astro-ph0610347/ms.tex
1: % SAMPLE2.TEX -- AASTeX macro package tutorial paper.
2: 
3: %\documentclass[12pt,preprint]{aastex}
4: %\documentstyle[12pt,aasms4]{article}
5: %\documentclass[10pt,preprint2]{aastex}
6: \documentclass{emulateapj}
7: 
8: \def\sun{\odot}
9: 
10: \slugcomment{to appear in the Astrophysical Journal Letters}
11: 
12: \lefthead{Tomita et al.}
13: \righthead{Applicability of the multiple regression analysis to the variable component in the near-infrared region for type I AGNs}
14: 
15: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
16: \begin{document}
17: 
18: \title{Multiple regression analysis of the variable component in the near-infrared region 
19: for type 1 AGN MCG+08-11-011}
20: 
21: \author{Hiroyuki Tomita\altaffilmark{1,2},
22:    Yuzuru Yoshii\altaffilmark{3,6},
23:    Yukiyasu Kobayashi\altaffilmark{4},
24:        Takeo Minezaki\altaffilmark{3},
25:            Keigo Enya\altaffilmark{7},
26:     Masahiro Suganuma\altaffilmark{4},
27:          Tsutomu Aoki\altaffilmark{1},
28:          Shintaro Koshida\altaffilmark{4,5},
29: and         
30:        Masahiro Yamauchi\altaffilmark{4,5}
31: }
32: 
33: \altaffiltext{1}{Kiso Observatory, Institute of Astronomy, School of Science,
34:                  University of Tokyo, 10762-30
35:                  Mitake, Kiso-machi, Kiso-gun, Nagano 397-0101, Japan}
36: \altaffiltext{2}{Email: tomitahr@cc.nao.ac.jp}
37: \altaffiltext{3}{Institute of Astronomy, School of Science,
38:                  University of Tokyo, 2-21-1
39:                  Osawa, Mitaka, Tokyo 181-0015, Japan}
40: \altaffiltext{4}{National Astronomical Observatory, 2-21-1
41:                  Osawa, Mitaka, Tokyo 181-8588, Japan}
42: \altaffiltext{5}{Department of Astronomy, School of Science,
43:                  University of Tokyo, 7-3-1
44:                  Hongo, Bunkyo-ku, Tokyo 113-0033, Japan}
45: \altaffiltext{6}{Research Center for the Early Universe,
46:                  School of Science, University of Tokyo, 7-3-1
47:                  Hongo, Bunkyo-ku, Tokyo 113-0033, Japan}
48: \altaffiltext{7}{Japan Aerospace Exploration Agency, 3-1-1
49:                  Yoshinodai, Sagamihara, Kanagawa, 229-8510, Japan}
50: 
51: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
52: 
53: \begin{abstract}
54: 
55: We propose a new method of analysing a variable component for type 1 active galactic nuclei (AGNs) in the near-infrared wavelength region.  This analysis uses a multiple regression technique and divides the variable component into two components originating in the accretion disk at the center of AGNs and from the dust torus that far surrounds the disk. Applying this analysis to the long-term $VHK$ monitoring data of MCG+08-11-011 that were obtained by the MAGNUM project, we found that the $(H-K)$-color temperature of the dust component is $T = 1635$K $\pm20$K, which agrees with the sublimation temperature of dust grains, and that the time delay of $K$ to $H$ variations is $\Delta t\approx 6$ days, which indicates the existence of a radial temperature gradient in the dust torus.  As for the disk component, we found that the power-law spectrum of $f_\nu \propto \nu^\alpha$ in the $V$ to near-infrared $HK$ bands varies with a fixed index of $\alpha\approx -0.1$ --- +0.4, which is broadly consistent with the irradiated standard disk model. The outer part of the disk therefore extends out to a radial distance where the temperature decreases to radiate the light in the near-infrared.
56: 
57: 
58: \end{abstract}
59: \keywords{galaxies: active---galaxies: individual
60: (MCG+08-11-011)---accretion disks---dust}
61: 
62: 
63: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
64: \section{Introduction}
65: 
66: The existence of massive black holes at the center of accretion 
67: disks in AGNs is well accepted in current AGN research.  The so-called big 
68: blue bump, which is observed in the UV to optical wavelength region 
69: of spectral energy distribution (SED) of many AGNs, is regarded as an 
70: evidence of the accretion disk (e.g., Shields 1978; Malkan \& Sargent 1982; 
71: Czerny \& Elvis 1987).  Based on another distinct feature of near-infrared 
72: bumps in their SED, the existence of a surrounding dust torus, which is 
73: heated to the 
74: sublimation temperature of dust by the accretion disk, is also accepted now 
75: (e.g., Sanders et al. 1989; Barvainis 1987; Kobayashi et al. 1993).  
76: 
77: In practice, however, it is known that composite SEDs constructed from a 
78: huge number of AGNs obtained by large surveys show their optical colors 
79: to be significantly redder than predicted by the standard accretion disk model 
80: (see Francis et al. 1991 for the composite SED from the low-redshift bright 
81: quasar survey (LBQS); Vanden Berk et al. 2001 for the composite SED from the 
82: Sloan digital sky survey (SDSS)).  Since AGNs are superimposed on host 
83: galaxies, how serious this discrepancy is depends on how reliably the 
84: nuclear SED has been separated from the SED of the galaxy component. 
85: 
86: In addition to the problem of SED decomposition, a possible contamination 
87: of near-infrared light from accretion disk and dust torus further complicates 
88: the interpretation of the SED.  In fact, it is reported that the SED of the accretion 
89: disk extends to the near-infrared wavelength region (Kishimoto et al. 2005; Minezaki et al. 2006), 
90: and that the temperature gradient exists in the dust torus of which the inner 
91: part is hotter than the outer part (e.g. Barvainis 1987; Glass 1992; Alloin 
92: et al. 1995). These two factors cause similar effects such that variations 
93: at shorter near-infrared wavelengths are less delayed behind the optical 
94: variations, when compared to variations at longer near-infrared wavelengths. 
95: Therefore, unless the above factors are taken into account simultaneously in 
96: the analysis, the time delay of near-infrared variations behind the optical 
97: is liable to be underestimated.
98: 
99: In order to overcome such complications as well as the problem of SED 
100: decomposition, we propose a new method of analysing the variable SEDs of 
101: accretion disks and dust tori, based on a multiple regression technique.  
102: In section 2, we describe the assumptions and methods of the analysis.  
103: In section 3, we 
104: apply the analysis to the long-term $VHK$ monitoring data of MCG+08-11-011 
105: obtained by the MAGNUM project (Yoshii 2002; Kobayashi et al. 2003; Yoshii et al. 2003).  
106: Finally, in section 4, we describe the applicability of the analysis and discuss the results.
107: 
108: 
109: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
110: \section{Multiple Regression Analysis}
111: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
112: 
113: \subsection{Overview of the Analysis}
114: 
115: Let $F_\lambda(t)$ be the observed flux at wavelength $\lambda$ 
116: and time $t$, then it is decomposed as 
117: \begin{equation}
118: F_\lambda(t)=F_\lambda^{var,disk}(t)+F_\lambda^{var,dust}(t)+
119: F_\lambda^{const}, 
120: \end{equation}
121: where the first two are the variable components from the accretion disk 
122: and dust torus, respectively, and the last component is a constant, partly
123:  offset including the contribution from the host galaxy. 
124: 
125: Kobayashi et al. (1993) showed that all 14 QSOs in their sample have 
126: almost the same blackbody spectra of 1500K. This indicates that the flux 
127: in the near-infrared wavelength region is mainly radiated from the 
128: innermost part of the dust torus, where the temperature is near the sublimation 
129: temperature of dust grains.  Hence, as we observe for each AGN, the 
130: near-infrared light curves in different bands have almost the same shape 
131: and overlap each other by shifting the lag time between the variations in 
132: different bands (e.g., Suganuma et al. 2006).  When there is a short
133: lag time of $\tau$ between variations in two near-infrared bands, 
134: $H(1.7\mu m)$ and $K(2.2\mu m)$ for example, which are intrinsic to the 
135: dust torus, the situation considered here corresponds to 
136: \begin{equation}
137: F_H^{var,dust}(t-\tau)/F_K^{var,dust}(t)=a .
138: \end{equation}
139: Winkler et al. (1992) and Winkler (1997) claimed that the optical colors of 
140: the variable component for Seyfert galaxies remain constant as their brightness
141: changes, while recent studies report that QSOs get bluer as they get brighter 
142: (e.g., Trevese \& Vagnetti 2002; Giveon et al. 1999). For our analysis 
143: focusing on Seyfert galaxies below, we then assume that the variable 
144: component synchronized 
145: with optical variation, which comes from the accretion disk, has constant 
146: optical $V(0.55\mu m)$ to near-infrared $H$ and $K$ colors, i.e., 
147: \begin{equation}
148: F_H^{var,disk}(t)/F_V(t)=b \;\;\; {\rm and} \;\;\; 
149: F_K^{var,disk}(t)/F_V(t)=c . 
150: \end{equation}
151: Substitution of equations 2 and 3 in equation 1 therefore gives
152: \begin{equation}
153: F_H(t-\tau)-b\times F_V(t-\tau)=a\times (F_K(t)-c\times F_V(t))+d , 
154: \end{equation}
155: where $a$, $b$, $c$, and $d$ are the constant parameters to be determined 
156: by the fit to the $VHK$ monitoring data, based on a multiple regression 
157: technique described below.
158: 
159: 
160: \subsection{Details of Calculation}
161: 
162: For calculation of the multiple regression, Jefferys (1980) provides the 
163: most generalized multivariate least-square process. This process allows errors 
164: to appear in all variables, which is desirable for astronomical observations. 
165: In practice, we adopted the Mardquardt method described in Jefferys (1981), 
166: which is robust for the case that the initial trial for the fit is far from 
167: the true solution.
168: 
169: The calculation of the regression with a given value of lag time $\tau$ 
170: requires interpolation of the monitoring data in time, because such data 
171: are usually discrete.  We here adopt the linear interpolation scheme, and 
172: smaller weights are assigned to the data produced with large interpolation.  
173: We estimate the uncertainties of the interpolation, using the structure 
174: function (SF) of AGN variabilities (Collier \& Peterson 2001; Suganuma et 
175: al. 2004):
176: \begin{equation}
177: \sigma_{\rm SF}^2={\rm SF}(l)=\frac{1}{N(l)}\sum_{i<j}[ f(t_i)-f(t_j)]^2-2\sigma^2 ,
178: \end{equation}
179: where the summation runs over all $(i,j)$-pairs, $l=t_j-t_i$ is the length of 
180: interpolation, 
181: $N(l)$ is the number of pairs at $l$, $f(t)$ is the flux at time $t$, and 
182: $\sigma$ is the mean observational error for the data used for the SF 
183: calculation. We adopted a length nearer to the neighboring data point as the length of interpolation $l$. While we naturally use the observational error as the weight 
184: for the data with no interpolation, we adopt the larger of the SF error or 
185: the observational error as representing the uncertainty of the interpolated 
186: data in the regression analysis. 
187: In order to determine whether or not the fit well represents the observations,
188:  the $\chi^2$-test is performed. The $\chi^2$-value is computed by 
189: doubling equation 8 of Jefferys (1980), and this doubled value follows the 
190: $\chi^2$ distribution of $N=m-4$ degrees of freedom, where $m$ is the number 
191: of data sets.
192: 
193: Finally, the value of lag time $\tau$ that gives the best fit is estimated 
194: at the peak of multiple correlation coefficient that measures the strength 
195: of association between the independent and dependent variables.
196: 
197: 
198: \vspace{0.5cm}
199: \centerline{{\vbox{\epsfxsize=7.5cm\includegraphics[angle=-90,width=7.5cm]{f1.eps}}}}
200: \figcaption{Light curves of MCG+08-11-011 for the period from 8 September 2001 
201: to 31 December 2002 (UTC).  Shown are the monitoring data for the $B$ band 
202: (open circles), $V$ band (crosses), $H$ 
203: band (open rectangles), and $K$ band (filled rectangles), where the flux from 
204: the host galaxy has been subtracted.\label{f1}}
205: \vspace{0.5cm}
206: 
207: 
208: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
209: \section{Application of the Analysis}
210: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
211: 
212: The $BVHK$ monitoring data for MCG+08-11-011 were obtained by the MAGNUM 
213: project in the period from 8 September 2001 to 31 December 2002 (Figure 1).
214: Aperture photometry was used throughout. An aperture size of 8.3 arcsec was
215: adopted for all images. 
216: The flux of MCG+08-11-011 was measured relative to a reference star 
217: ($\alpha_{2000}= 05^h 54^m47.^s9, \delta_{2000}=+46^{\circ}23'5''$) whose 
218: flux was absolutely calibrated using Landolt standard stars (Landolt 1992) in 
219: $BV$ and Hunt standard stars (Hunt et al. 1997) in $HK$  
220: during the monitoring observations. The host galaxy flux was estimated using  
221: the GALFIT program (Peng et al. 2002) and was subtracted from the measured 
222: flux of MCG+08-11-011. The estimated magnitudes of the host galaxy were 
223: 16.5 mag at $B$, 15.5 mag at $V$, 12.0 mag at $H$ and 11.5 mag at $K$. 
224: A small error of 0.003 mag caused by the seeing effect, arising from 
225: the profile difference between the AGN and the reference star, was taken 
226: into account. Galactic 
227: extinction was corrected according to Schlegel, Finkbeiner, \& Davis (1998).
228: 
229: We performed the regression analysis comparing the observed $H$-band light 
230: curve with the predicted $H$-band light curve in equation 4. We adopted 
231: only the data set where observations in $V$, $H$ and $K$ were carried out 
232: on the same night in order to decrease the effect of interpolation. 
233: The parameters used for SF$(l)=S_0l^\beta$ were $S_0=0.033\; \mathrm{mJy}^2$ 
234: and $\beta=1.1$ for $V$, and $S_0=0.0077\; \mathrm{mJy}^2$ and $\beta$=2.0 for 
235: $K$.  Zero-correlation was assumed between the data in any two bands, and  
236: the standard error was used for the error of the fitted parameters.
237: 
238: We carried out the regression analysis for a range of $\tau$ from $-10$ to 
239: 60 days. The value of the multiple correlation coefficient was calculated as a 
240: function of $\tau$, and its peak was found to occur at $\tau=6.0$ days with 
241: the $\chi^2$-value of 12.5 for $m=26$ (Figure 2). The reduced chi-square 
242: value becomes $\approx$ 0.5, which indicates that the fit is good enough 
243: to represent the observations. The regression at 
244: $\tau=6.0$ days yielded that the $(H-K)$-color temperature for the dust 
245: component is 1635$\pm$20K, and the index of power-law SED for the accretion 
246: disk component ($f_\nu \propto \nu^\alpha$) is $\alpha=+0.36_{-0.25}^{+0.34}$ 
247: from $V$ to $H$ and $\alpha=-0.06_{-0.24}^{+0.35}$ from $V$ to $K$. 
248: 
249: \vspace{0.5cm}
250: \centerline{{\vbox{\epsfxsize=7.5cm\includegraphics[angle=-90,width=7.5cm]{f2.eps}}}}
251: \figcaption{Multiple correlation coefficient as a function of lag time $\tau$ 
252: between variations in the $H$ and $K$ bands for the case of MCG+08-11-011.  
253: The coefficient value is peaked at $\tau=6.0$ days.\label{f2}}
254: \vspace{0.5cm}
255: 
256: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
257: \section{Discussion}
258: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
259: 
260: The $(H-K)$-color temperature of 1635K here derived for MCG+08-11-011 agrees 
261: well with the sublimation temperature of $1500-1800$K for graphite/silicate 
262: grains (Salpeter 1977; Huffman 1977).  Hence, this component represents the 
263: innermost part of the dust torus heated to the sublimation temperature by the 
264: accretion disk.
265: 
266: Following Winkler et al. (1992), we performed the flux variation gradient 
267: (FVG) analysis for the $BV$ monitoring data (Figure 3). We converted the 
268: FVG $(B-V)$-color to the index of the power-law SED and obtained 
269: $\alpha$=+0.21$\pm$0.10.  Combining this and the results in section 3 
270: from the regression analysis, we found that the power-law SED with 
271: $\alpha=-0.1\sim +0.4$ prevails in the optical to near-infrared wavelength 
272: region (Figure 4).  
273: 
274: For comparison, we computed theoretical colors of the 
275: standard accretion disk model (Shakura \& Sunyaev 1973) with a black hole 
276: mass of $M=10^7 M_{\odot}$ and an outer disk radius of $R=10^4 R_g$, where 
277: $R_g = GM/c^2$.  Figure 4 shows $\alpha$ as a function of frequency $\nu$ 
278: for two cases of accretion rate $\dot{M}=0.1 M_{\odot}$yr$^{-1}$ 
279: (upper thick line) and $0.001 M_{\odot}$yr$^{-1}$ (lower thick line). 
280: These predicted SEDs have an approximate value of $\alpha=1/3$, and agree 
281: with our finding of $\alpha=-0.1\sim +0.4$ for MCG+08-11-011. 
282: 
283: The standard disk model has a critical difficulty for explaining the 
284: variability.  Although many UV/optical monitoring observations have so 
285: far been made, all of them have failed to detect significant lags between UV and 
286: optical variations, except for NGC7469 (Peterson et al. 1998; Wanders et al. 
287: 1997; Collier et al. 1998) and Ark564 (Collier et al. 2001).  The expected lag 
288: time is too short for the energy to flow at sound speed within the standard 
289: viscous disk.  However, such a lag time is sufficient for the incident 
290: radiation to transfer the energy to any distant parts of the disk 
291: (Krolik et al. 1991).  Hence, the irradiated accretion disk model has been 
292: invoked, where the central X-ray energy source illuminates a geometrically 
293: thin disk, and UV/optical light is radiated at the outer part of the disk. 
294: The predicted spectrum of the 
295: irradiated standard disk model has an index of $\alpha=1/3$ at the radii 
296: where the height of the X-ray source is negligible compared to the radial 
297: distance from the center. This spectrum is exactly the same as that of the 
298: standard viscous disk model. Furthermore, the detected wavelength-dependent 
299: lag time agrees with $\tau\propto\lambda^{4/3}$ in the UV to optical region 
300: for NGC7469 and Ark564, which is the same as the prediction for the irradiated 
301: standard disk model (Collier et al. 1998; Collier et al. 2001).  Consequently, 
302: we conclude that the power-law SED obtained by our analysis 
303: reinforces evidence that irradiated standard disks exist in the center 
304: of AGNs. 
305: 
306: \vspace{0.5cm}
307: \centerline{{\vbox{\epsfxsize=7.5cm\includegraphics[angle=-90,width=7.5cm]{f3.eps}}}}
308: \figcaption{The $B$-flux to $V$-flux diagram for MCG+08-11-011. 
309: The straight line is the least-square fit to the data.}
310: 
311: \vspace{0.5cm}
312: \centerline{{\vbox{\epsfxsize=7.5cm\includegraphics[angle=-90,width=7.5cm]{f4.eps}}}}
313: \figcaption{Values of power index $\alpha$ plotted against the rest-frame 
314: frequency of radiation. Filled circles are our results for MCG+08-11-011 
315: from the data of $BV$, $VH$ and $VK$. Vertical bar indicates  the estimated 
316: error of 
317: $\alpha$, and horizontal bar corresponds to the wavelength range between 
318: two bands used for calculation of $\alpha$. For reference, open circles show 
319: the values of $\alpha$ for other AGNs by Winkler et al. (1992), and two thick 
320: lines show the predictions of two standard accretion disk models. 
321: See text for detail.}
322: 
323: \vspace{0.5cm}
324: 
325: 
326: Our analysis showed that the power-law SED extends to the near-infrared 
327: wavelength region. By the standard accretion disk model, the near-infrared 
328: $K$-emitting region in the disk is located at a radial distance of $R\approx 
329: 10^3R_g$, from the center of typical Seyfert 
330: galaxies, which is close to the 
331: inner part of the broad line region (BLR) where high-ionization lines 
332: originate.  
333: Collin \& Hure (2001) showed that the disk would become gravitationally 
334: unstable at $R>R_{crit}\approx 2 \times 10^4R_g(M/10^7M_{\odot})^{-0.46}$. 
335: Based on this 
336: estimation, the $K$-emitting region in the disk could be gravitationally 
337: stable, and the existence of such a stable extended disk could distort the Kepler 
338: motion of BLR gas clouds.  This may explain the significant scatter of 
339: observed virial relationships of the broad-emission lines, in accord with the 
340: suggestion of Peterson et al. (2004) that the possible existence of some 
341: additional components may affect the BLR gas motion. 
342: Thus, near-infrared observations of AGNs bring information of physical 
343: coupling between the extended disk and the BLR.  Particularly, in the case 
344: of AGNs with high Eddington ratio, the $K$-emitting region in the disk 
345: would be further away from the centers of AGNs, and the outer disk 
346: edges might be observable (Collin et al. 2002; Kawaguchi 2003; 
347: Kawaguchi et al. 2004). 
348: 
349: The accretion disk component contributes to the near-infrared flux by 
350: 15-30\% in the $H$ band and by 15-25\% in the $K$ band (Figure 5).  
351: Since the observational errors are of order 1\% only, these disk 
352: contributions are significant, and cannot be ignored in deriving  
353: accurate and reliable results from near-infrared observations of AGNs. 
354: For a long time, however, separation of accretion disk components 
355: in the near-infrared 
356: region has been considered difficult, because the thermal emission 
357: from the dust torus component is the dominant source there.  In this context, 
358: the multiple regression analysis for multicolor monitoring observations is 
359: indeed a promising method for extracting the accretion disk component from 
360: near-infrared data, and will be useful for investigating the physical conditions 
361: and emission mechanisms of both components of accretion disk and dust torus 
362: simultaneously. 
363: 
364: %%%%%%%
365: 
366: 
367: 
368: \vspace{0.5cm}
369: \centerline{{\vbox{\epsfxsize=7.5cm\includegraphics[angle=-90,width=7.5cm]{f5.eps}}}}
370: \figcaption{Fractional contribution of near-infrared flux from the accretion 
371: disk component throughout the monitoring observations of MCG+08-11-011. Shown 
372: are the values of $F_\lambda^{var,disk}(t)/F_\lambda(t)$ for the $H$ band 
373: (open squares) and for the $K$ band (filled squares).}
374: \vspace{0.5cm}
375: 
376: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
377: \acknowledgements
378: We thank B. A. Peterson for helpful discussions. This research has been supported partly by the Grant-in-Aid of Scientific Research (10041110, 10304014, 11740120, 12640233, 14047206, 14253001, 14540223, 16740106) and COE Research (07CE2002) of the Ministry of Education, Science, Culture and Sports of Japan.
379: 
380: 
381: %%%%%%%%%%%%%%%%%%%
382: % references
383: %%%%%%%%%%%%%%%%%%%
384: 
385: \begin{references}
386: 
387: \reference{}Alloin, D., et al. 1995, \aap, 293, 293
388: 
389: \reference{}Barvainis, R. 1987, \apj, 320, 537
390: 
391: \reference{}Collier, S. \& Peterson, B. M. 2001, \apj, 555, 775
392: 
393: \reference{}Collier, S., et al. 1998, \apj, 500, 162
394: 
395: \reference{}Collier, S., et al. 2001, \apj, 561, 146
396: 
397: \reference{}Collin, S., \& Hure, J. M. 2001, \aap, 372, 50
398: 
399: \reference{}Collin, S., et al. 2002, \aap, 388, 771
400: 
401: \reference{}Czerny, B. \& Elvis, M., 1987, \apj, 321, 305
402: 
403: \reference{}Francis, P. J., Hewett, P. C., Foltz, C. B., Chaffee, F. H., Weymann, R. J., \& Morris, S. L. 1991, \apj, 373, 465
404: 
405: \reference{}Giveon, U., Maoz, D., Kaspi, S., Netzer, H., \& Smith, P. S. 1999, \mnras, 306, 647
406: 
407: \reference{}Glass, I. S., 1992, \mnras, 256, 23
408: 
409: \reference{}Huffman, D. R. 1977, Adv. Phys., 26, 129
410: 
411: \reference{}Hunt, L. K., Malkan, M. A., Salvati, M., Mandolesi,
412:    N., Palazzi, E., \& Wade, R. 1997, \apjs, 108, 229
413: 
414: \reference{}Jefferys, W. H. 1980, \aj, 85, 177
415: 
416: \reference{}Jefferys, W. H. 1981, \aj, 86, 149
417: 
418: \reference{}Kawaguchi, T. 2003, \apj, 593, 69
419: 
420: \reference{}Kawaguchi, T., Perens, A., \& Hure, J. M. 2004, \aap, 415, 47
421: 
422: \reference{}Kishimoto, M., Antonucci, R., \& Blaes, O. 2005, \mnras, 364, 640
423: 
424: \reference{}Kobayashi, Y., Sato, S., Yamashita, T., Shiba, H., Takami, H. 1993, \apj, 404, 94
425: 
426: \reference{}Kobayashi, Y., Yoshii, Y., Peterson, B. A., 
427:     Minezaki, T., Enya, K., Aoki, T., Suganuma, M., Tomita, H., 
428:     Doi, M., Motohara, K., Smith, C., Little, J., Greene, B. 
429:     2003, Proc. SPIE Vol. 4837,954
430: 
431: \reference{}Krolik, J. H., et al. 1991, \apj, 371, 541
432: 
433: \reference{}Landolt, A. U. 1992, \aj, 104, 340
434: 
435: \reference{}Malkan, M. A. \& Sargent, W. L. W. 1982, \apj, 254, 22
436: 
437: \reference{}Minezaki, T., et al. 2006, \apj, 643, L5
438: 
439: \reference{}Peng, C. Y., Ho, L. C., Impey, C. D., \& Rix, H.-W. 2002, \aj, 124 266
440: 
441: \reference{}Peterson, B. M., Wanders, I., Horne, K., Collier, S., Alexander, T., Kaspi, S., \& Maoz, D. 1998, \pasp, 110, 660
442: 
443: \reference{}Peterson, B. M., et al. 2004, \apj, 613, 682
444: 
445: \reference{}Salpeter, E. E. 1977, \araa, 15, 267
446: 
447: \reference{}Sanders, D. B., Phinney, E. S., Neugebauer, G., Soifer, B. T. \& Matthews, K. 1989, \apj, 347, 29
448: 
449: \reference{}Schlegel, D. J., Finkbeiner, D. P., \& Davis, M. 1998, \apj, 500, 525
450: 
451: \reference{}Shakura, N. I. \& Sunyaev, R. A. 1973, \aap, 24, 337
452: 
453: \reference{}Shields, G. A. 1978, \nat, 272, 706
454: 
455: \reference{}Suganuma, M., et al. 2006, \apj, 639, 46
456: 
457: \reference{}Suganuma, M., et al. 2004, \apj, 612, L113
458: 
459: \reference{}Trevese, D., \& Vagnetti, F. 2002, \apj, 564, 624
460: 
461: \reference{}Vanden Berk, D. E., et al., 2001, \aj, 122, 549
462: 
463: \reference{}Wanders, I., et al. 1997, \apjs, 113, 69
464: 
465: \reference{}Winkler, H., Glass, I. S., Van Wyk, F, Marang, F., Spencer Jones, J, H., Buckley, D. A. H., Sekiguchi, K. 1992, \mnras, 257, 659
466: 
467: \reference{}Winkler, H. 1997, \mnras, 292, 273
468: 
469: \reference{}Yoshii, Y. 2002, in New Trends in Theoretical and Observational 
470:     Cosmology, ed. K. Sato and T. Shiromizu (Tokyo: Universal Academy Press),
471:      p235
472: 
473: \reference{}Yoshii, Y., Kobayashi, Y. \& Minezaki, T. 2003, American Astronomical Society, 202, 3803
474: 
475: \end{references}
476: 
477: 
478: 
479: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
480: %   table
481: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
482: 
483: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
484: \end{document}
485: 
486: 
487: 
488: