: \section{Statistical Analysis}
180: We have evaluated the strength and significance of correlation between
181: monochromatic luminosities using non-parametric methods, such as
182: Kendall's $\tau$ \citep{ken90}. These results are given in
183: Table 2. Column(1) gives the independent variable, Column(2) the
184: dependent variable, Column(3) the number in the sample,
185: Column(4) Kendall's $\tau$ statistic, Column(5) the probability of
186: the null result for this statistic, Column(6) the Spearman's $\rho$ statistic,
187: Column(7) the probability of the null result for this statistic, and in
188: Column(8) the
189: regression technique used. Here BJ refers to the Buckley-James method,
190: EM refers to the EM method and SB refers to Schmitt's binned
191: regression, as explained in the references that follow.
192: Column(9) gives the linear regression slope, and
193: Column(10) the linear regression Y-intercept. In some cases, we have upper and lower limits to measured
194: values, and to take this into account, we use the proper correlation
195: and regression techniques as discussed in \citet{fei85,iso86} and utilized in the ASURV
196: software package. In reporting correlation coefficients, probabilities and regression fits we adopt the format of \citet{muc97}. That is, we report three significant digits down to 0.100. Between 0.010 and 0.100 we report two significant digits, and for all numbers with values below 0.01, we only report with 1 significant digit.
197:
198: Most notably, we determine that there is a strong
199: correlation between gamma ray and radio luminosity (Figure 1 and Table
200: 2). In order to
201: determine whether the correlation is affected by including the sources with
202: FoM $<1$ (the lower confidence identifications), we recalculate the correlation and regression coefficients
203: while retaining only the 76 sources with FoM $>1$. The strength of the
204: correlation is similar, with a very slight steepening of the regression
205: slope (0.78 instead of 0.77, which is within the range of the
206: uncertainty of the correlation slopes). If the sources with declination below -40 $^\circ$ are also excluded, the the correlation slope flattens slightly to 0.73. In all though, correlation is not significantly altered by such changes to the source list.
207:
208: For some correlations, we must account for the effect of a third
209: variable, using partial correlation analysis \citep{pad92}. Results of our partial correlation analysis
210: are given in Table 3. Column(1) gives the independent variable, Column(2) the dependent variable,
211: Column(3) the third variable, Column(4) the number in sample, Column(5) Kendall's $\tau$, and
212: Column(6), the probability of the null result. Examples of this are the need to take into account the correlation of
213: all luminosities on redshift, and the dependence of several broad band
214: spectral indices on optical luminosity. In particular we note that the
215: strong correlation between gamma ray and radio luminosity is
216: significantly weaker, though still significant, once we account for
217: the partial correlation with redshift. Similarly, the correlation
218: between the $\alpha_{ro}$ and $\alpha_{o \gamma}$ still persists, even
219: after the effect of the common dependence on optical luminosity is
220: negated. We use the Kendall's $\tau$ coefficient for this purpose, since
221: to date it is the only coefficient for which there are methods to take
222: into account censored data \citep{akr96}. A detailed discussion of the
223: physical models for the luminosity correlation is addressed in {\S 4,
224: \S 5}. Further discussion of the remaining correlations is deferred
225: to {\S 6}.
226: \section{Physical Models}
227: Our approach in this section is to see if particular models can be used to
228: explain our statistical results above, particularly the radio and gamma-ray luminosity relationship.
229: A radio and gamma-ray luminosity correlation would naturally be
230: expected given any of the non-thermal inverse Compton models linking
231: low energy emission with gamma rays, such as in \citet{blo96}. As
232: noted above, this luminosity correlation
233: holds even after we take the effect of redshift into account. A linear or nearly linear correlation is possible within the synchrotron self-Compton
234: (SSC) model if the range of possible physical parameters, particularly the
235: optically thin spectral index, Thomson optical depth and electron
236: energy cutoffs do not vary
237: too greatly in the sample. This can be illustrated by following
238: \citet{mar87} for a spherical homogeneous source:
239:
240: $$ {L_{\nu c} \over L_{\nu s}}=
241: {{c_3(\alpha)}{N_0}R ln{({\nu_m \over \nu_2})}{{( {\nu_c \over
242: \nu_s})}^\alpha}}\eqno {(1)} $$
243:
244:
245: Where $L _{\nu c}$ is the frequency dependent Compton luminosity, $L_{\nu s}$ is the
246: frequency dependent synchrotron luminosity. $c_3$ is a function of
247: spectral index, tabulated in \citet{mar87}, $N_O$ is the normalization
248: factor of the energy dependent electron density distribution, and $R$
249: is the radius of a spherical source. $\nu_m$ refers to the spectral turnover frequency
250: in a plot of flux density versus frequency, and $\nu_2$ is the upper frequency cutoff
251: in a similar plot. A correction to this formula needs to be applied if the observed gamma-rays are
252: at a frequency beyond the high end spectral cutoff for Compton scattering, or if the synchrotron photons are observed at
253: a frequency at which the source is believed to be optically thick or partially opaque.
254: The logarithmic term only varies a small amount for large changes in the parameters,
255: so this term can usually be neglected relative to the others.
256: The function of $\alpha$ can vary over several orders of magnitude,
257: but can be roughly constant if the optically thin spectral index of
258: the sample is narrowly distributed. The factor of ${N_O \,R}$,
259: proportional to the Thomson optical depth, would also have to be
260: tightly constrained. Using a somewhat different
261: formulation, \citet{ghi98} calculate a particle injection
262: compactness parameter, which is also proportional to the Thompson optical
263: depth. They show that this parameter, though distributed over 4 orders
264: of magnitude for a
265: sample of 51 EGRET blazars, has a
266: clear peak value (over half of the blazars with a value within 1 order
267: of magnitude of the peak). Thus, in adopting a linear model for use
268: in the following Monte Carlo analysis, this parameter can
269: be drawn randomly from a sharply peaked distribution.
270: It is not as evident that these arguments can be extended to
271: a relationship such as that observed, ${L_{\gamma}} \propto {{L_r}^{0.77}} $. We explore this possibility below.
272:
273: Alternatively, the gamma-ray luminosity can be produced by an
274: external Compton scattering process
275: (ECS)\citep{der93,lis99a}. In this particular model the source of the soft
276: photons is the accretion disk, and these photons scatter off of a plasma of
277: relativistic electrons in a blob. Adapting the formulas of \citet{der92} to
278: our notation:
279:
280: $${L_{\nu c} \over L_{\nu,acc}}=
281: { {R^3\over 6r^2} {\sigma_T}{(m_ec^2)^{-2\alpha}}{N_0}{\delta^{2\alpha+4}}{(\nu/\nu_o)^{-\alpha}}}$$
282:
283: Here, the luminosity ratio is between the observed Compton luminosity
284: in the gamma-ray range and the observed accretion luminosity. $R$
285: refers to the radius of the blob, assuming it is spherical, and $r$
286: refers to the distance between the blob and the accretion
287: disk. $\sigma_T$ refers to the Thomson scattering cross-section,
288: ${m_e}c^2$ refers to the electron rest mass, $N_0$ has the same meaning as above. $\delta$ is the Doppler factor
289: to correct for bulk relativistic motion of the emitting plasma and is
290: defined by:
291:
292: $$\delta \equiv {1 \over {\Gamma(1-\beta cos\theta)}} \eqno{(3)} $$
293: Here $\beta$ has the usually meaning of ratio of bulk speed to that of light in a vacuum, and $\theta$ is the angle between direction of the bulk flow and the line of sight of the observer.
294: The power to which the Doppler factor is raised is dependent on the structure of the emitting region (discrete blobs vs. continuous jet). The formula above is correct for discretely emitting blobs. For a continuous jet, the power is reduced by one \citep{der95}.
295: It has been suggested that the
296: synchrotron luminosity can be proportional to the ultraviolet luminosity
297: of the seed photons \citep{ghi96}, thus providing a direct relationship
298: (though not necessarily linear) between the gamma-ray and radio luminosity within the framework of the
299: ECS model. Furthermore, the underlying radio and gamma-ray correlation may
300: also be explained by other means, since the gamma-ray luminosity in
301: all relativistic Compton scattering models is proportional to the
302: Thomson optical depth and the intrinsic
303: radio luminosity can be proportional to the Compton optical depth via
304: the parameter $N_0$ and $R$. Other source parameters,
305: such as the magnetic field, would have to fall into a narrow range for
306: the sample in order for this correlation to hold.
307: \section{Monte Carlo Simulations}
308: The goal of our Monte Carlo simulations is to generate the
309: luminosities and flux densities of a hypothetical sample that,in principle,
310: mimics the sample of gamma-ray detected blazars discussed above. To do
311: that we need to define a radio luminosity function from which we will
312: randomly draw to create the sample. We will also need a distribution of Lorentz factors
313: which will be used in determining Doppler beaming factors needed to
314: transform between observer frame luminosities and intrinsic luminosities.
315: In addition, we will need a model that determines what the gamma-ray
316: luminosity will be for a given radio luminosity. We will limit
317: ourselves to the models discussed above. Lastly, we will need to adopt
318: a cosmological model in order to determine redshifts corresponding to randomly
319: chosen enclosed volumes of space. We can then determine both radio and gamma-ray flux densities using these redshifts and the adopted cosmological model.
320: Following \citet{lis97} we have used an
321: assumed radio luminosity function of the form:
322:
323: $$ \rho(L) \propto L^{-g}\eqno{(4)}$$
324:
325: valid between $L_1$ and $L_2$. We have adopted a luminosity evolution function:
326:
327: $$L(z)=L(z=0)exp[T(z)/\tau] \eqno{(5)} $$
328:
329: Where $T(z)$ is the look-back time and $\tau$ is a constant in the approximate range of 0.1-0.5 that determines the degree of evolution.
330:
331: We have also assumed
332: a distribution of Lorentz factors for which:
333:
334: $$N(\Gamma) \propto \Gamma^{-s} \eqno{(6)}$$
335:
336: valid between $\Gamma_1$ and $\Gamma_2$ and a linear model connecting
337: gamma-ray and radio luminosities, i.e., the inverse Compton models discussed above,
338: to generate a simulated sample with Monte Carlo techniques and then use the corresponding distributions
339: (luminosity, redshift,flux density, etc.) and correlation diagrams to assess
340: the validity of the model. We express these
341: models using the following equations :
342:
343: $$ L_{\gamma,o,ssc}=K_{ssc}L_{r,i}\delta^{3+\alpha} \eqno{(7)}$$
344: $$ L_{\gamma,o,ecs}=K_{ecs}f(\beta,\theta)L_{r,i}{\delta^{4+2\alpha}}\eqno{(8)}$$
345:
346: Here, $L_{\gamma,o,ssc}$ and $L_{\gamma,o,ecs}$ refer to the predicted
347: gamma-ray luminosities in the observer's frame, using the SSC and ECS
348: models respectively. The K constants depend on the parameters
349: discussed above, $L_{r,i}$ refers to the intrinsic radio luminosity,
350: and the remaining factors are the appropriate functions of Doppler
351: factors for SSC or ECS assuming that the radiation is from discrete blobs \citep{lis99a}. As mentioned above, the powers in each case are reduced by one for the case of a continuous jet.
352:
353: \citet{lis97} rigorously determines radio luminosity function parameters based on simulations of the
354: Caltech-Jodrell Bank sample (CJ-F, as labeled by \citet{lis97}), taking into account distributions of
355: bulk Lorentz factors, redshifts, and luminosities. \citet{car07} have
356: recently arrived at comparable results for the similar MOJAVE sample
357: (though the luminosity and redshift ranges differ somewhat). We
358: begin by adopting the parameters that \citet{lis97} derived for the
359: CJ-F sample. We expect the CJ-F parent population to be very similar to that of our sample, because it is selected for flat spectrum, compact structure and 5 GHz total flux density $>$ 0.35 Jy. Additionally we take into account beaming effects appropriate for the SSC or ECS
360: model. Then, assuming both SSC and ECS in
361: turn, the gamma-ray luminosity and flux density is determined for each source.
362: In generating the gamma-ray luminosities from the intrinsic radio luminosities
363: we assume a power-law distribution of the K values discussed
364: above in Equations 7 and 8. This is an arbitrary choice of distribution, but the precise values of power-law
365: slope and cutoffs do not significantly effect the results.
366: We then generate 122 sources with gamma-ray flux that could have been
367: detected by EGRET. We adjust the gamma-ray flux density limit
368: to be slightly lower than that of EGRET (about 85 \%) to allow for a
369: proportion of upper limits that is similar to that we see in our observed sample.
370: In looking for agreement between the simulated
371: population and the observed one, we use the two distribution
372: Kolmogorov-Smirnov test on the distributions of luminosity, redshift
373: and flux density. The $D_{KS}$ statistic is then used to measure the
374: discrepancy between the two distributions. $D_{KS}$ is defined as \citep{pre86}:
375:
376: $$ D_{KS} \equiv { {\rm max} \vert S_{N1}(x)-S_{N2}(x)\vert } \eqno{(9)} $$
377:
378: That is, $D_{KS}$ is the maximum difference (absolute value) between the two distribution functions, $S_{N1}(x)$ and $S_{N2}(x)$. The distribution function is defined to be the probability that the particular value of the parameter in question is less than x. A value of $D_{KS}$ near zero indicates
379: agreement between the distribution functions, whereas a value near 1 indicates a large discrepancy. The
380: associated probability is also determined following \citet{pre86}. In this case, a probability approaching 1 indicates a high
381: probability for agreement between the distributions. For distributions
382: that include upper limits we use the corresponding tests using the
383: ASURV program (this is only an issue for the gamma ray flux densities
384: and luminosities).Among these tests are the Gehan generalized Wilcoxon
385: test and the Peto \& Peto Generalized Wilcoxon test. The results of
386: these tests are discussed below.
387: We have applied these methods using
388: the parameters of \cite{lis97}, summarized in Table 4.
389: In Table 4, Column (2) gives $L_1$, Column(3) gives $L_2$ and they refer to the lower and upper
390: luminosity limits of the intrinsic luminosity function. Column (4) gives
391: g and refers to
392: the luminosity function power law slope. Column(5) gives $\Gamma_1$,
393: Column(6) gives $\Gamma_2$ and
394: Column(7)gives s. These labels refer to the lower bulk Lorentz factor
395: distribution limit, upper bulk Lorentz factor distribution limit and the slope of the power law,
396: respectively. Column (8) gives $\tau$ and in this case refers to the evolution parameter, assuming an
397: exponential evolution model (see Equation 5). Column(9) gives $z_{min}$ and is the
398: lower limit to the redshift distribution and Column(10) gives $z_{max}$ and is the upper limit. Column (11) gives $h$ as discussed in {\S 2}, Column (12) gives the minimum value of $K$, Column (13) the maximum value of $K$ and Column (14) gives the slope of the power-law distribution of $K$, where $K$ refers to either $K_{ssc}$ or $K_{ecs}$ (Equations 7 and 8), whichever is appropriate for the simulation indicated. Applying the values assumed by \citet{lis97} does not
399: immediately lead to good agreement between the observed and modeled
400: distributions (see Table 5). For models marked "Lister SSC" we have adopted
401: an SSC model to generate simulated gamma ray luminosities and for "Lister ECS" we have adopted an ECS model to generate simulated gamma ray luminosities.
402: Next we have explored parameter space to determine where
403: better models might lie. Since the number of parameters and the
404: possible range of parameters is large, there are likely to be
405: multiple models that would have similar agreement as seen by a K-S test.
406: The results that give the lowest value of $D_{KS}$ (Equation 9) with high significance are summarized
407: in Table 5 for the SSC and ECS models. Column(2) gives the $D_{KS}$ value
408: for the radio luminosity function and Column(3) gives the associated
409: probability that the distributions are drawn from the same parent
410: population. Column(4) and Column(5) give the same results for the
411: gamma ray luminosity distributions, whereas Columns (6) and (7) do so
412: for redshift, and Columns(8)-(11) do so for the radio and gamma-ray
413: flux density distributions, respectively. These results can also be
414: visually inspected in Figures 2-11. There is only weak agreement
415: between the observed and predicted distributions for either SSC or ECS.
416: For the special case of the gamma-ray flux and luminosity
417: distributions (which have limits as well as measured values), we have
418: used the tests discussed above, and generally the associated
419: probabilities (for agreement of the two distributions) for these
420: tests fall in the range of 0.0002-0.0015 for SSC and 0.0001-0.0010 for ECS, thus also only very weakly confirming
421: the SSC or ECS hypothesis. The parent population in these simulations was approximately 30
422: million for SSC and 300 million for ECS. Both models require a range of
423: Lorentz factors of about 1-30. This is the approximate range inferred by VLBI observations \citep{coh07}. However, some authors have made the argument that these bulk Lorentz factors may not pertain to the same portion of the jet from which the gamma ray
424: emission originates \citep{geo03}.
425:
426: A closer look at Figures 2-11 show some features that are not as evident from just comparing results of Kolgomorov-Smirnov tests. A comparison of the radio luminosity distributions (Figures 2a,2b,\& 7) shows that assuming an SSC or ECS model lead to similar results. Both the SSC and ECS distributions have peak values near $10^{27} {\rm Watts \, Hz^{-1}}$, with approximately 24 \% of the sample having values near the peak for SSC and 21 \% for ECS. The observed distribution of radio luminosity peaks near $10^{28} {\rm Watts \,Hz^{-1}}$ with approximately 33 \% of the sample in the bin nearest to the peak. However, in conjunction with the redshift and flux density distributions we do see some important differences between the predictions of the SSC and ECS model, and can pin point particular observations that will be necessary in the GLAST era for differentiating these models using the methods we have discussed above.
427: For instance, when looking at the simulated distributions of radio flux density (Fig. 4a, Fig. 9 ), the simulated distribution assuming the SSC model is clearly peaked close to about 0.3 Jy, whereas the same distribution for ECS is peaked close to 1.5 Jy. The observed distribution is peaked closer to 0.3 Jy and is more similar in overall shape to the predicted SSC distribution (Fig. 4b). Ongoing radio monitoring of GLAST detected sources (or at least updated to these older survey flux densities) will be critical for determining the true observed distribution to be used for comparison to models. In addition, we can see that the redshift distribution (Fig. 6a) with SSC assumed is nearly identical to the corresponding distribution with ECS assumed (Fig. 11). Though the observed redshift distribution (Fig. 6b) is also peaked near z=1, this is at least in part due to the assumed redshifts of z=1 for 13\% of the sample. An additional challenge in th GLAST era will be obtaining re!
428: dshifts for entire large samples (possibly thousands of new quasars that have not yet been optically identified or observed spectroscopically). It is a bit harder to discern significant differences between the ECS and SSC predictions for either gamma-ray luminosity or gamma-ray flux density. However, a significant increase in sensitivity for GLAST should clarify the nature of the low-end of each of these diagrams. For example, we now can only see
429: a very large drop-off in all of the flux densities near $10^{-11}$ Jy. This is caused by the flux detection limit of EGRET. In addition, the dramatic increase in number of gamma-ray detected sources to thousands \citep{pad07} will clarify the shape of of these distributions and increase the conclusiveness of comparisons to theoretical predictions.
430:
431: We have also used the Monte Carlo results to generate simulated
432: luminosity-luminosity correlation plots between $log L_{\gamma}$ and $log L_r$ by assuming, in turn, the SSC and
433: ECS models. The SSC simulations generate the relationship $L_{\gamma} \propto{L_r}^{0.74 \pm 0.02}$, in agreement with observations. However, assuming the ECS model generates results with $L_{\gamma} \propto
434: {L_r}^{0.95 \pm 0.02}$ (the results are plotted in Figures 12a and 12b). This clearly shows that though the assumed underlying physical models in both the SSC and ECS case are linear between $L_{\gamma}$ and $L_r$, selection effects can lead to a simulated sample that shows a non-linear relationship between the luminosities. This effect is more pronounced for the SSC case.
435: This is potentially an effect caused by Doppler beaming factors raised to different exponents (dependent on spectral shape). For instance, in the SSC case, if the exponents for the beaming factors are identical for both the radio and gamma ray part of the spectrum, then we would expect no change in going from the intrinsic luminosity correlations and the observed ones. However, in our model we do account for the possibility that the gamma-ray spectrum is considerably steeper than the radio spectrum. This leads to somewhat different exponents for the Doppler beaming factors that are going into an overall beaming correction that is
436: applied to the intrinsic radio and gamma ray luminosities. There are two additional effects which we have not modeled, but may contribute to the observed luminosity correlation. In particular, the radio source and gamma-ray source may not be spatially coincident, and may thus have different bulk Lorentz factors, as discussed above. Also, the bulk Lorentz factors may be dependent on radio luminosity \citep{lis97}.
437:
438: Though similar effects are also at work with ECS process, it is less pronounced in our case due to a selection effect. That is, because the range of $K_{ecs}$ is very narrow, and also has relatively low values, only the sources with the highest values of Lorentz factors are detected in gamma rays. Thus, relative to the SSC case, the Doppler factors are closer to being constant for the sample. Overall, this leads to a relationship between the gamma-ray and radio luminosities that is closer to being linear.
439:
440: %Though the value of the exponent (slope in a logarithmic
441: %plot) is somewhat larger than that for the observed sample, there is
442: %only aprroximately a 1 $\sigma$ difference between the 0.04 uncertainties
443: %of each exponent determination. Interestingly, both simulated
444: %exponents deviate from 1.0, the underlying intrinsic correlation
445: %exponent, though once the uncertainties are taken into
446: %account, these differences are also $<1 \sigma$
447: \section{Additional Correlations}
448: In addition to the luminosity correlation that we have discussed at
449: length we have also observed other significant correlations for the sample.
450: The origin of the broad band spectral index correlation, $\alpha_{og}$
451: vs. $\alpha_{ro}$ (Table 2), becomes clear if we consider the conclusions of
452: \citet{fos98}. They show that gamma-ray dominance (the ratio of gamma-ray luminosity to luminosity at the
453: model dependent synchrotron peak of the SED) is positively correlated with the maximum
454: frequency of the lower energy peak of the SED. In a simple
455: homogeneous source, this lower energy peak corresponds to the
456: upper synchrotron cutoff frequency discussed above. \citet{ghi98} show that this correlation is a proxy for
457: the underlying physical relationship between the maximum relativistic electron energy (on which the low frequency peak strongly
458: depends) and the ratio of the Compton and synchrotron luminosities. For a model in which both an internal (SSC)
459: and external (ECS) radiation field are present, an increase in the energy density of the external field leads to
460: an increase in Compton cooling and thus a decrease in the maximum electron energy and an increase in the Compton gamma-ray
461: luminosity \citep{ghi98}.
462: \citet{fos98} further show that observed optical flux density can be substituted for the (model dependent) synchrotron peak flux
463: density without loss of the Compton dominance/spectral peak frequency correlation. We utilize this useful substitution
464: below.
465: \citet{fos98} also show that both $\alpha_{rx}$ and
466: $\alpha_{ro}$ are separately but similarly correlated with this same spectral maximum frequency, so that the
467: latter may be substituted for the former in the absence of X-ray data.
468: Though we have limited spectral data for many sources, and thus
469: can't confirm this related correlation for our own sample, the same
470: reasoning should apply to our sample, and thus we make this substitution
471: as well. Thus the correlation of $\alpha_{og}$ and $\alpha_{ro}$ is a
472: possible indicator of the relationship between dominance of gamma-rays
473: produced by the ECS process and maximum electron energy.
474: However, it is also possible that the broad band spectral index correlation
475: is not independent of the radio/gamma-ray luminosity correlation
476: discussed above. This is especially true if we take into account the
477: common dependence of both spectral indices on optical luminosity
478: (Table 3). Recent surveys have also suggested that the aforementioned correlations discussed in \citet{fos98} and \citet{ghi98} may be due in part to selection effects \citep{ant05,nie06}. For instance, the low luminosity objects of \citet{fos98} were mainly X-ray selected, favoring selection of objects with SEDs peaked at higher frequencies, whereas the high luminosity objects were mainly radio selected, thus favoring the selection of objects with SEDs peaked at lower frequencies \citep{ant05}.
479:
480: The correlation between radio luminosity and gamma-ray dominance
481: (which in our analysis would be characterized by $\alpha_{og}$) ,
482: originally seen by \citet{fos98} is also seen here, after more than doubling the sample size. Here, because of
483: the way the broad band indices are defined, a lower value of $\alpha_{og}$ corresponds to greater gamma-ray dominance.
484: This correlation is further evidence for a blazar sequence defined by relative importance of the ECS model
485: at higher luminosities.
486:
487: We have also examined a possible correlation between all parameters
488: individually and the gamma-ray variability \citep{nol03} and find
489: only a weak anti-correlation between gamma-ray spectral index, $\alpha_{\gamma}$, and variability (as parameterized by \citet{nol03}; see Table 1). This may be related to a spectral hysteresis effect observed by \citet{nan07} for a sample of 26 particularly bright and well observed blazars. These authors find that during gamma ray flares the spectral index first tends to flatten with increasing flux, and then returns to a steeper index as the flare ends. Thus, it is plausible that if a blazar is more highly variable there will be an increasing chance of observing it while its spectral index is flatter than for other blazars that are not gamma ray variables.
490: \section{Conclusions and Implications for GLAST}
491: After taking into account statistical tests and Monte Carlo analysis,
492: we find the following:
493: \begin{enumerate}
494: \item For this sample of 122 gamma-ray blazars there is a strong
495: correlation between radio and gamma ray
496: luminosity which persists even after the effects of redshift and
497: limits are taken into account. The correlation is of the form
498: $L_{\gamma} \propto {L_r}^{0.77}$. This correlation remains with
499: similar regression coefficients even when only the strongest
500: 76 candidates are included in the sample.
501: \item There is a correlation between $\alpha_{og}$ and $\alpha_{ro}$
502: as well as a correlation between $L_r$ and $\alpha_{og}$. Each
503: correlation is consistent
504: with the increasing dominance of the SED by gamma ray luminosity as
505: the maximum relativistic electron energy decreases and the increasing
506: importance of the ECS process at high radio luminosities and low
507: maximum electron energy.
508: \item A detailed simulation of source statistics using Monte Carlo
509: techniques shows that the relationship $L_{\gamma} \propto {L_r}^{0.77}$ can only be reproduced assuming the SSC model. Though the assumed intrinsic physical model in the source frame is of the form $L_{\gamma} \propto L_r$, selection effects lead to a simulated sample with the relationship described above. This effect is much less significant when ECS is assumed as the underlying model. However, upon also comparing the observed and simulated distributions of the luminosities, flux densities and redshifts, both the SSC and ECS models are only weakly consistent with the data, assuming linear dependence of intrinsic
510: gamma ray luminosity
511: on intrinsic radio luminosity. Taken together with the previous
512: results, this would suggest that if either SSC or ECS is indeed
513: responsible for the gamma ray emission, a more complex model is
514: likely needed. In addition, effects of evolution and assumed
515: cosmologies can be explored in more detail.
516: \end{enumerate}
517: In the GLAST era, our findings can be clarified in the following ways. Acquiring more gamma-ray data, including, particularly, detections of new dim sources near the GLAST detection limit will clarify whether the correlations we find are caused by truncation effects due to the flux limits of our instruments, or whether this is due, at least in part, to physical causes, at least down to the new limit established by GLAST.
518:
519: In order to predict what we might see with GLAST, we can extend our observed radio/gamma ray luminosity correlation down to lower luminosities than what are covered in Figure 1. A radio luminosity of $1.25 \times 10^{24} {\rm Watts\,Hz^{-1}}$ would lead to a gamma ray luminosity of approximately $4.98 \times 10^{13} {\rm Watts \, Hz^{-1}}$. We can then covert these luminosities to flux densities for a range of potential redshifts. For z=0.1-1.1 the radio flux density would be in the range of 0.8-60 mJy and the gamma-ray flux density would be in the range,
520: $3.3 \times 10^{-14}$-$2.4 \times 10^{-12}$ Jy. Greater values of redshift would lead to even lower flux density values. Most of the range of radio flux densities ($>$ few mJy) would be detectable by the Green Bank Telescope (GBT), with the lower limit detectable with the Very Large Array (VLA)\citep{con08}. Assuming that the sensitivity of GLAST is approximately $2.8 \times 10^{-13}$ Jy at 100 MeV, these sources would only be be detected if $z< 0.2$. A radio luminosity of $1.58 \times 10^{23} {\rm Watts \, Hz^{-1}}$ corresponds to a gamma-ray luminosity of $1 x10^{13} {\rm Watts \, Hz^{-1}}$ . The radio flux density would fall in the range of 0.1-8 mJy for z=0.1-1.1. These sources could potentially be radio detectable with the VLA under optimal conditions. At the higher flux density limit it would also be possible to detect with GBT, especially at higher frequencies (i.e., 40-50 GHz)\citep{min08}. However, if the regression slope were actually closer to 1, implying a d!
521: irect proportionality between radio and gamma ray luminosities, then radio sources in the same luminosity, redshift and flux ranges discussed above would not be detected in gamma rays at all. In short, for the observed luminosity relationship for EGRET blazars to be appreciably extended down to lower luminosities with GLAST detections, then a large percentage of GLAST sources near the detection limit would have radio flux densities about 10 mJy or greater (and $z< 1$). Several authors have stated that the predicted number of blazars to be detected by GLAST rougly matches with the sky density of flat spectrum radio sources down to 50 mJy at 5 GHz \citep{pad07} and 65 mJy at 8.4 GHz \citep{hea07}. However, our results show that a large number of GLAST sources would be detected at even lower radio flux densities if the luminosity relationship we observe for EGRET sources also holds up for GLAST sources. It is very likely that many new radio observations would be required in a!
522: ny of the cases mentioned above, but especially for cases in w!
523: hich the
524: putative radio source may not even be in any previous catalog.
525:
526: >From a theoretical perspective, confirmation of the previously determined correlation between gamma-ray and radio luminosities will lead to confirmation of SSC, though the precise agreement would have to be re-analyzed with these new data.
527:
528: \acknowledgements
529: The author thanks M. Lister for many useful discussions and the
530: anonymous referee for abundant constructive comments.
531: S. Bloom would like to acknowledge the generosity of the National Radio
532: Astronomy Observatory (NRAO) during his stay as a Visiting Scientist at NRAO
533: Headquarters in Charlottesville, VA. S. Bloom would also like to acknowledge several Summer Faculty Fellowship grants from Hampden-Sydney College.
534:
535: This work has also made extensive use of the NASA Extragalactic Database
536: (NED) and NASA's Astrophysics Data System Bibliographic Services (ADS).
537:
538: \begin{thebibliography}{}
539:
540: \bibitem[Akritas \& Siebert(1996)]{akr96} Akritas, M. G. \& Siebert,
541: J. 1996,\mnras, 278, 919
542:
543: \bibitem[Anton \& Browne(2005)]{ant05} Anton, S. \& Browne, I. W. A 2005, \mnras,356, 225
544:
545: \bibitem[Bessel(1979)]{bes79} Bessel,M. 1979, \pasp, 91, 589
546:
547: \bibitem[Bloom \& Marscher(1996)]{blo96} Bloom, S. D. \& Marscher,
548: A. P., \apj, 461, 657
549:
550: %\bibitem[Bloom {\it et al.}(1997)]{blo97} Bloom {\it et al.} 1997, \apj, 488, L23
551:
552: %\bibitem[Bloom {\it et al.}(2000)]{blo00} Bloom {\it et al.} 2000, \apj, 529, 675
553:
554: \bibitem[Bloom {\it et al.}(2004)]{blo04} Bloom {\it et al.} 2004,
555: \aj, 128, 56
556:
557: \bibitem[Cara \& Lister(2007)]{car07} Cara, M. \& Lister, M. L. 2007, \apj, in press
558:
559: \bibitem[Caroll \& Press(1992)]{car92} Caroll, S. M. \& Press,
560: W. H. 1992, \araa, 30, 499
561:
562: \bibitem[Cohen {\it et al.}(2007)]{coh07} Cohen, M. H., Lister, M. L., Homan, D. C., Kadler, M., Kellermann, K. I., Kovalev, Y. Y., Vermeulen, R. C.
563: 2007, \apj, 659, 232
564:
565: \bibitem[Condon(2008)]{con08} Condon, J. J. 2008, private communication
566:
567: \bibitem[Dermer, Schlickeiser,\& Mastichiadis(1992)]{der92} Dermer,
568: C. D., Schlickeiser, R., Mastichiadis, A. 1992, \aap,256, L27
569:
570: \bibitem[Dermer \& Schlickeiser(1993)]{der93} Dermer, C. D. \&
571: Schlickeiser, R. 1993, \apj, 416,458
572:
573: \bibitem[Dermer(1995)]{der95} Dermer, C. D. 1995, \apj, 446, L63
574:
575: \bibitem[Fan {\it et al.}(1998)]{fan98} Fan, J. H., Adam, G., Xie,
576: G. Z., Cao, S. L., Lin, R. G., Copin, Y. 1998, \aap, 338, 27
577:
578: \bibitem[Feigelson \& Nelson(1985)]{fei85} Feigelson, E. D. \& Nelson,
579: P. I. 1985, \apj, 293, 192
580:
581: \bibitem[Fossati {\it et al.}(1998)]{fos98} Fossatti, G., Maraschi, L., Celotti, A., Comastri, A., Ghisellini, G. 1998, \mnras, 299, 433
582:
583: \bibitem[Fouque {\it et al.}(1992)]{foq92} Fouque, P., Durand, N.,
584: Bottinelli, L., Gougenheim, L., Patureli, G. 1992, {\it Catalogue of
585: Optical Radial Velocities}
586:
587: \bibitem[Ghisellini, Maraschi, \& Treves(1985)]{ghi85} Ghisellini, G.,
588: Maraschi, L., Treves, A. 1985,\aap, 146, 204
589:
590: \bibitem[Ghisellini \& Madau(1996)]{ghi96} Ghisellini, G. \& Madau, P.
591: 1996,\mnras, 280, 67
592:
593: \bibitem[Ghisellini {\it et al.}(1998)]{ghi98} Ghisellini, G.,
594: Celotti, A., Fossati, G., Maraschi, L., Comastri, A. 1998, \mnras,
595: 301, 451
596:
597: \bibitem[Georganopoulos \& Kazanas(2003)]{geo03} Georganopoulos, M. \&
598: Kazanas, D. 2003, \apj, 594, L27
599:
600: \bibitem[Halpern, Eracleous \& Mattox(2003)]{hal03} Halpern, J. P.,
601: Eracleous, M., Mattox, J. R. 2003, \aj, 125, 572
602:
603: \bibitem[Hartman {\it et al.}(1999)]{har99} Hartman, R. C. {\it et al.}
604: 1999, \apjs, 123, 79
605:
606: \bibitem[Healey {\it et al.}(2007)]{hea07} Healey, S. E. {\it et al.} 2007,
607: \apjs, 171, 611
608:
609: \bibitem[Hewitt \& Burbidge(1987)]{hew87} Hewitt, A. \& Burbidge,
610: G. 1987, \apjs, 63, 1
611:
612: \bibitem[Impey(1996)]{imp96} Impey, C. 1996, \aj, 112, 2667
613:
614: \bibitem[Isobe, Feigelson \& Nelson (1986)]{iso86} Isobe, T.,
615: Feigelson, E. D., Nelson, P. I. 1986, \apj, 306, 490
616:
617: \bibitem[Kendall \& Gibbons(1990)]{ken90} Kendall, M. \& Gibbons,
618: J. D. 1990, {\it Rank Correlation Methods} (New York: Oxford University Press)
619:
620: \bibitem[Landt {\it et al.}(2001)]{lan01} Landt, H., Padovani, P.,
621: Perlman, E. S., Giommi, P., Bignall, H., Tziomis, A. 2001, \mnras,
622: 323, 757
623:
624: \bibitem[Lauberts \& Valentijn(1989)]{lau89} Lauberts, A. \& Valentijn,
625: G. A. 1989, {\it A Surface Photometry Catalog of ESO-Uppsala
626: Galaxies}, European Space Observatory, Garching
627:
628: \bibitem[Lister \& Marscher(1997)]{lis97} Lister, M. \& Marscher,
629: A. P. 1997, \apj,476, 572
630:
631: \bibitem[Lister(1999)]{lis99a} Lister, M. 1999, Ph. D. dissertation, Boston University
632:
633: %\bibitem[Lister \& Marscher(1999)]{lis99b} Lister, M. \& Marscher, A. P. , Astroparticle Physics
634:
635: \bibitem[Liang \& Liu(2003)]{lia03} Liang, E. W. \& Liu, H. Y. 2003, \mnras, 340, 632
636:
637: \bibitem[Marscher(1987)]{mar87} Marscher, A. P. 1987, in {Superluminal Radio Sources}, eds. J. A. Zensus \& T. J. Pearson (Cambridge: Cambridge University Press), 280
638:
639: \bibitem[Mattox {\it et al.}(1997)]{mat97} Mattox, J. R. {\it et al.}
640: 1997, \apj, 481, 95
641:
642: \bibitem[Mattox, Hartman \& Reimer(2001)]{mat01} Mattox, J. R., Hartman, R. C., \& Reimer, O. 2001, \apj, 135, 155
643:
644: \bibitem[McMahon {\it et al.}(2002)]{mcm02} McMahon, R. G. White,
645: R. L., Helfand, D. J., Becker, R. H. 2002, \apjs, 143, 1
646:
647: \bibitem[Minter(2008)]{min08} Minter, T. 2008, private communication
648:
649: \bibitem[Monet {\it et al.}(2003)]{mon03} Monet, D. G. {\it et al.} 2003,
650: \aj, 125, 984
651: J., Mirabal, N., Gotthelf, E. V. 2002, \apj, 574, 694
652:
653: \bibitem[Mucke {\it et al.}(1997)]{muc97} Mucke, A. {\it et al.} 1997, \aap, 320, 33
654:
655: \bibitem[Nandikotkur {\it et al.}(2007)]{nan07} Nandikotkur, G. {\it et al.} 2007, \apj, 657, 706
656:
657: \bibitem[Nieppola, Tornikoski, \& Valtaoja(2006)]{nie06} Nieppola, E., Tornikoski, M., \& Valtaoja, E. 2006, \aap, 445, 441
658:
659: \bibitem[Nolan {\it et al.}(2003)]{nol03} Nolan, P. L., Tompkins, W. F., Grenier, I. A., Michelson, P. F. 2003, \apj, 597, 615
660:
661: \bibitem[Padovani(1992)]{pad92} Padovani, P. 1992, \aap, 256, 399
662:
663: \bibitem[Padovani(2007)]{pad07} Padovani, P. 2007, {\sl The First GLAST Symposium}, eds. S. Ritz, P. Michelson, C. Meegan, p. 19
664:
665: \bibitem[Perlmutter {\it et al.}(1997)]{per97} Perlmutter, S. {\it et al.} 1997, \apj, 483, 565
666:
667: \bibitem[Pen(1999)]{pen99} Pen, U. L. 1999, \apjs,120, 49
668:
669: \bibitem[Press {\it et al.}(1986)]{pre86} Press, W. H., Flannery, B. P., Teukolsky, S. A., \& Vetterling, W. T. (Cambridge: Cambridge University Press), 472
670:
671: \bibitem[Salamon \& Stecker (1996)]{sal96} Salamon, M. H. \& Stecker,
672: F. W. 1996, \apjl, 430, 21
673:
674: \bibitem[Sguera {\it et al.}(2004)]{sgu04} Sguera, V., Malizia, A.,
675: Bassani, L., Stephen, J. B., Di Cocco 2004, \aap, 414, 839
676:
677: \bibitem[Sguera {\it et al.}(2006)]{sgu06} Sguera, V., Bassani, L.,
678: Malizia, A., Dean, A. J., Landi, R., Stephen, J. B. 2006, \aap, 430, 107
679:
680: \bibitem[Sowards-Emmerd, Romani, \& Michelson(2003)]{sow03} Sowards-Emmerd, D., Romani, R. W., Michelson, P. F. 2003, \apj, 590, 109
681:
682: \bibitem[Sowards-Emmerd, Romani, \& Michelson(2004)]{sow04} Sowards-Emmerd, D. , Romani, R. W., Michelson, P. F. 2004, \apj, 609, 564
683:
684: \bibitem[Stecker, Salamon, \& Malkan(1993)]{ste93} Stecker, F. W.,
685: Salamon, M. H., Malkan, M. A. 1993, \apjl, 410, 71
686:
687: \bibitem[Stocke {\it et al.}(1991)]{sto91} Stocke, J. T. {et al.}
688: 1991, \apjs, 76, 813
689:
690: \bibitem[Thompson {\it et al.}(1996)]{tho96} Thompson, D. J. {\it et al.} 1997, \apjs, 107, 227
691:
692: \bibitem[Torres, Pessah \& Romero(2001)]{tor01} Torres, D. F., Pessah, M. E., \& Romero, G. E. 2001, {\it Astron. Nachr.}, 4, 223
693:
694: \bibitem[Veron-Cetty \& Veron(2001)]{ver01} Veron-Cetty, M. P. \&
695: Veron, P. 2001, \aap, 374, 92
696:
697: \bibitem[Wisotski {\it et al.}(2000)]{wis00} Wisotzki, L. {\it et
698: al.} 2000, \aap,358,77
699:
700: \bibitem[Wright \& Otrupcek(1990)]{wri90} Wright, A. \& Otrupcek, R 1990,
701: {\it Parkes Radio Catalogue}, Australian Telescope National Facility
702:
703: \bibitem[Zhang, Cheng, \& Fan(2001)]{zha01} Zhang, L., Cheng, K. S.,
704: Fan, H. 2001, \pasj, 53, 207
705:
706: \bibitem[Zhou {\it et al.}(1997)]{zho97} Zhou, Y. Y., Lu, Y. J., Wang,
707: T. G., Yu, K. N., Young, E. C. M. 1997, \apjl,484, 47
708: \end{thebibliography}{}
709:
710: \clearpage
711:
712: %\centerline{\bf FIGURE CAPTIONS}
713: \begin{figure}
714: \figurenum{1}
715: \plotone{f1.eps}
716: \caption{The observed radio/gamma ray luminosity correlation. Arrows
717: indicate upper limits and the dashed line represents the
718: regression fit.}
719: \end{figure}
720: \begin{figure}
721: \figurenum{2}
722: \plottwo{f2a.eps}{f2b.eps}
723: \caption{Simulated (left) and Observed(right) Radio Luminosity
724: Distributions (SSC model assumed for simulations)}
725: \end{figure}
726: \begin{figure}
727: \figurenum{3}
728: \plottwo{f3a.eps}{f3b.eps}
729: \caption{Simulated (left) and Observed (right) Gamma Ray Luminosity
730: Distributions (SSC model assumed for simulations)}
731: \end{figure}
732: \begin{figure}
733: \figurenum{4}
734: \plottwo{f4a.eps}{f4b.eps}
735: \caption{Simulated (left) and Observed (right) Radio Flux Density
736: Distributions (SSC model is assumed for simulations)}
737: \end{figure}
738: \begin{figure}
739: \figurenum{5}
740: \plottwo{f5a.eps}{f5b.eps}
741: \caption{Simulated (left) and Observed (right) Gamma Ray Flux Density
742: Distributions (SSC model is assumed for the simulations)}
743: \end{figure}
744: \begin{figure}
745: \figurenum{6}
746: \plottwo{f6a.eps}{f6b.eps}
747: \caption{Simulated (left) and Observed (right) Redshift Distributions
748: (SSC model is assumed for the simulations)}
749: \end{figure}
750: \begin{figure}
751: \figurenum{7}
752: \plotone{f7.eps}
753: \caption{Simulated Radio Luminosity Distribution (ECS model is assumed)}
754: \end{figure}
755: \begin{figure}
756: \figurenum{8}
757: \plotone{f8.eps}
758: \caption{Simulated Gamma Ray Luminosity Distribution (ECS model is assumed)}
759: \end{figure}
760: \begin{figure}
761: \figurenum{9}
762: \plotone{f9.eps}
763: \caption{Simulated Radio Flux Density Distribution (ECS model is assumed)}
764: \end{figure}
765: \begin{figure}
766: \figurenum{10}
767: \plotone{f10.eps}
768: \caption{Simulated Gamma Ray Flux Density Distribution (ECS model is assumed)}
769: \end{figure}
770: \begin{figure}
771: \figurenum{11}
772: \plotone{f11.eps}
773: \caption{Simulated Redshift Distribution (ECS model is assumed)}
774: \end{figure}
775: \begin{figure}
776: \figurenum{12}
777: \plottwo{f12a.eps}{f12b.eps}
778: \caption{Simulated Radio/Gamma-Ray Luminosity Correlation Diagrams for
779: SSC (left) and ECS (right). The arrows indicate upper limits and
780: the dashed line indicates the regression fit.}
781: \end{figure}
782:
783: \clearpage
784:
785: \begin{deluxetable}{lllccccccccc}
786: \tabletypesize{\tiny}
787: %\rotate
788: \tablecolumns{12}
789: \tablewidth{0pt}
790: \tablenum{1}
791: \tablecaption{Multiwaveband Data for EGRET Blazars}
792: \tablehead{\colhead {EGRET Source}& \colhead {Radio Source}&\colhead
793: {z}& \colhead {log $L_r$}&\colhead {log $L_o$} &\colhead{log
794: $L_g$}&\colhead{$\alpha_r$}&\colhead{$\alpha_g$
795: }&\colhead{$\alpha_{ro}$}&\colhead {$\alpha_{og}$}&\colhead
796: {$\delta_{var}$} &\colhead {Ref}\\
797: \colhead {} &\colhead {} &\colhead {} &\colhead {($Watts \, Hz^{-1}$)} &\colhead {($Watts \,Hz^{-1}$)} &\colhead
798: {($Watts \,Hz^{-1}$)}&\colhead {}&\colhead {}&\colhead{} &\colhead{} &\colhead{} &\colhead{}\\
799: \colhead{(1)} &\colhead{(2)} &\colhead{(3)} &\colhead{(4)}
800: &\colhead{(5)} &\colhead{(6)}&\colhead{(7)}& \colhead {(8)} &
801: \colhead{(9)}& \colhead {(10)} &\colhead {(11)}& \colhead {(12)}\\}
802: \startdata
803: 3EG J0038-0949& J0039-0942 &2.101&28.030&24.41&16.96&-0.12&1.7 &0.69&0.83&0&2,1,6,3\\
804: 3EG J0118+0248& J0113+0222 & 0.047& 24.26& 23.38 &13.15 & -0.1 & 1.63 & 0.18& 1.24 & 1.18&1,1,1,3\\
805: 3EG J0130-1758& J0132-1654& 1.022& 27.74 &23.80& 16.44& -0.08& 1.5&
806: 0.78& 0.86&0& 2,2,4,3\\
807: 3EG J0204+1458& J0204+1514& 0.405 &27.11& 21.29& 15.53& 0.09& 1.23&
808: 1.20&0.68&1.29& 7,1,5,3\\
809: 3EG J0210-5055& J0210-5101& 0.999 &28.25& 24.39& 17.57& -0.07& 0.99&
810: 0.78& 0.81& 0.31& 14,15,4,3\\
811: 3EG J0215+1123& J0213+1213&0.252& 25.31 & 21.57 & $<$ 15.16 &-0.1& 1.03& 0.77& 0.77& 1.27&1,1,1,3\\
812: 3EG J0222+4253& J0222+4302 & 0.444& 26.54 & 24.11& 16.04& 0.17& 1.01& 0.50& 0.97&0 &4,1,4,3\\
813: 3EG J0239+2815& J0237+2848&1.213& 28.48& 23.83& 16.65 &-0.1 &
814: 1.53&0.93&0.83 &0& 4,1,4,3 \\
815: 3EG J0237+1635& J0238+1636&0.94 & 28.51& 23.48& 17.05 & -0.5&
816: 0.85&0.98&0.77& 0.89& 4,1,4,3 \\
817: 3EG J0245+1758& J0242+1742&0.551& 26.38& 22.22& 15.66 &-0.1 &
818: 1.61&0.84&0.77& 1.14 & 1,1,1,3 \\
819: 3EG J0329+2149& J0325+2224&2.066& 28.39& 24.24& 16.82 & 0 & 1.61&
820: 0.81& 0.83& 0 & 1,1,1,3\\
821: 3EG J0340-0201& J0339-0146&0.852& 28.02& 23.61& 16.71 & -0.13&
822: 0.84&0.88&0.72 & 0& 4,1,4,3 \\
823: 3EG J0404+0700& J0407+0742& 1.133& 27.75& 24.29 & 16.42 & -0.3& 1.65&
824: 0.67& 0.91& 0.39& 1,1,1,3 \\
825: 3EG J0412-1853& J0416-1851& 1.536& 28.07& 24.27& $<$ 16.14& 0.23& 2.25& 0.77& 0.90& 0&4,1,4,3 \\
826: %3EG J0416+3650& J0418+3801 & 0.049& 25.61& 21.07& 13.88& 0&1.59&
827: %0.93& 0.88& 0.59& \\
828: 3EG J0422-0102& J0423-0120& 0.915& 28.28& 24.95& 16.51& -0.2& 1.44&
829: 0.65& 0.99& 0 & 4,1,4,3\\
830: 3EG J0423+1707&J0422+1741& 0.908&26.95&23.14&16.50&-0.5&1.43&0.73&0.77&0.42& 1,1,1,3 \\
831: 3EG J0433+2908&J0433+2905&\nodata & 27.34&23.43& 17.03& 0&0.9& 0.78&
832: 0.76& 0.40 &\nodata,1,3\\
833: 3EG J0442-0033&J0442-0017& 0.844&27.32& 23.27& 16.36& 0.39& 1.37&
834: 0.84&0.81 &1.59& 4,1,4,3\\
835: 3EG J0450+1104&J0449+1121&1.207&28.11& 23.44& 16.85&-0.2& 1.27& 1.06&
836: 0.69&1.13 & 5,1,5,3\\
837: 3EG J0456-2338&J0457-2324& 1.003& 27.98& 23.66& 15.84&0&2.14& 0.84&
838: 0.9&0.21& 5,1,5,3\\
839: 3EG J0458-4635&J0455-4615&0.858 &27.66& 24.02& 15.94&0.18&1.75&0.74&
840: 0.94&0.41 & 4,15,4,3\\
841: 3EG J0459+0544&J0502+0609&1.106 &27.48& 23.40& 16.32& 0.28&1.36& 0.82&
842: 0.83&0.74& 5,1,5,3\\
843: 3EG J0459+3352&J0503+3403&0.149 &25.16& 22.02 & 14.64&0.38&1.54&0.65&
844: 0.89&0.59& 1,1,1,3\\
845: 3EG J0500-0159&J0501-0159&2.2286& 29.53& 24.76& 17.20&-0.25&1.45&0.91&
846: 0.86&1.09& 4,2,4,3 \\
847: 3EGJ0510+5545&J0514+5602&2.19&28.02&23.50&17.66&0.18&1.19&0.91&0.67&0&1,1,1,3\\
848: 3EG J0512-6150&J0506-6109& 1.093&27.90& 24.52 & 16.36&0.21& 1.40&
849: 0.68&0.95&0&4,15,4,1 \\
850: 3EG J0530-3626&J0529-3555&\nodata&27.29&24.08&16.47&-0.42&1.63&0.64&1.04&0.58& \nodata,2,7,3\\
851: 3EGJ0530+1323&J0530+1331&2.07&29.31&24.03&18.04&-0.3&1.46&1.01&0.55&0.74&
852: 5,1,5,3\\
853: 3EG J0531-2940&J0539-2839&3.104&29.07 &24.52&17.27&0.06&1.47&0.89&0.81&0.85& 4,2,4,3\\
854: 3EG J0533+4751&J0533+4822&1.16&27.68&23.67&16.61&-0.1&1.55&0.75&0.84&
855: 0.0& 1,1,1,3\\
856: 3EGJ0540-4402&J0538-4405&0.896&28.56&25.20&17.08&-0.17&1.41&0.66&0.95&0.75&4,15,4,3\\
857: 3EGJ0542-0655&J0541-0541&0.839&27.49&22.95&$<$16.45&-0.06&1.0&0.91&0.77&1.38&5,2,5,3\\
858: 3EGJ0542+2610&J0540+2507&0.62&26.40&24.56&15.96&0.12&1.67&0.90&0.70&0.57&
859: 1,1,1,3\\
860: 3EG J0721+7120&J0721+1721&0.3&26.02&23.61&15.56&0.09&1.19&0.49&0.97&0&4,1,4,3\\
861: 3EG J0737+1721&J0738+1742&0.424&27.14&24.24&15.69&-0.1&1.6&0.58&1.02&0&4,1,4,3\\
862: 3EG J0743+5447&J0742+5444&0.723&26.36&24.84&16.32&0.36&1.03&0.55&0.88&1.21&5,1,5,3\\
863: %3EG J0808+4844&J0807+4849&1.43&28.45&24.96&17.33&0.12&1.15&0.69&0.89&0\\
864: 3EG J0808+5114&J0807+5117&1.14&27.57&23.95&16.24&-0.4&1.76&0.69&0.88&0&5,1,5,3\\
865: 3EG J0828+0508&J0831+0429&0.1736&25.77&22.65&14.88&0&1.47&0.64&0.93&0&4,1,4,3\\
866: 3EG J0829+2413&J0830+2410&0.939&27.48&24.18&16.73&0&1.42&0.66&0.87&\nodata&4,1,4,3\\
867: 3EG J0845+7049&J0841+7053&2.172&28.77&25.49&16.99&0.42&1.62&0.67&0.96&0.62&4,1,4,3\\
868: 3EG J0852-1216&J0850-1213&0.566&26.87&23.77&16.31&-0.37&0.58&0.61&0.90&1.21&5,2,5,3\\
869: 3EG J0853+1941&J0854+2006&0.306&26.80&24.23&15.41&-0.3&1.03&0.51&1.06&0&4,1,4,3\\
870: 3EG J0917+4427&J0920+4441&2.18&28.93&24.49&17.47&-0.1&1.19&0.86&0.81&0.0&4,1,4,3\\
871: 3EG J0952+5501&J0957+5522&0.901&27.64&23.87&16.42&0.39&1.12&0.78&0.88&0.39&4,1,4,3\\
872: 3EG J0958+6533&J0958+6533&0.368&26.64&24.70&15.33&-0.3&1.08&0.38&0.88&0.96&4,1,4,3\\
873: %3EG J1009+4855&J1015+4926&0.2&25.44&23.02&14.89&0.21&0.9&0.47&1.00&0&\\
874: 3EG J1052+5718&J1058+5628&0.14&25.88&23.54&15.18&0.09&1.51&0.31&1.09&0.24&5,1,5,3\\
875: 3EG J1104+3809&J1104+3812&0.031&23.85&22.32&13.46&0&0.57&0.32&1.07&0.35&4,1,4,3\\
876: 3EG J1133+0033&J1133+0040&1.633&27.84&24.07&16.21&0.05&1.73&0.74&0.89&0.82&1,1,1,3\\
877: 3EG J1200+2847&J1159+2914&0.729&27.34&23.99&16.18&0.24&0.98&0.68&0.93&1.17&4,1,4,3\\
878: 3EG J1219-1520&J1222-1645&\nodata&26.87&23.42&15.95&0.03&1.52&0.65&0.89&1.10&\nodata,2,8,3\\
879: 3EG J1222+2841&1221+2813&0.102&25.25&22.13&14.45&-0.2&0.73&0.64&0.93&0.62&4,1,4,3\\
880: 3EG J1224+2118&J1224+2122&0.435&26.66&23.20&15.79&0.34&1.28&0.71&0.88&0.45&4,1,4,3\\
881: 3EG J1227+4302&J1224+4335&1.872&27.69&23.74&$<$16.97&0.26&1.5&0.80&0.79&1.36&1,1,1,3\\
882: 3EG J1229+0210&J1229+0203&0.158&27.20&24.02&14.74&0.04&1.58&0.65&1.12&0.46&4,1,4,3\\
883: %3EG J1230-0247&J1232-0224&1.045&27.65&24.04&16.77&0.5&1.85&0.64&0.96&0.59&\\
884: 3EG J1236+0457&J1231+0418 &1.03&27.21&23.99&16.21&0.05 &1.48 &0.64&0.91&0.53&1,1,1,3\\
885: 3EG J1246-0651&J1246-0730&1.286&27.86&24.26&16.42&-0.08&1.73&0.71&0.6&\nodata&4,2,4,3\\
886: 3EG J1255-0549&J1256-0547&0.538&28.16&23.34&16.86&-0.2 &0.96&0.97&0.78&0.9&4,2,4,3\\
887: 3EG J1310-0517&J1312-04242&0.824&26.96&23.24&16.15&-0.15&1.34&0.76&0.82&\nodata&1,1,1,3\\
888: 3EG J1323+2200& J1327+2210&1.40&28.66&24.06&16.79&-0.5 &0.86&0.74&0.95&1.09&5,1,5,3\\
889: 3EG J1324-4314& J1325-4301&0.0018&21.78&22.00&10.74&0.66 &1.58&-0.05&1.36&0&16,7,13,3\\
890: 3EG J1329+1708& J1333+1649&2.09&29.71&25.36&16.76&-0.1 &1.41&0.57&0.99&0.62&4,1,4,3\\
891: 3EG J1339-1419 &J1337-1257&0.539&27.56&23.56&15.67&-0.21&1.62&0.80&0.93&0.81&4,2,4,3\\
892: 3EG J1347+2932 &J1343+2844&0.91&26.81&24.24&16.24&0.13&1.51&0.51&0.93&0.64&9,1,9,3\\
893: 3EG J1409-0745 &J1408-0752&1.494&28.03&24.28&17.32&0&1.29&0.74&0.81&1.42&4,2,4,3\\
894: 3EG J1424+3734 &J1419+3821&1.83&28.47&24.08&16.35&-0.1&2.25&0.85&0.84&\nodata&4,1,4,3\\
895: 3EG J1429-4217& J1427-4206&1.522&28.59&24.16&17.08&0.15&1.13&0.80&0.88&0.93&4,1,4,3 \\
896: 3EG J1457-1903 &J1459-1810&\nodata &26.98&$<$24.60&16.15&-0.24&1.67&0.78&0.79&0.49&\nodata,2,7,3\\
897: 3EG J1500-3509&J1457-3539&1.422& 27.94&23.97&16.36&0.04&1.99&0.79&0.85&0&2,2,2,3\\
898: 3EG J1504-1537&J1502-1508&\nodata&26.88&$<$23.02&$<$16.58&-0.18&\nodata&0.76&0.76&1.24&\nodata,2,\nodata,3\\
899: 3EG J1512-0849&J1512-0905&0.360&26.78&23.39&15.63&0.12&1.47&0.69&0.92&0&4,2,4,3\\
900: 3EG J1517-2538&J1517-2422&0.049&24.75 & 19.84 &13.39& 0.02&1.66&1.02&0.75&0&2,2,2,3\\
901: 3EG J1527-2358&J1532-2310&2.289&28.06&24.11&$<$16.81&-0.09&1.67&0.76&0.81&0.92&2,2,2,3\\
902: 3EG J1605+1553&J1603+1554&0.11&24.66&23.77&14.50&-0.5&1.47&0.17&1.12&0&1,1,1,3\\
903: 3EG J1608+1055&1608+1029&1.23&28.27&23.59&16.86&-0.1&1.06&0.84&0.82&1.09&5,1,5,3\\
904: 3EG J1607-1101&J1612-1133&\nodata&27.11&$<$23.02&$<$16.36&-0.58&\nodata&0.78&0.79&1.45&\nodata,2,\nodata,3\\
905: 3EG J1612-2618&J1611-2612&\nodata&26.80&$<$23.02&16.10&-0.17&1.71&0.75&0.79&1.04&\nodata,2,\nodata,3\\
906: 3EG J1614+3424&J1613+3412&1.40&28.58&24.56&17.15&0.13&1.42&0.81&0.86&0&4,1,4,3\\
907: 3EG J1621+8203&J1632+8232& 0.02&23.53&21.93&12.67&0.05& 1.29 &0.33&1.11&0&5,2,7,3\\
908: 3EG J1625-2955&J1626-2951&0.815&27.75&23.18&17.06&0.00&1.07&1.02&0.67&1.62&5,2,5,3\\
909: 3EG J1626-2519&1625-2527&0.786&27.74&22.63&16.60&0.45&1.21&1.17&0.65&1.05&5,2,5,3\\
910: 3EG J1634-1434&J1628-1415&1.025&27.15&23.08&16.65&0.14 &1.15&0.82&0.76&0&4,2,4,3\\
911: 3EG J1635-1751&J1629-1720&\nodata&27.27&$<$22.96&$<$16.36&-0.56&\nodata&0.83&0.78&0.83&\nodata,2,\nodata,3\\
912: 3EG J1635+3813&J1635+3808&1.81&28.86&24.79&17.94&0.04&1.15&0.80&0.79&0&4,1,4,3\\
913: 3EG J1646-0704&J1644-0743 &0.139&24.58&20.44&14.59&-0.08&1.39 &0.85&0.70&0.75&2,2,2,3\\
914: 3EG J1718-3313&J1717-3342&\nodata&27.56&$<$22.96&16.57&-0.06&1.59&0.92&0.74&0.82&\nodata,2,\nodata,3\\
915: 3EG J1720-7820&1723-7713&\nodata&27.38&23.61&16.12&-0.55&1.74&0.69&0.88&0.64&\nodata,7,8,3\\
916: 3EG J1727+0429&J1728+0427&0.29&26.01&22.97&15.33&0.04&1.67&0.62&0.91&0&4,1,4,3\\
917: 3EG J1733+6017&J1722+6105&2.06&28.04&24.07&16.55&-0.1&2.0&0.75&0.83&0.29&1,1,1,3\\
918: 3EG J1733-1313&1733-1304&0.9&28.37&23.63&16.96&-0.02&1.23&0.95&0.78&0.37&4,2,4,3\\
919: 3EG J1735-1500&J1738-1503&\nodata&27.43&23.33&$<$16.23&0.15&2.24&0.83&0.80&0.87&10,2,10,3\\
920: 3EG J1738+5203&J1740+5211&1.38&28.33&24.14&16.97&-0.2&1.42&0.81&0.83&0.51&4,1,4,3\\
921: 3EG J1744-0310&J1743-0350&1.05&28.16&23.78&16.52&-0.22&1.42&0.86&0.84&0.59&4,1,4,3\\
922: 3EG J1800-3955&J1802-3940&\nodata&26.79&22.80&15.27&0.07&2.1&0.82&0.91&0&17,2,\nodata,3\\
923: 3EG J1806-5005&J1808-5011&1.61&27.80&23.63&16.25&-0.33&1.93&0.84&0.82&0.88&12,7,12,3\\
924: 3EG J1824+3440&J1826+3431&1.81&27.78&25.23&$<$17.16&0.25&1.03&0.51&0.95&0.01&1,1,1,3\\
925: 3EG J1828+0142&J1826+0149&1.77&28.41&25.98&$<$16.91&-0.2&1.76&1.01&0.70&1.6&1,1,1,3\\
926: 3EG J1832-2110&J1833-2103&2.510&29.63&23.46&17.56&0.25&1.59&1.25&0.64&0.62&5,2,5,3\\
927: 3EG J1850-2652&J1848-2718&\nodata&27.71&$<$22.96&16.29&-1.26&1.19&0.88&0.78&0.81&\nodata,2,\nodata,3\\
928: 3EG J1904-1124&J1905-1153&\nodata &26.81&$<$22.96&16.51&0.26&1.6&0.79&0.75&0.08&\nodata,2,\nodata,3\\
929: 3EG J1911-2000&J1911-2006&1.119&28.16&23.73&16.77&0.11&1.39&0.89&0.81&0.3&11,2,11,3\\
930: 3EG J1921-2015&J1923-2104&0.871&27.82&24.11&$<$16.15&0.19&0.75&0.95&1.24&\nodata&11,2,11,3\\
931: 3EG J1935-4022&J1937-3958&0.965&27.82&23.58&16.02&-0.15&1.86&0.77&0.91&1.34&5,2,5,3\\
932: 3EG J1937-1529&J1939-1525&1.66&28.20&24.76&$<$15.94&0.0&2.45&0.83&0.88&1.35&4,2,4,3\\
933: 3EG J1959+6342&J2006+6424&1.57&28.41&$<$23.50&16.94&-0.3&1.45&0.95&0.75&0&5,1,5,3\\
934: 3EG J2006-2321&J2005-2310&0.830&26.79&23.29&16.13&0.14&1.33&0.71&0.84&1.31&2,2,2,3\\
935: 3EG J2025-0744&J2025-0735&1.388&27.88&24.15&17.07&0.33&1.38&0.76&0.82&0.87&5,2,5,3\\
936: 3EG J2027+3429&J2025+3343&0.22&26.32&21.08&15.39&-0.4&1.28&1.07&0.68&0&1,1,1,3\\
937: 3EG J2034-3110&J2030-3039&\nodata &27.08&25.10&15.44&-0.05&2.43&0.82&0.84&1.18&12,2,12,3\\
938: 3EG J2036+1132&J2034+1154&0.60&26.82&23.13&15.80&0.46&1.83&0.73&0.85&0&5,1,5,3\\
939: 3EG J2046+0933&J2049+1003&\nodata &27.84&$<$22.96&16.40&-0.6&1.22&0.94&0.77&0&\nodata,\nodata,1,3\\
940: 3EG J2055-4716&J2056-4714&1.49&28.38&23.99&17.02&0.32&1.04&0.89&0.82&0.93&4,15,4,3\\
941: 3EG J2100+6012&J2102+6015&4.57&28.59&26.13&18.33&0.35&1.21&0.52&0.88&0.13&2,2,2,3\\
942: 3EG J2158-3023&J2158-3013&0.12&24.61&23.29&14.51&0.44&1.35&0.27&1.06&0.59&4,2,4,3\\
943: 3EG J2202+4217&J2202+4216 &0.07&25.31&22.58&13.85&0.31&1.6&0.57&1.05&0.8&1,1,4,3\\
944: 3EG J2206+6602&J2208+6519&1.12&27.14&22.69&16.98&0.33&1.29&0.91&0.66&0&1,1,1,3\\
945: 3EG J2209+2401&J2212+2355&1.13&27.76&23.70&16.32&-0.1&1.48&0.80&0.86&0.91&1,1,1,3\\
946: 3EG J2232+1147&J2232+1143 &1.04&28.08&24.27&16.70&0.48&1.45&0.79&0.88&0.48&4,1,4,3\\
947: 3EG J2254+1601&J2253+1608&0.86&28.50&24.54&17.10&0.1&1.21&0.80&0.88&0.52&4,1,4,3\\
948: 3EG J2255+1943&J2253+1942&0.28&25.75&22.53&14.94&-0.1&1.36&0.66&0.91&1.18&5,1,4,3\\
949: 3EG J2321-0328&J2323-0317&1.411&28.12&24.13&$<$16.78&0.0&\nodata&0.79&0.87&1.36&4,2,4,3\\
950: %3EG J2352+3752&2346+385&1.03&27.30&23.27&16.71&0.22&1.47&0.71&0.86&1.37&\\
951: 3EG J2358+4604&J2354+4553&1.99&28.45&23.74&17.25&0.34&1.38&0.96&0.74&0&4,1,4,3 \\
952: 3EG J2359+2041&J0003+2129&0.45&26.25&21.58&15.67&-0.6&1.09&0.93&0.71&0.74&1,1,1,3\\
953: \enddata
954: \tablerefs{(1) Sowards-Emmerd {\it et al.} 2003 (2)Sowards-Emmerd {\it
955: et al} 2004 2004 (3) Hartman {\it et al.} 1999 (4)Hewitt \& Burbidge 1989
956: (5)Veron-Cetty \& Veron 2001 (6)Bloom {\it et al} 2004 (7) NED
957: (8) Monet {\it et al.} (9) Stocke {\it et al.} 1991 (10) Combi {\it et al.} (11)
958: Halpern {\it et al.} 2003 (12) Landt {\it et al.} 2001 (13)
959: Lauberts \& Valentijn 1989 (14) Wisotzki {\it et al.} 2000 (15)
960: Wright \& Otrupcek 1990 (16) Fouque {\it et al.} (17) Liang \& Liu 2003}
961: \end{deluxetable}
962: \begin{deluxetable}{lllccccccc}
963: \tabletypesize{\scriptsize}
964: %\rotate
965: \tablecolumns{10}
966: \tablewidth{0pt}
967: \tablenum{2}
968: \tablecaption{Correlation and Regression Analysis for EGRET Blazars}
969: \tablehead{\colhead {Ind. Var.}& \colhead {Dep. Var.}&\colhead {N}
970: &\colhead {$\tau$}&\colhead{Prob.}&\colhead{$\rho$}&\colhead{Prob.}&\colhead{ Technique}&\colhead {Slope}&\colhead{Intercept} \\
971: \colhead{(1)} &\colhead{(2)} &\colhead{(3)} &\colhead{(4)}
972: &\colhead{(5)} &\colhead{(6)}&\colhead{(7)}& \colhead {(8)} &
973: \colhead{(9)}& \colhead {(10)}\\}
974: \startdata
975: log $L_r$& log $L_{\gamma}$ &122&0.628 & $1 \times 10^{-10}$ &
976: 0.806 & $1 \times 10^{-10}$ & EM & 0.766 &-4.81 \\
977: log z& log $L_r$ & 122 &0.650 & $1 \times 10^{-10}$ &0.825 &$1
978: \times 10^{-10}$ & EM &2.47 &27.7 \\
979: log z& log $L_{\gamma}$&122&0.569 & $1 \times 10^{-10}$ &0.705
980: &$1 \times 10^{-10}$ &EM &2.12 & 16.4\\
981: log $L_o$ &log $L_{\gamma}$ & 122&0.299&$1 \times 10^{-6}$ &0.437
982: & $2 \times 10^{-6}$ & SB & 0.566 & 2.71\\
983: log z & log $L_o$ & 122 &0.404 & $1 \times 10^{-10}$ &0.564 & $ 7
984: \times 10^{-10}$& EM&1.26 & 23.7\\
985: $\alpha_{ro}$&$\alpha_{o \gamma}$&122&-0.556 &$1 \times
986: 10^{-10}$&-0.701&$1 \times 10^{-10}$ &BJ&-0.519&1.23\\
987: log$L_o$ & $\alpha_{ro}$ &122 & -0.094& $0.122$&-0.111&0.222&SB&-0.007 &0.892 \\
988: log $L_o$ &$\alpha_{o \gamma}$&122 &0.214& $7 \times
989: 10^{-4}$&0.309& $7. \times 10^{-4}$ &SB&0.032&0.113\\
990: %$\alpha_{\gamma}$ & $\delta_{var}$ &118 &-0.099 &0.112 &-0.156 &0.0923 &EM &-0.166 &0.989\\
991: \enddata
992: \end{deluxetable}
993: \begin{deluxetable}{lllccc}
994: \tabletypesize{\scriptsize}
995: %\rotate
996: \tablecolumns{6}
997: \tablewidth{0pt}
998: \tablenum{3}
999: \tablecaption{Partial Correlation Analysis for EGRET Blazars}
1000: \tablehead{\colhead {Ind. Var.}& \colhead {Dep. Var.}&\colhead {Third
1001: Var.} &\colhead {N}&\colhead {$\tau$} &\colhead{Prob.} \\
1002: \colhead{(1)} &\colhead{(2)} &\colhead{(3)} &\colhead{(4)}
1003: &\colhead{(5)} &\colhead{(6)}\\}
1004: \startdata
1005: log $L_r$& log $L_{\gamma}$ & log z &122& 0.341 & $1 \times 10^{-10}$ \\
1006: log $L_o$ &log $L_{\gamma}$ & log z &122&-0.089 &0.148 \\
1007: $\alpha_{ro}$& $\alpha_{o \gamma}$ & log $L_o$ & 122& 0.550&$2 \times
1008: 10^{-8}$\\
1009: \enddata
1010: \end{deluxetable}
1011: \begin{deluxetable}{lllccccccccccc}
1012: \tabletypesize{\scriptsize}
1013: %\rotate
1014: \tablecolumns{14}
1015: \tablewidth{0pt}
1016: \tablenum{4}
1017: \tablecaption{Parameters for Monte Carlo Simulations}
1018: \tablehead{\colhead {Identifier}& \colhead {$log L_{1}$}&\colhead
1019: {$log L_{2}$}& \colhead {g}&\colhead { $\Gamma_1$} &\colhead{
1020: $\Gamma_2$}&\colhead{s} &\colhead {$\tau $}&\colhead {$z_{min}$}
1021: &\colhead {$z_{max}$}&\colhead {$\rm h$} &\colhead {$K_{min}$} &\colhead {$K_{max}$}
1022: &\colhead {$\rm slope$}\\
1023: \colhead{(1)} &\colhead{(2)} &\colhead{(3)} &\colhead{(4)}
1024: &\colhead{(5)} &\colhead{(6)}&\colhead{(7)} & \colhead {(8)} &\colhead
1025: {(9)} &\colhead {(10)} &\colhead {(11)} &\colhead {(12)} &\colhead {(13)} &\colhead {(14)}\\ }
1026: \startdata
1027: Lister SSC& 22.4 & 26.0 & 2.5& 1.001& 25.0&1.5 &0.26&0.0& 4.0& 0.65 &$5 \times 10^{-14}$ & $1.0 \times 10^{-11}$ &1.6\\
1028: Lister ECS & 22.4 & 26.0 &2.5 &1.001& 25.0& 1.5 &0.26&0.0 & 4.0& 0.65& $1.0 \times 10^{-16}$ &$ 1.0 \times 10^{-15}$ & 2.3\\
1029: Bloom SSC & 21.5 & 30.0 & 2.0& 1.001& 30.0 & 1.5 & 0.26& 0.0&5.0 &1.0& $5 \times 10^{-14}$ & $1 \times 10^{-11}$ & 1.6\\
1030: Bloom ECS & 21.3 & 30.0 & 2.0 &1.001& 30.0 & 1.5 & 0.26& 0.0&5.0& 1.0& $4.8
1031: \times 10^{-17}$ & $5 \times 10^{-17}$& 2.3\\
1032: \enddata
1033: \end{deluxetable}
1034: \begin{deluxetable}{lllcccccccc}
1035: \tabletypesize{\scriptsize}
1036: %\rotate
1037: \tablecolumns{11}
1038: \tablewidth{0pt}
1039: \tablenum{5}
1040: \tablecaption{Kolmogorov-Smirnov Tests for Monte Carlo Simulations}
1041: \tablehead{\colhead {Identifier}& \colhead {$D_{Lr}$}&\colhead
1042: {Prob}& \colhead {$D_{L,\gamma}$}&\colhead {Prob} &\colhead{
1043: $D_z$}&\colhead{Prob} &\colhead {$D_{fr} $}&\colhead {Prob}
1044: &\colhead {$D_{f\gamma}$}&\colhead {Prob}\\
1045: \colhead{(1)} &\colhead{(2)} &\colhead{(3)} &\colhead{(4)}
1046: &\colhead{(5)} &\colhead{(6)}&\colhead{(7)} & \colhead {(8)} &\colhead
1047: {(9)} &\colhead {(10)} &\colhead {(11)}\\}
1048: \startdata
1049: Lister SSC&0.516 &$1 \times 10^{-15}$ &0.172& 0.054
1050: &0.328 & $4 \times 10^{-6}$&0.303& $3 \times 10^{-5}$& 0.148
1051: & 0.140\\
1052: Lister ECS & 0.426 & $5 \times 10^{-10}$ &0.148 & 0.398& 0.271&0.004 &0.156 &0.103 & 0.189& 0.026\\
1053: Bloom SSC & 0.205 & 0.012 & 0.230& 0.003&
1054: 0.180 & 0.038& 0.205&0.012& 0.164& 0.075 \\
1055: Bloom ECS & 0.164 & 0.075&0.246 & 0.001& 0.213& 0.008& 0.271& 0.003& 0.197& 0.018\\
1056: \enddata
1057: \end{deluxetable}
1058: \end{document}
1059:
1060:
1061:
1062:
1063:
1064:
1065:
1066:
1067:
1068:
1069: