0810.3359/ms.tex
1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %
3: % The alignment between the host and the nearest neighbor Groups
4: % Yougang Wang, Changbom Park, Xiaohu Yang, Yun-Young Choi, Xuelei Chen
5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
6: \documentclass[apj]{emulateapj}
7: %\documentclass[preprint2]{aastex}
8: \usepackage{amsfonts}
9: 
10: \newcommand{\etal}{{et al.~}}
11: \newcommand{\msunh}{\>h^{-1}\rm M_\odot}
12: \newcommand{\Msun}{\>{\rm M_\odot}}
13: \newcommand{\Lsunhh}{\,h^{-2}\rm L_\odot}
14: \newcommand{\Lsun}{\>{\rm L_{\odot}}}
15: \newcommand{\lta}{\la}
16: \newcommand{\mpch}{\>h^{-1}{\rm {Mpc}}}
17: \newcommand{\kms}{\>{\rm km}\,{\rm s}^{-1}}
18: \newcommand{\kpc}{\>{\rm kpc}}
19: \newcommand{\calC}{{\cal C}}
20: \newcommand{\rmag}{\>^{0.1}{\rm M}_r-5\log h}
21: 
22: \shorttitle{Alignment of groups}
23: 
24: \shortauthors{Wang et al.}
25: 
26: 
27: \begin{document}
28: 
29: \title{Alignments of Group Galaxies with Neighboring Groups}
30: \author{Yougang Wang\altaffilmark{1}, Changbom Park\altaffilmark{2}, Xiaohu
31:   Yang\altaffilmark{3}, Yun-Young Choi\altaffilmark{4}, Xuelei
32:   Chen\altaffilmark{1}}
33: 
34: \altaffiltext{1}{National Astronomical Observatories, Chinese Academy of
35:   Sciences, Beijing 100012, China; E-mail: wangyg@bao.ac.cn}
36: %
37: \altaffiltext{2}{Korea Institute for Advanced Study, Dongdaemun-gu, Seoul
38:   130-722, Korea; cbp@kias.re.kr}
39: %
40: \altaffiltext{3}{Key Laboratory for Research in Galaxies and Cosmology,
41:   Shanghai Astronomical Observatory, the Partner Group of MPA, Nandan Road 80,
42:   Shanghai 200030, China}
43: %
44: \altaffiltext{4}{Astrophysical Research Center for the Structure and Evolution
45:   of the Cosmos, Sejong University, Seoul 143-747, Korea}
46: 
47: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
48: 
49: \begin{abstract} Using  a sample of galaxy  groups found in  the Sloan Digital
50:   Sky Survey Data Release 4, we  measure the following four types of alignment
51:   signals: (1)  the alignment between  the distributions of the  satellites of
52:   each group  relative to the direction  of the nearest  neighbor group (NNG);
53:   (2) the alignment between the major  axis direction of the central galaxy of
54:   the host group (HG) and the  direction of the NNG; (3) the alignment between
55:   the major axes  of the central galaxies of  the HG and the NNG;  and (4) the
56:   alignment  between the  major  axes of  the  satellites of  the  HG and  the
57:   direction  of the  NNG.   We find  strong  signal of  alignment between  the
58:   satellite distribution and the orientation of central galaxy relative to the
59:   direction of the NNG, even when  the NNG is located beyond $3r_{\rm vir}$ of
60:   the host group.  The  major axis of the central galaxy of  the HG is aligned
61:   with the direction of the NNG.  The alignment signals are more prominent for
62:   groups that are  more massive and with early type  central galaxies. We also
63:   find  that there  is a  preference for  the two  major axis  of  the central
64:   galaxies of  the HG and NNG  to be parallel  for the system with  both early
65:   central galaxies,  however not for the  systems with both  late type central
66:   galaxies.  For  the orientation  of satellite galaxies,  we do not  find any
67:   significant alignment  signals relative to  the direction of the  NNG.  From
68:   these four types of alignment measurements, we conclude that the large scale
69:   environment traced  by the nearby group  affects primarily the  shape of the
70:   host dark matter halo, and  hence also affects the distribution of satellite
71:   galaxies  and the  orientation of  central galaxies.   In addition,  the NNG
72:   directly  affects the  distribution of  the satellite  galaxies  by inducing
73:   asymmetric  alignment signals.  And NNG  at very  small separation  may also
74:   contribute a second  order impacts on the orientation  of the central galaxy
75:   in the HG.
76: \end{abstract}
77: 
78: \keywords{methods: statistical-galaxies: haloes-galaxies: structure-dark
79:   matter-large scale structure of universe}
80: 
81: %111111111111111111111111111111111111111111111111111111111111111111
82: \section{Introduction}
83: 
84: The distribution of satellites in the groups of galaxies holds
85: important clues to the  assembly history  of dark matter  halos.
86: Since satellite  galaxies are typically  distributed over the entire
87: dark  matter halo,  they are  a useful tracer  of  the  dark matter
88: distribution  on  the  scale  of the  group.  In particular, their
89: position  provides information  on  the shape  of the  dark matter
90: halo (Carter \& Metcalfe 1980; Plionis, Barrow \& Frenk 1991; Fasano
91: et al. 1993; Basilakos, Plionis \& Maddox 2000; Orlov, Petrova \&
92: Martynova 2001; Plionis  et al.  2004,  2006; Bailin  \& Steinmetz
93: 2005;  Wang  et al.  2008, hereafter W08), and their kinematics
94: could be used to estimate the mass of the haloes  (e.g., Zaritsky et
95: al.  1993, 1997;  McKay et  al. 2002;  Brainerd \& Specian  2003;
96: Katgert, Biviano \& Mazure 2004;  van den  Bosch  et al.  2004; More
97: et al. 2009a, 2009b).
98: 
99: 
100: One way to characterize  quantitatively the distribution of satellite galaxies
101: is  to  measure the  alignment  between  their  spatial distribution  and  the
102: orientation of their central galaxies.  Extensive studies with high-resolution
103: simulations have  shown that sub-haloes tend  to align with the  major axis of
104: their host  halos (Knebe et  al. 2004, 2008a,  2008b; Libeskind et  al.  2005,
105: 2007;  Wang  et al.  2005;  Zentner  et al.  2005;  Kang  et  al.  2007).  The
106: observational search for  a possible alignment between the  central galaxy and
107: satellites  has a  long and  serpentine history.  The first  study of  such an
108: alignment  was performed  by Holmberg  (1969), who  found that  satellites are
109: preferentially   located    along   the   minor   axes    of   isolated   disc
110: galaxies.  Holmberg's  study  was  restricted to  projected  satellite-central
111: distances of $r_p  \lta 50 \kpc$. Subsequent studies,  however, were unable to
112: confirm this so-called ``Holmberg effect'' (Hawley \& Peebles 1975; Sharp, Lin
113: \& White 1979;  MacGillivray et al. 1982). Zaritsky et  al. (1997) studied the
114: distribution of satellites around spiral  hosts and were also unable to detect
115: any significant alignment for $r_p \lta  200 \kpc$, but they found a preferred
116: minor-axis alignment for $300 \kpc \lta  r_p \lta 500 \kpc$. The satellites of
117: our Milk Way galaxy  and the nearby M31 galaxy lie in  planes which are highly
118: inclined with respect  to their discs.  This were  noted by Lynden-Bell (1976,
119: 1982),  Majewski (1994),  Hartwick (1996,  2000)  and Kroupa,  Theis \&  Boily
120: (2005) for  the Milky Way, by Koch  \& Grebel (2006) and  McConnachie \& Irwin
121: (2006) for M31, and by Metz, Kroupa \& Jerjen (2007) for both galaxies.
122: 
123: With large redshift  surveys, such as the 2dF  Galaxy Redshift Survey (2dFGRS;
124: Colless  et  al.  2001) and  the  Sloan  Digital  Sky  Survey (SDSS;  York  et
125: al. 2000),  much larger samples  of galaxy groups  can be used  to investigate
126: this alignment problem.  Sales \& Lambas (2004; 2009) used a  set of 1498 host
127: galaxies  with  3079 satellites  from  the  2dFGRS,  and found  a  large-scale
128: alignment of the  satellites along the host major axes for  $300 \kpc \lta r_p
129: \lta 500 \kpc$.   Brainerd (2005) studied a sample  of isolated SDSS galaxies,
130: and found that the distribution of satellite galaxies is strongly aligned with
131: the major  axis of the disc host  galaxy. Yang et al.   (2006, hereafter Y06),
132: using a galaxy group catalogue similar to  the one used here, but based on the
133: SDSS Data Release  2 (DR2), studied the alignment signal as  a function of the
134: color of  the central and satellite  galaxies.  They found  that the alignment
135: strength  is strongest  between red  centrals  and red  satellites, while  the
136: satellite distribution  in systems  with a blue  central galaxy  is consistent
137: with being isotropic.  Y06 also found that the  alignment strength is stronger
138: in  more  massive haloes  and  at smaller  projected  radii  from the  central
139: galaxy. These results have  subsequently been confirmed by several independent
140: studies  (Donoso, O'Mill  \& Lambas  2006; Azzaro  et al.  2007;  Agustsson \&
141: Brainerd 2006,  2007; W08;  Steffen \& Valenzuela  2008; Bailin et  al. 2008).
142: These studies have focused on whether the satellites are distributed along the
143: major axis or minor axis of the central galaxy.
144: 
145: 
146: Besides the alignment  between the distribution of satellite  galaxies and the
147: orientation of their  central galaxy, other forms of  alignment have also been
148: studied. These  include the  alignment between neighboring  clusters (Binggeli
149: 1982; West 1989; Plionis 1994),  between brightest cluster galaxies (BCGs) and
150: their parent clusters (Carter \&  Metcalfe 1980; Binggeli 1982; Struble 1990),
151: between  the orientation  of satellite  galaxies  and the  orientation of  the
152: cluster  (Dekel 1985; Plionis  et al.  2003), and  between the  orientation of
153: satellite galaxies  and the orientation of  the BCG (Struble  1990). Using the
154: same group catalogue as the one used here, Faltenbacher et al. (2007) examined
155: three  different  types of  intrinsic  galaxy  alignment  within groups:  halo
156: alignment between  the orientation of  the brightest group galaxies  (BGG) and
157: the  distribution of  its  satellite galaxies,  radial  alignment between  the
158: orientation of a satellite galaxy and the direction toward its BGG, and direct
159: alignment between the orientation of the  BGG and that of its satellites. They
160: found  that the  orientations  of red  satellites  are preferentially  aligned
161: radially  in the direction  of the  BGG. In  addition, they  found a  weak but
162: significant  indication  that  the  orientations  of  satellite  galaxies  are
163: directly  aligned   with  that  of   their  BGG  on   scales  $r<0.1R_{\mathrm
164: vir}$. Based on a cosmological $N$-body simulation, Faltenbacher et al. (2008)
165: analyzed  the spatial  and kinematic  alignments of  satellite halos  within 5
166: times the virial  radius of group-sized host halos. They  found that the tidal
167: forces on the large scales can gives  rise to a halo alignment out to at least
168: $5R_\mathrm{vir}$. This means that  the alignment signal is strongly dependent
169: on the large-scale environment. It is also found that the orientations of dark
170: matter halos can be related to their surrounding structures, such as filaments
171: and large-scale  walls (e.g., Faltenbacher et  al. 2002; Einasto  et al. 2003;
172: Avila-Reese et al.  2005; Hopkins, Bahcall \& Bode 2005;  Kasun \& Evard 2005;
173: Basilakos et al. 2006; Altay et al. 2006; Aragon-Calvo et al. 2006; Maulbetsch
174: et al. 2007; Ragone-Figueroa \& Plionis 2007; Hahn et al.  2007a, 2007b). Paz,
175: Stasyszyn \& Padilla (2008) used  numerical simulations and the real data from
176: the SDSS  Data Release  6 (DR6)  to study the  alignments between  the angular
177: momentum of individual objects and  the large-scale structure. They found that
178: the angular momentum of dark  matter haloes are preferentially oriented in the
179: direction perpendicular to the distribution of matter, and more massive haloes
180: show a higher degree of alignment. Okumura, Jing \& Li (2009) investigated the
181: correlation  between  the  orientation  of  giant  elliptical  galaxies,  they
182: measured the intrinsic ellipticity correlation function of 83773 SDSS luminous
183: red galaxies (LRGs) and found that there is a positive alignment between pairs
184: of the LRGs up to  $~30h^{-1}$Mpc scales. Recently, Faltenbacher et al. (2009)
185: used the SDSS  DR6 and the Millennium simulation  to determinate the alignment
186: between  galaxies   and  large-scale  structure,  and  found   that  there  is
187: significant  alignment  between  the  major  axes of  red  galaxies  with  the
188: surrounding large-scale structure.
189: 
190: %In this paper, we  aim to study the effects of the  large scale environment on
191: %the various  properties of  the central and  satellite galaxies, i.e.,  on the
192: %distribution of satellite galaxies and the shapes of the central and satellite
193: %galaxies. In our approach, the large scale environment is characterized by the
194: %nearest neighbor groups (hereafter NNG).
195: 
196: In this paper, we aim to study the impacts of the nearest neighbor
197: groups (hereafter NNGs) and  possibly the large scale environments
198: beyond the NNGs (e.g., NNG has a higher probability to be
199: distributed along the direction of the  filament)  on the  various
200: properties  of  the central  and  satellite galaxies, i.e., on the
201: distribution  of satellite galaxies and the shapes of the central
202: and satellite galaxies.   Throughout this paper,  unless stated
203: otherwise, when  we refer to the  impacts of the NNG,  possible
204: impacts from the large  scale environments  are not excluded. For
205: our purposes, the following  alignment  signals are  measured:  (1)
206: the alignment between  the distributions of the satellites relative
207: to the direction of the NNG; (2) the alignment between the major
208: axis of  the central galaxy of the host group (HG) and the direction
209: of the NNG; (3)  the alignment between the two major axes of the
210: central galaxies of the HG and  the NNG; and (4) the alignment
211: between the major axes of the satellites of the  HG and the
212: direction of the NNG. Here the alignment signals (2), (3) and (4)
213: are measured to probe the impact of the NNG on the shapes of the
214: central and satellite galaxies, respectively.
215: 
216: This  paper  is  organized  as   follows.   In  Section  2,  we  describe  the
217: observational  data used for  this study.   Section 3  presents the  method to
218: quantify the alignment signal. Section 4 shows the results of various kinds of
219: alignment signals we  measured, and their dependence on  the morphology of the
220: central galaxies of  the HG and the NNG.  Section 5  summarize our results and
221: discuss various related issues.
222: 
223: 
224: 
225: \section{Observational Data Set}
226: \subsection{Groups: central and satellite galaxies}
227: 
228: The analysis  presented in this  paper is based  on the SDSS DR4  galaxy group
229: catalogue of Yang \etal (2007)\footnote{In  this paper, we refer to systems of
230:   galaxies as groups regardless of their richness, including isolated galaxies
231:   (i.e., systems with a single  member) and rich clusters of galaxies.}.  This
232: group catalogue is constructed by applying the halo-based group finder of Yang
233: \etal  (2005)  to  the   New  York  University  Value-Added  Galaxy  Catalogue
234: (NYU-VAGC;  see  Blanton  \etal  2005),   which  is  based  on  the  SDSS  DR4
235: (Adelman-McCarthy  \etal 2006). From  this catalogue  Yang \etal  selected all
236: galaxies in  the Main Galaxy Sample with  redshifts in the range  $0.01 \leq z
237: \leq 0.20$ and  with a redshift completeness greater than  0.7. This sample of
238: galaxies is  further divided  into three group  samples: sample I,  which only
239: uses the $362,356$ galaxies with  measured redshifts from the SDSS; sample II,
240: which also  includes $7091$ galaxies  with SDSS photometry but  with redshifts
241: taken from alternative  surveys; and sample III, which  includes an additional
242: $38,672$ galaxies which do not have measured redshifts due to fiber-collision,
243: but were assigned the redshift  of nearest neighbors (cf.  Zehavi \etal 2002).
244: The present  analysis is based on  the sample II, which  consists of $369,447$
245: galaxies distributed over 301,237 groups  with a sky coverage of $~4514\, {\rm
246:   deg^2}$.  Details of  the group  finder and  the general  properties  of the
247: groups can be found in Yang et al.  (2007).
248: 
249: The halo  mass of  each group is  estimated using  the ranking of  the group's
250: characteristic stellar  mass, $M_{\rm stellar}$, defined as  the total stellar
251: masses  of  all  group  members  with  absolute  magnitude  ${^{0.1}}M_r-5\log
252: h\leq-19.5$.  More details of the mass estimations can be found in Yang et al.
253: (2007).  For those  groups  with  all members  that  have absolute  magnitudes
254: ${^{0.1}}M_r-5\log h > -19.5$, which are not assigned halo masses in the group
255: catalogue, we use  the mean stellar-to-halo mass relation  obtained in Yang et
256: al. (2009) to  assign their halo masses. In this study,  only groups with halo
257: masses $M\geq  10^{11.5} \msunh$ are selected.  Note also that  in these group
258: catalogues the survey  edge effects have been taken into  account (Yang et al.
259: 2007). We use only those groups  with $f_{\rm edge} \geq 0.6$, where $1-f_{\rm
260: edge}$ is  the fraction  of galaxies  in a group  that are  missed due  to the
261: survey edges.
262: 
263: 
264: Applying  all the  above  mentioned selection  criteria  (magnitude, mass  and
265: $f_{\rm  edge}$),  we have  a  total of  27,173  central  galaxies and  64,366
266: satellite galaxies. Here the central galaxy  is defined to be the most massive
267: (in  terms of  stellar  mass) galaxy  in  each group  and  other galaxies  are
268: satellites.  These  galaxies are  used to detect  the first kind  of alignment
269: signals where position angles of galaxies are not required.
270: 
271: However,  in  order to  study  other three  kinds  of  alignment
272: signals,  the position angles of the galaxies are  required. For
273: these studies, we keep only those  galaxies with  $\emph{b/a}<0.75$
274: whose  isophotal position  alngles are well defined.   Here $a$ and
275: $b$  are the isophotal semi-major  and minor axis lengths,
276: respectively.  Thus  within our final sample with  27,173 central
277: and 64,366 satellite  galaxies, 13,890 central and 44,219  satellite
278: galaxies have well measured position angles.
279: 
280: 
281: Finally,  note  that the  selected  galaxy  groups  contain some  interlopers,
282: i.e.  false  members  assigned  to  a  group. If  the  distribution  of  these
283: interlopers is uncorrelated (or anti-correlated) with that of the true members
284: of the group,  our results on the first kind of  alignment would be negatively
285: biased.  According  to  Yang et  al.  (2007;  2005)  the average  fraction  of
286: interlopers in the group is less than 20\%. We have tested the effects of such
287: interlopers  by   assuming  that  the  distribution  of   the  interlopers  is
288: uncorrelated with the shape of the  group and is spherical, and found that the
289: presence  of the  interlopers can  decrease the  first type  of  the alignment
290: signal by $\sim 10\%$.
291: 
292: 
293: \subsection{Galaxies: early and late types, position angles}
294: 
295: In our study we  follow the prescription of Park \& Choi  (2005) to divide our
296: sample into the early (ellipticals and lenticulars ) and the late (spirals and
297: irregulars) morphological types.  This division is based on  their location in
298: the $u  - r$ versus $g  - i $ color  gradient space, and also  in the $i$-band
299: concentration index  space. The  early type galaxies  are classified  as those
300: lying above the boundary lines  passing through the points (3.5, -0.15), (2.6,
301: -0.15), and  (1.0, 0.3) in  the $u-r$ versus  $\Delta(g - i)$ space.  They are
302: also   required   to  have   the   (inverse)   concentration  index   $c\equiv
303: R_{50}/R_{90}<0.43$, where $R_{50}$ and $R_{90}$ are the radii from the center
304: of galaxy containing $50\%$ and $90\%$  of the Petrosian flux. The rest of the
305: galaxies  are   classified  as   late-type  galaxies.  The   completeness  and
306: reliability of this  classification scheme reaches 90\%.  For  more details of
307: the morphology classification, we refer the reader to Park \& Choi (2005).  In
308: this paper,  we adopt the $r$  band isophotal position angle  for each galaxy,
309: which is given in the SDSS-DR4 (Adelman-McCarthy et al. 2006). We have checked
310: the distributions of these position angles and found them to be isotropic.
311: 
312: 
313: \subsection{The nearest neighbor group}
314: 
315: A very  important step in  our investigation is  to find the  nearest neighbor
316: group, so that  we can define/estimate the direction of  the tidal force. Here
317: we  combine the  pairwise  velocity differences  and  the projected  distances
318: between the central galaxies of the HG and its neighboring groups in selecting
319: the nearest neighbor.  Note that the {\it pairwise  velocity differences} here
320: refers  to  the line-of-sight  velocity  difference  between  the two  central
321: galaxies of  the groups.   For a  given host group,  its NNG  is found  in the
322: following way.
323: 
324: 
325: 
326: We first inspect the distribution  of the pairwise velocity difference between
327: central galaxies of  the HG and the neighboring groups to  set up the velocity
328: difference criteria (e.g., Park  \& Choi 2009). Figure~\ref{fig:v_d} shows the
329: distribution of the velocity difference  $\Delta v$ for the early (filled dot)
330: and late (open circle) type centrals. These profiles are obtained by measuring
331: the velocity  difference $\Delta v$  distribution of all central  galaxy pairs
332: with the projected  distance $r_p$ between 0.4 and $1  \mpch$. Note that these
333: profiles  are contributed  by two  components: (1)  the  randomly distributed,
334: un-correlated pairs (constant component);  and (2) the correlated pairs (e.g.,
335: by neighbor  groups; enhanced  component). Since the  impact of  pairwise {\it
336: peculiar}  velocity  of  galaxies  is  to  broaden  the  distribution  of  the
337: correlated pairs, we may use the  measured profile to probe the {\it up-limit}
338: of the pairwise  peculiar velocity assuming that the  correlated pairs in real
339: space have a very compact distribution.
340: 
341: \begin{figure}[tbp]
342: %\resizebox{\hsize}{!}{\includegraphics{figure1.ps}}
343: \includegraphics[width=0.45\textwidth]{figure1.ps}
344: \caption{Probability  distribution  of  the  velocity  difference  $\Delta  v$
345:   between early  (filled dot)  or late (open  circle) type centrals  and other
346:   central  galaxies at  projected separation  $r_p=0.4\sim1 \mpch$.  The solid
347:   line is the best fit curve for the early (E/S0) morphological type centrals,
348:   while the dashed one is for the late (S/Irr) type centrals.}
349:   \label{fig:v_d}
350: \end{figure}
351: 
352: 
353: According   to  the  measured   $\Delta  v$   distribution  shown   in  Figure
354: \ref{fig:v_d},  we  fit to  the  data using  an  exponential  function plus  a
355: constant:
356: %
357: \begin{equation}
358: f(\Delta v) = f_1 {\rm exp}(-\Delta v/\sigma_{\Delta v})+f_2,
359: \end{equation}
360: %
361: where $f_1, f_2$ and $\sigma_{\Delta v}$ are fitting parameters.  The best fit
362: characteristic ({\it  up-limit}) velocity differences  $\sigma_{\Delta v}$ are
363: 342  and 243  km/s for  early and  late type  centrals, respectively,  and the
364: fitting curves  are shown  as the  solid (early type)  and dashed  (late type)
365: lines in Figure~\ref{fig:v_d}.
366: 
367: 
368: 
369: 
370: We  use the following  criteria to  select the  NNG: (1)  for a
371: given central galaxy  of a HG,  if the  central galaxy  of the
372: neighbor group  has velocity difference  less than  800  (for early
373: type  central) or  600 (for  late-type central)  km/s (about 2.4
374: times the  characteristic velocity  difference) and projected
375: separation $r_p  < 3\mpch$, the neighbor with  the smallest $r_p$ is
376: set  as the  NNG;  (2)  if there  is  no central  galaxy  within the
377: velocity difference or  $r_p$ limit, the  neighbor with the smallest
378: three dimensional separation $s$ is set to as the NNG, since this
379: group has a higher probability of being the true NNG in real space
380: than the other ones with smaller $r_p$ but beyond the velocity
381: difference limit.  These velocity values (800 km/s and 600 km/s,
382: roughly 2.4 times the  characteristic velocity difference) are
383: chosen as a  reasonable compromise  between obtaining  a large
384: sample and  reducing the fraction  of  interlopers in  the  sample.
385: Since  the line-of-sight  pairwise peculiar velocity of galaxies  is
386: typically 300 km/s (see Figure~\ref{fig:v_d}), it is very unlikely
387: for redshift  distortion to make real space close pairs to be
388: separated in  redshift space with velocity difference  much larger
389: than 300 km/s,  so the three  dimensional distance  in redshift
390: space between  the two central galaxies can be used as the distance
391: indicator. On the other hand, the central galaxy pairs with velocity
392: difference less than 800 km/s (early types) or 600 km/s (late types)
393: and $r_p$ less than $3\mpch $  are modeled using the projected
394: separation $r_p$ as the distance indicator. However, we have checked
395: varying  these velocity  differences  (800 km/s  for early type  and
396: 600  for late-type) and  found that our results  are not sensitive
397: to  the exact values used.  Finally,  note that  one can  also
398: define  the  NNG  according to  the projected separation relative to
399: the size  of the host group (e.g., its virial radius). However, we
400: have made  tests  and again  found that  using such  an alternative
401: definition of the NNG will  not have any significant impact on our
402: results.
403: 
404: 
405: \section{Quantifying the alignment}
406: %
407: \begin{figure*}
408: \begin{center}
409:   \includegraphics[scale=0.72,angle=270] {figure2.ps}\caption{Illustration of
410:     the first three alignment angles $\theta_1$, $\theta_2$ (left panel) and
411:     $\theta_3$ (right panel).} \label{fig:ang}
412: \end{center}
413: \end{figure*}
414: %
415: 
416: To study the impacts  of the NNG on the various properties  of the central and
417: satellite galaxies,  we first  measure the different  alignment signals  as we
418: listed in Section 1.  The alignments  of objects are obtained by computing the
419: distribution  functions  of  the   alignment  angles  (e.g.,  Brainerd  2005),
420: $P(\theta_i)  (i=1,2,3,4)$, where  $\theta_i$  is the  angle  between the  two
421: directions.
422: 
423: The angle  $\theta_1$ is the projected  angle between the  line
424: connecting the central galaxy  to the  satellite galaxy and  the
425: line connecting  the central galaxy to  the NNG  (See the  left
426: panel of  Figure \ref{fig:ang}).  The angle $\theta_1$  is
427: constrained  in   the  range  $0^{\degr}  \leq  \theta_1  \leq
428: 180^{\degr}$, where $\theta_1 <  90^{\degr} (>90^{\degr})$ implies a
429: satellite is at the near (far) side of the HG with respect to the
430: NNG. We also define an angle
431: \begin{equation}
432: \tilde{\theta_1} \equiv \left\{
433: \begin{array}{ll}
434: \theta_1, & \theta_1 < 90^{\degr}\\
435: 180^{\degr} - \theta_1, & \theta_1 >90^{\degr}
436: \end{array}\right.
437: \end{equation}
438: The range of $\tilde{\theta_1}$ is $0^{\degr} \leq \tilde{\theta_1}
439: \leq 90^{\degr} $, which is more useful when making average.
440: 
441: The angle $\theta_2$ is the angle between the major axis of the
442: central galaxy of  the HG  and  the  direction of  the  NNG (See the
443: left  panel of  Figure \ref{fig:ang}),  which is  constrained to  be
444: in  the  range $0^{\degr}  \leq \theta_2 \leq90^{\degr}$. $\theta_2
445: = 0^{\degr}  (90^{\degr})$ suggests that the major (minor) axis of
446: the central  galaxy of the HG  is perfectly aligned with the
447: direction of the NNG.
448: 
449: The angle  $\theta_3$ is the angle between  the two major axes  of
450: the central galaxies of the  HG and the NNG. As shown  in
451: Figure~\ref{fp_sep}, the angular separations  between HG  and  NNG
452: on  the  celestial sphere  can  reach up  to $2^{\degr}$ for some
453: pairs, so for  the sake of accuracy, we have included the curvature
454: effect in the determination  of $\theta_3$ by parallel transport the
455: angles  along  great  circles  on  the  celestial  sphere.  The
456: procedure  is illustrated in  the right panel  of
457: Figure~\ref{fig:ang}, and explained  in more details in  the
458: Appendix. Note however that  neglecting this effect  does not have
459: any significant  impact on  our results.  The angle  $\theta_3$  is
460: also constrained in the range $0^{\degr} \leq \theta_3 \leq
461: 90^{\degr}$.
462: 
463: \begin{figure}[htb]
464: %  \resizebox{\hsize}{!}{\includegraphics{figure3.ps}}
465: \includegraphics[width=0.45\textwidth]{figure3.ps}
466: \caption{The distribution of angular separations  between HGs and NNGs for our
467:   sample.  } \label{fp_sep}
468: \end{figure}
469: 
470: 
471: 
472: The angle $\theta_4$  is similar to $\theta_2$ but defined  for the major axes
473: of  the satellite  galaxies, i.e.,  the angle  between the  major axis  of the
474: satellite galaxy of the HG and the direction of the NNG.
475: 
476: 
477: In the following we will take  $\theta_1$ as an example to explain the process
478: of     measuring    the    alignment     signals,    the     measurement    of
479: $\theta_2,\theta_3,\theta_4$ are similar.
480: 
481: For  a  given  set  of  the  HG-NNG  pairs,  we  first  count  the  number  of
482: satellite-central-NNG  pairs, $N(\theta_1)$,  that have  the  angle $\theta_1$
483: between  the central-satellite direction  and the  direction of  the NNG  in a
484: number of $\theta_1$ bins.  Next, we  construct 100 random samples in which we
485: randomize  the  positions of  all  NNGs,  and  compute $\langle  N_R(\theta_1)
486: \rangle$, the  average number of satellite-central-NNG pairs  for the randomly
487: located NNGs as a function  of $\theta_1$. The random samples constructed this
488: way  suffer exactly  the same  selection effects  as the  real sample,  so any
489: significant  difference between $N(\theta_1)$  and $N_R(\theta_1)$  reflects a
490: genuine alignment between the distribution of the satellite galaxies in the HG
491: and the direction of the NNG.
492: 
493: Following  Y06  and  W08,  we  quantify  the alignment  signal  by  using  the
494: distribution of normalized pair counts:
495: %
496: \begin{equation}\label{eq:fpairs}
497: f_{\rm pairs}(\theta_1)=\frac{N(\theta_1)}{\langle N_R(\theta_1)\rangle}.
498: \end{equation}
499: %
500: In the  absence of any  alignment, $f_{\rm pairs}(\theta_1)=1$,  while $f_{\rm
501: pairs}(\theta_1) >  1$ near  $\theta =  0$ or 180  degree implies  a satellite
502: distribution preferentially aligned along the direction of the NNG.
503: 
504: We  may quantify  the fluctuation  using $\sigma_{\rm  R}(\theta_1)
505: / \langle N_{\rm R}(\theta_1) \rangle$, where $\sigma_{\rm R}$ is
506: the standard deviation of $N_{\rm R}(\theta)$, which could  be
507: estimated from the 100 random samples. In addition to  this
508: normalized pair count, we also  compute the average angle $\langle
509: \theta_1    \rangle$.    In   the   absence    of   any   alignment
510: $\langle\tilde{\theta_1}  \rangle   =  45^{\degr}$.  If   one  finds
511: $\langle \tilde{\theta_1}  \rangle <45^{\degr}$  ($\langle
512: \tilde{\theta_1} \rangle  > 45^{\degr}$), it means that the
513: satellites are parallel (perpendicular) to the line connecting the
514: host-neighbor pairs.
515: 
516: 
517: 
518: \section{Results}
519: 
520: \subsection{Satellite galaxy distribution relative to the direction of the
521:   NNG}\label{sec:theta1}
522: 
523: As have  been found in recent  papers, the satellite  galaxies are distributed
524: preferentially along the major axis  of the central galaxies, especially those
525: with red  central galaxies (e.g.,  Brainerd 2005; Agustsson \&  Brainerd 2006;
526: Y06;  Azzaro  et al.  2007;  Kang  et al.  2007;  W08).  This  shows that  the
527: distribution  of the  satellite  galaxies  is not  completely  random, but  is
528: correlated  with the shapes  of the  central galaxies.  In this  subsection we
529: check whether similar correlations exist beyond the single dark matter halo of
530: the group.   In order to describe  the deviation of the  alignment signal from
531: the null, a $\chi^2$ test is applied here
532: %
533: \begin{equation}\label{eq:chi2}
534: \chi^2=\sum_{i=1}^{N_\mathrm{bin}}\frac{(f_\mathrm{pairs}-1.0)^2}{\sigma_\mathrm
535: {R}^2}
536: \end{equation}
537: %
538: where $N_\mathrm{bin}$ denotes the bin number of the angular distribution.
539: 
540: 
541: \begin{figure}[htb]
542: %  \resizebox{\hsize}{!}{\includegraphics{figure4.ps}}
543: \includegraphics[width=0.49\textwidth]{figure4.ps}
544: \caption{The  normalized probability  distribution for  $\theta_1$,  the angle
545:   between the directions of the satellite galaxies in the HG and the NNG.  The
546:   left  panel shows  results  for the  whole  sample (filled  circle) and  the
547:   subsample with projected distance smaller than $3 r_{vir}$ of the host group
548:   (open circle).  The right panel shows  results for the  subsamples where the
549:   mass of the  HG are larger (filled diamond) and  smaller (open diamond) than
550:   that of the NNG.  The open symbols are shifted slightly along the horizontal
551:   axis for clarity.}
552: \label{fp_ns1}
553: \end{figure}
554: 
555: 
556: 
557: The filled  circles in  the left-hand panel  of Figure~\ref{fp_ns1}
558: shows the distribution of  $f(\theta_1)$ for all  satellite-NNG
559: pairs in our  SDSS group catalogue.   One  can see  that  $f_{\rm
560: pairs}(\theta_1)>1$  at small  (near $0\degr$) and large $\theta_1$
561: (near $180\degr$) values, while it is less than 1 at middle values
562: (near $90\degr$).  From the figure, we see clearly that the
563: distribution of HG  satellites are not completely uniform  or
564: isotropic, there is a  small (in absolute strength)  but highly
565: significant  preference for the direction along  the line connecting
566: HG and  NNG. The deviation  from uniform distribution has
567: $\chi^2=97.39$, corresponding  to  $\rm{CL}>99.99\%$ for  8 degree
568: of freedom  (hereafter dof) . The average  value of $
569: \tilde{\theta_1}$ is $\langle \tilde{\theta_1}\rangle =
570: 43.8^{\degr}\pm0.1^{\degr}$, which again shows that the distribution
571: of satellites  of HG are not isotropic or uniform, but slightly
572: prefers the direction of the NNG.  The effect is small but highly
573: significant  ($\sim10\sigma$). For  comparison, the  open circle  in
574: the left panel  of  Figure~\ref{fp_ns1}  shows   $f_{\rm  pairs}
575: (\theta_1)$  for  the host-neighbor pairs  with the projected
576: distance  smaller than 3  times of the virial radius of the HG.
577: There is no significant difference between the filled circle  and
578: open  circle lines,  the confidence  of this  difference  is below
579: $0.5\sigma$ level.  {\it In other words,  not only the NNGs but also
580: the large
581:   scale environments, e.g. the filaments,  represented by the NNGs that affect
582:   (or at least  play an important role in) the  distribution of the satellites
583:   (and  the mass  distribution within  the  host halo).}   Otherwise we  would
584: expect distance dependent signals.
585: 
586: 
587: 
588: Note that in the above discussion, the  NNG can be either more
589: massive or less massive than  the HG,  it may  be interesting to
590: investigate these  two cases separately. For  this purpose, in the
591: right-hand  panel of Figure~\ref{fp_ns1} we show  the resulting
592: $\theta_1$ distributions for  HG-NNG systems  with the mass of HG
593: larger (filled diamond) and smaller (open diamond) than that of the
594: NNG. In  the rest part  of the paper,  unless stated otherwise, we
595: use filled circle symbol to represent the sample which is not
596: constrained by the distance limit  while the  open circles  to
597: represent  the results  with  the projected distance smaller than 3
598: times the  virial radius of the HG; the filled diamond (open
599: diamond) symbol represents the  results for subsamples where the
600: mass of the HG are larger (smaller) than  that of the NNG. The
601: stronger alignment signals shown  in the right-hand panel in
602: Figure~\ref{fp_ns1} for satellite galaxies in more massive HGs may
603: caused by the fact that their distributions are flatter  than those
604: in smaller HGs  (e.g., Jing  \& Suto 2002;  Yang et al. 2006).
605: 
606: 
607: 
608: \begin{figure}[tb]
609: %  \resizebox{\hsize}{!}{\includegraphics{figure5.ps}}
610: \includegraphics[width=0.5\textwidth]{figure5.ps}
611: \caption{Similar   to  Fig.~\ref{fp_ns1},  but   here  for   different  HG-NNG
612:   systems. Left panels: results for HG-NNG systems where the mass of the HG is
613:   larger than that of the NNG.  Right panels: results for HG-NNG systems where
614:   the mass of the HG is smaller  than that of the NNG.  The asterisk, triangle
615:   and  square symbols represent  the mass  of the  NNG in  the range  $M_n \in
616:   [11.5,12.5]$,           $[12.5,13.5]$           and           $[13.5,14.5]$,
617:   respectively. } \label{fp_mass_alig1}
618: \end{figure}
619: 
620: We  can further  check  this by  measuring  the alignment  signals
621: for  HG-NNG systems that  are divided  into subsamples of  different
622: masses.  In  the left panel of  Figure~\ref{fp_mass_alig1}, we show
623: the alignment  signal of $f_{\rm pairs}(\theta_1)$ for the submaples
624: with mass of the HG  that are larger than that of the NNG, while in
625: the right  panel of the Figure~\ref{fp_mass_alig1}, we show  the
626: results of the  subsamples with the  mass of the HG smaller than
627: that of  the NNG.  In  each panel, the  asterisk, triangle and
628: square symbols represent the mass  of the NNG in the range $M_n
629: \equiv \log_{10} (h M/\Msun) \in [11.5,12.5]$, $[12.5,13.5]$ and
630: $[13.5,14.5]$, respectively. We found some interesting    trends
631: here.    According    to    the    left panel    of
632: Figure~\ref{fp_mass_alig1},  for  the most  massive groups  ($M_n
633: >13.5$),  the alignment  signal (i.e.   deviation of $f_{\rm
634: pairs}(\theta_1)$  from  1) is strongest ($5\sigma$), perhaps
635: indicating the  very strong attraction  by the NNG  and the flatter
636: distribution of  satellite galaxies  within the  HG. For system with
637: less massive HG and  NNGs ($M_n \in [12.5,13.5]$),  the signal is
638: weaker(deviation from  null signal hypothesis by $3\sigma$),
639: However, for the smallest groups  discussed here ($M_n \in
640: [11.5,12.5]$),  the alignment signal is again rather  strong
641: (deviation from null signal  hypothesis by 5$\sigma$). This  is also
642: evident from  the  value  of  $\chi^2$, 58.22,  32.31,  40.60,
643: corresponding  to the  mass of  the NNG  in the  range $M_n  \in
644: [13.5,14.5]$, [12.5,13.5] and  [11.5,12.5], respectively.
645: Interestingly,  for those systems with NNG  heavier than HG  and NNG
646: in  the range $M_n\in  [11.5,12.5]$ (bottom right panel), there are
647: slightly  more satellite galaxies distributed near the NNG (with
648: $\theta_1 < 90\degr$) than HG (the confidence level is $1.5\sigma$),
649: which may  indicate that  the NNG near  a small  HG may attract  its
650: satellite galaxies and affect their distribution.
651: 
652: \begin{figure}[tb]
653: %  \resizebox{\hsize}{!}{\includegraphics{figure6.ps}}
654: \includegraphics[width=0.5\textwidth]{figure6.ps}
655: \caption{Same as
656:     Fig.~\ref{fp_ns1}, but for different subsamples, divided by the type of
657:     the central galaxies of the HG or NNG.} \label{fp_ns2}
658: \end{figure}
659: 
660: 
661: \begin{figure}[tb]
662: %  \resizebox{\hsize}{!}{\includegraphics{figure7.ps}}
663: \includegraphics[width=0.5\textwidth]{figure7.ps}
664: \caption{Same as
665:     Fig.~\ref{fp_ns2}, except that here we split the sample according to the
666:     morphological types of both the central galaxies of the HG and
667:     NNG. } \label{fp_ns3}
668: \end{figure}
669: 
670: 
671: In order to study how this alignment depends on the morphological types of the
672: central galaxies of the  HG and the NNG, following Y06 and  W08, we divide our
673: sample  of host  and  neighbor groups  into  different morphology  subsamples.
674: Figure~\ref{fp_ns2} shows the  alignment signals $f_{\rm pairs}(\theta_1)$ for
675: HG  with early (lower-left  panel) and  late-type (lower-right  panel) central
676: galaxy,  and NNG  with  early (upper-left  panel)  and late-type  (upper-right
677: panel) central galaxy, respectively.
678: 
679: The alignment  signal does  seem to  depend on the  morphological type  of the
680: central galaxy slightly:  the groups with early type  central galaxies or with
681: the NNGs having an early  central galaxy show slightly stronger alignment than
682: those with late-type  centrals ($1.0\sigma$ for the early  centrals of HG than
683: late  centrals of  HG, $1.7\sigma$  for the  early centrals  of NNG  than late
684: centrals  of NNG).  In  Figure~\ref{fp_ns3}, we  show  the alignment,  $f_{\rm
685: pairs}(\theta_1)$,  for four  combinations  of  the HG  and  NNG with  central
686: galaxies of  different morphological  types.  As one  can see, pairs  with the
687: late-type HG  and late-type  NNG show the  smallest strength of  the alignment
688: signal.
689: 
690: The HG satellite distribution is  also slightly asymmetric with
691: respect to the near  side or  far  side  from the  NNG.   There is a
692: small but  significant preference for  the satellite  galaxies in a
693: group with  a late  type central galaxy to be distributed  near the
694: side of the NNG, with  either an early or a late type central
695: galaxy.  This asymmetry is significant by $\sim 2\sigma$ for both
696: the whole  sample and  and for  the case  where the  NNG  with
697: projected separation is smaller than 3 times  of the virial radius.
698: For the samples with the  early-type HG  and late-type  NNG,
699: however,  the trend  is  opposite: the satellite galaxies  are
700: preferentially  distributed on the  far side  from the NNG, again
701: this effect is  small but significant ($2.4\sigma$ for all samples,
702: $2.7\sigma$ for the close pairs). The  reason of this is not
703: completely clear, but it  may be due  to (i) the  groups with late
704: type central galaxy  is more probably located in the outskirts of  a
705: high density region with groups having preferentially early type
706: central  galaxies; (ii) probably smaller than groups with early type
707: central galaxies, thus have smaller impact on its NNG.  Before we
708: proceed, it is quite interesting to check whether or not the weak
709: asymmetry is induced by those groups near the survey edges. As we
710: have tested using only groups with  $f_{\rm edge}>0.9$ by performing
711: exactly the same  analysis, the weak asymmetry is  almost the same.
712: Moreover, one would  only expect that the groups near the edge may
713: slightly induce the weak  asymmetry in the direction of the {\it
714: near} side of the NNG,  however not in the {\it far}  side of the
715: NNG, as shown in the lower-right  panel of Figure~\ref{fp_ns3}.
716: Therefore, we believe that our results (e.g., the weak asymmetry)
717: are robust against the impact of groups near the survey edges.
718: 
719: 
720: 
721: \begin{figure}[tbp]
722: %  \resizebox{\hsize}{!}{\includegraphics{figure10.ps}}
723: \includegraphics[width=0.5\textwidth]{figure8.ps}
724: \caption{The average value of $\theta_1$ as a function of the projected
725:   distance between the HG and the NNG.  Left: as a function of $r_{vir, n}$;
726:   Right: as a function of $r_{vir, h}$.} \label{fp_r1}
727: \end{figure}
728: 
729: 
730: To what  extent does  the large  scale environment as  represented by  the NNG
731: affect the  distribution of the satellite  galaxies in the HG?   To study this
732: problem,  we  divided  the  total  sample into  subsamples  according  to  the
733: projected distance between  the HG and the NNG.   Figure~\ref{fp_r1} shows the
734: dependence of the  average value $\theta_1$ on the  projected distance between
735: the HG and the  NNG.  Here we take both the virial  radius of the NNG, $r_{\rm
736:   vir, n}$(left  panel), and the  HG, $r_{\rm vir,  h}$ (right panel),  as the
737: unit. There is  no significant difference between the results  in the left and
738: right panels.  The  value of $\langle \theta_1 \rangle$  depends weakly on the
739: distance between  the HG  and NNG. The  alignment between the  distribution of
740: satellite galaxies and  the direction of the NNG  is significant (greater than
741: $1\sigma$) up to  separations as large as about  12$r_{\rm vir, n}$(or $r_{\rm
742:   vir,h}$), which is also evident from the fact that $\langle \tilde{\theta_1}
743: \rangle  = 43.1\degr  \pm 0.5\degr$  (or $43.6\degr  \pm 0.6\degr$ )
744: for the systems with  separation of the  HG and the  NNG being
745: 12$r_{\rm vir,  n}$ (or $r_{\rm vir, h}$). From these alignment
746: signals for $\theta_1$ at so large separations, we conclude that the
747: distributions of the satellite galaxies in HGs are (also) affected
748: by the large scale environments.
749: 
750: 
751: \subsection{The position angle of the central galaxy}
752: 
753: \subsubsection{relative to the direction of the NNG}
754: 
755: 
756: Both Y06  and W08 found that  there is a  preference for the satellites  to be
757: distributed along the major axis of  their central galaxy, and in the previous
758: section, we found a prominent  alignment signal between the orientation of the
759: satellite system of the HG and the direction of the NNG.  Therefore, we expect
760: there is also an alignment between the major axis of the central galaxy of the
761: HG and the direction of the NNG.
762: 
763: 
764: 
765: \begin{figure}[tbp]
766: %\resizebox{\hsize}{!}{\includegraphics{figure11.ps}}
767: \includegraphics[width=0.5\textwidth]{figure9.ps}
768: \caption{The  normalized probability distribution,  $f_{\rm pairs}(\theta_2)$,
769:   of the angle $\theta_2$ between the  major axes of the central galaxy of the
770:   HG and the direction of the NNG. In the left panel, results are measured for
771:   all samples (filled  circle) and the subsamples with  the projected distance
772:   smaller than 3  times the virial radius of the host  group (open circle). In
773:   the right panel, we show results for subsamples where the mass of the HG are
774:   larger (smaller)  than that  of the NNG  with filled diamond  (open diamond)
775:   symbol. Formal rejection confidence levels  from the $\chi^2$ test are shown
776:   in each panel. The open circles and open diamonds have been shifted slightly
777:   along the horizontal axis for clarity.} \label{fp_chn1}
778: \end{figure}
779: 
780: 
781: Indeed, as shown in the left  panel of Figure~\ref{fp_chn1}, there
782: is a strong signal  of alignment  between  these two  directions:
783: the major  axis of  the central galaxy of  the HG is preferentially
784: aligned with  the direction of the NNG.   The  alignment signal  is
785: even stronger  for  the  pairs with  smaller projected   separations
786: as   shown   by   the   open   circle   symbols   of
787: Figure~\ref{fp_chn1}.  This result suggests that the major axis of
788: the central galaxy  of the  HG tends  to be  aligned  with the
789: direction of  the NNG  (or possibly the large scale structure
790: beyond).
791: 
792: 
793: The orientation of the central galaxy of HG is possibly affected by
794: (1)the potential of the HG; (2)NNG and (3) the large scale
795: environment. To understand the relative influence of these factors,
796: we checked separately the signals for all pairs of NNG and the close
797: pairs ($\le 3r_{vir}$), and find the following average values:
798: $\langle \theta_2 \rangle = 43.9\degr \pm 0.2\degr$ ($\langle
799: \theta_2 \rangle = 42.7\degr \pm  0.4\degr$ for the close pairs),
800: which depend quite significantly on  the pair separations. Thus it
801: is unlikely for the large scale environment to be main factor in
802: determining the orientation of the central galaxy, otherwise one
803: would expect $\theta_2$ to be insensitive to the distance. On the
804: other hand, in W08 it was found that the orientation of the central
805: galaxy is also aligned with the potential of the HG (provided that
806: the satellite  distributions trace  the  mass distribution of the
807: halo reasonably well). Therefore, we conclude both the
808:  the satellite distribution (or the mass
809: distribution) in the  HG and  the NNG at small  separation can
810: affect the orientation  of the central galaxy. However, from the
811: alignment signal measured for $\theta_1$, we find that the
812: distribution of the satellite is affected by both the NNG and the
813: large scale of environment (see the last part of section 4.1), so
814: there may be indirect correlation between the orientation of the
815: central galaxy and the large scale environment.
816: 
817: In  the right-hand   panel  of Figure~\ref{fp_chn1}  we show the
818: resulting $\theta_2$ distributions  for HG-NNG systems with the mass
819: of HG larger (filled diamond) and smaller (open diamond) than that
820: of the NNG.   The position angle of the central galaxy in more
821: massive host group is  slightly more aligned (by $1.4\sigma$) which
822: is very likely caused by its flatter distribution of the satellite
823: galaxies.
824: 
825: 
826: \begin{figure}[tbp]
827: %\resizebox{\hsize}{!}{\includegraphics{figure12.ps}}
828: \includegraphics[width=0.5\textwidth]{figure10.ps}
829: \caption{Similar  to   Fig.~\ref{fp_chn1},  but  here   for  different  HG-NNG
830:   systems. Left panels: results for HG-NNG systems where the mass of the HG is
831:   larger than that of the NNG.  Right panels: results for HG-NNG systems where
832:   the mass  of the  HG is smaller  than that  of the NNG.  In each  panel, the
833:   asterisk, triangle  and square points represent  the mass of the  NNG in the
834:   range    $M_n   \in    [11.5,12.5]$,   $[12.5,13.5]$    and   $[13.5,14.5]$,
835:   respectively. }
836: \label{fp_mass_alig2}
837: \end{figure}
838: 
839: %To check on what mass scale such an alignment signal reaches its
840: %maximum, we divide our sample into subsamples of three mass bins as
841: %in Section \ref{sec:theta1}. As shown in Figure~\ref{fp_mass_alig2},
842: %the alignment signal in the systems with the more massive HG is
843: %slightly stronger than that in the system with the lower mass of the
844: %HG for the sample with the small mass of the NNG. \textbf{The
845: %significance is about $2.0\sigma$($0.6\sigma$),
846: %$3.5\sigma$($0.0\sigma$) and $3.3\sigma$ ($4.0\sigma$) for the
847: %systems where the mass of the HG is larger (smaller) than that of
848: %the NNG and the mass of the NNG is in the range $M_n \in
849: %[11.5,12.5]$, $[12.5,13.5]$ and $[13.5,14.5]$, respectively. For the
850: %systems with the mass of the NNG between $10^{11.5}$ and
851: %$10^{12.5}\msunh$ (and HG is even less massive), the strength of the
852: %alignment signal is no longer significant relative to the null
853: %hypothesis $\langle \theta_2 \rangle = 45^{\degr}$ at $1\sigma$
854: %level for our sample.}
855: 
856: To examine the  dependence of the alignment signal  on the group
857: mass, we divide our  sample into three subsamples according to the
858: NNG mass as in Section \ref{sec:theta1}.  Figure~\ref{fp_mass_alig2}
859: shows the distribution of  $\theta_2$   for  systems   with  the NNG
860: mass  in  the   range  $M_n \in[13.5,14.5]$, $[12.5,13.5]$, and
861: $[11.5,12.5]$ from the top to the bottom panel.  The panels on the
862: left are  for the systems having a HG more massive than  its  NNG,
863: and  those  on  the  right  are  for  the  opposite  cases.
864: Figure~\ref{fp_mass_alig2}  clearly  shows  that  the  alignment
865: signal  is stronger  for more  massive systems.   The statistical
866: significance  of the alignment is $3.3\sigma,  3.5\sigma$, and
867: $2.0\sigma$ from top  to bottom in the left column.  It is
868: $4.0\sigma, 0.0\sigma$, and $0.6\sigma$ in the same order in the
869: right column. The top panels tell that the groups in a pair are
870: aligned  with  each  other  when  the  NNG  mass  exceeds
871: $10^{13.5}\msunh$ regardless of  the HG  mass. Such  alignment can
872: occur  by the  strong tidal force of the  NNG on the central galaxy
873: of the HG. This signal  due to NNGs disappears when the  NNG mass is
874: less than $10^{13.5}\msunh$  as can be seen in the middle  right and
875: bottom right panels.  Instead, the alignment signal is seen when the
876: HG mass is higher than the NNG  mass, which means that the NNG is
877: now aligned,  regardless of its  mass, along  the major axis  of the
878: central galaxy of the HG.
879: 
880: In  Figure~\ref{fp_chn2},  we  examine how  $f_{\rm
881: pairs}(\theta_2)$ depends on  the morphological  type of the central
882: galaxies. Note  that the early  types are  mainly ellipticals,  and
883: the  elongation of  an early-type galaxy can be due to  external
884: gravitational effects or internal kinematics, both of which are
885: closely  correlated with the distribution of the satellite galaxies
886: and the NNGs. On the other hand, the late types are dominantly disk
887: galaxies  and the  position angle  of  a disk  galaxy is  determined
888: by  the direction of  its spin axis. Therefore, for HG central
889: galaxies with different morphological types, the alignment with
890: respect to the NNG could have different physical origins. In the
891: bottom two panels of  Fig.~\ref{fp_chn2}, we show the  alignment
892: signals $f(\theta_2)$ for HGs  with early-type centrals. The left
893: panel is for the HGs  whose NNG has an early-type central while the
894: right is for those whose NNG contains a late-type  central.  The HGs
895: having an early-type central show  quite a  strong alignment signal,
896: particularly when the central galaxy of the NNG is also an early
897: type (a $7\sigma$ effect for  a sample of  all such  pairs and
898: $6\sigma$ for  close pairs). If the central galaxy of the NNG is a
899: late type one, the alignment signal weakens considerably, the
900: deviation from the case of no alignment is only $1.6\sigma$ and
901: $2.7\sigma$ for all pairs and close pairs, respectively (see the
902: lower-right panel of Figure~\ref{fp_chn2}).
903: 
904: 
905: 
906: 
907: 
908: % For the HGs having a late-type central
909: %galaxy we detect no statistically significant alignment signal
910: %regardless of the morphology of the central galaxy of their NNG.
911: 
912: 
913: 
914: 
915: 
916: \begin{figure}[tb]
917: %\resizebox{\hsize}{!}{\includegraphics{figure13.ps}}
918: \includegraphics[width=0.5\textwidth]{figure11.ps}
919: \caption{Same as Figure~\ref{fp_chn1}, but for different subsamples of central
920:   galaxies of the host and neighbor groups. } \label{fp_chn2}
921: \end{figure}
922: 
923: 
924: %In Figure~\ref{fp_chn2}, we examine how $f_{\rm pairs}(\theta_2)$
925: %depends on the morphological types of the central galaxies. Note
926: %that the position angles of the early types which are dominantly
927: %ellipticals can represent quite well the orientation of the stellar
928: %distribution of the central galaxies, while the position angles of
929: %the late types which are dominantly spirals are sensitive to the
930: %inclination of the galaxies, \textbf{thus the primary signification
931: %of the alignment signal for the early and late type galaxies is
932: %different.} In the bottom two panels of Fig.~\ref{fp_chn2}, we show
933: %the alignment signals $f(\theta_2)$ for the early type central
934: %galaxy of the HG (Lower Left: NNG with early type central; Lower
935: %Right: NNG with late type central). There is a quite strong
936: %alignment signal between the position angle of the early type
937: %central galaxy in the HG with the direction of the NNG, especially
938: %if the central galaxy of the NNG is also an early type one
939: %($7\sigma$ for all sample and $6\sigma$  for close pairs). If the
940: %central galaxy of the NNG is a late type one, the alignment signal
941: %weakens considerably, the deviation from the case of no alignment is
942: %only $1.6\sigma$ ($2.7\sigma$).
943: 
944: 
945: \begin{figure}[tbp]
946: %  \resizebox{\hsize}{!}{\includegraphics{figure14.ps}}
947: \includegraphics[width=0.5\textwidth]{figure12.ps}
948: \caption{The average value of $\theta_2$ as a function of the projected
949:   distance between the HG and the NNG. Left: as a function of $r_{vir, n}$;
950:   Right: as a function of $r_{vir, h}$.  } \label{fp_r2}
951: \end{figure}
952: 
953: In the case where the central galaxy of  the NNG is a late type one,
954: the major axis of  the galaxy is not  so well correlated with  its
955: stellar distribution, but more with the  inclination of the disk. On
956: the other  hand, the minor axis is quite well correlated with the
957: disk spin axis. Thus $\theta_2=0$ means that the  spin  axis   is
958: perpendicular  to  the  direction   of  the  NNG,  while
959: $\theta_2=90^{\degr}$ means  that the spin axis  tends to be aligned
960: with the direction of the NNG. In the upper panels of
961: Figure~\ref{fp_chn2}, we show the distribution $f(\theta_2)$ for the
962: late type central galaxies (Upper Left: NNG with early type central;
963: Upper Right: NNG with late type central). However, we do not find
964: significant alignment signal between the major axis of the central
965: late type galaxy and the direction of  the NNGs. As we have also
966: measured the  average alignment  angle $\langle  \theta_2 \rangle$
967: as a  function of distance  between the  HG and  NNG  as shown  in
968: Fig.~\ref{fp_r2},  $\langle \theta_2  \rangle$ increases  slowly and
969: approaches  the  null position  of $45\degr$ as the distance
970: increases.  The results shown in Fig.~\ref{fp_r2} indicate  that
971: only  NNGs at  separation  $\la 5r_{vir}$  may have  possible
972: impacts onto the orientations of the central galaxies in the HGs.
973: 
974: 
975: 
976: \subsubsection{relative to the orientation of the central galaxy of the NNG}
977: 
978: 
979: \begin{figure}[tb]
980: %\resizebox{\hsize}{!}{\includegraphics{figure15.ps}}
981: \includegraphics[width=0.5\textwidth]{figure13.ps}
982: \caption{The normalized probability distribution, $f_{\rm pairs}(\theta_3)$,
983:   of the angle $\theta_3$ between the two major axes of the central galaxies
984:   of the HG and the NNG. Upper panels: the central galaxies of the HG and NNG
985:   are both late types. Lower panels: the central galaxies of the HG and NNG
986:   are both early types. In the two left panels we show results for the whole
987:   sample (filled circle) and the subsample with the projected distance smaller
988:   than 3 times the virial radius of the host group (open circle). In the two
989:   right panels, we show the results for subsamples where the mass of the HG
990:   are large (smaller) than that of the NNG with filled diamond (open diamond)
991:   symbol.} \label{fp_nh1}
992: \end{figure}
993: 
994: 
995: 
996: In Figure~\ref{fp_nh1},  we present the  distribution of the  angle
997: $\theta_3$ between the major axes of the central  galaxy of the HG
998: and the central galaxy of the  NNG. In  the lower panels  we use
999: only the HG  and NNG  whose central galaxies are both early types.
1000: One can  see that the major axes of the central galaxies of the HG
1001: and the NNG tend to be parallel, though the signal is weak. We also
1002: checked  the case of alignment between all  types of central
1003: galaxies, and find there is no prominent  signal, as shown by the
1004: fact $\langle \theta_3 \rangle  =  44.6\degr \pm  0.3\degr$
1005: (deviates from  the  null  case by  only $1.3\sigma $).  However,
1006: the $\theta_3$ alignment signal for  the case of the HG more massive
1007: than the NNG is stronger than the opposite case by $1.8\sigma$ (see
1008: the  lower-right panel). Combined  with our probe  of the alignment
1009: signals  for  $\theta_1$  and  $\theta_2$,  we find  that  the large
1010: scale environments tend to impact the  distribution of satellite
1011: galaxies (and the mass distribution) in the HGs,  while the
1012: distribution of satellite galaxies and possibly  NNGs at  small
1013: separations may  impact the orientations  of the (early-type)
1014: central galaxies  in the HGs. Thus the  alignment signals shown in
1015: $\theta_3$ for early type galaxies are expected.
1016: 
1017: 
1018: The upper panels of  Figure~\ref{fp_nh1} shows the distributions of $\theta_3$
1019: when the  morphological types of  the central galaxies  of the HG and  NNG are
1020: both  late  types.  The  spin  axis  of the  two  galaxies  are  parallel  for
1021: $\theta_3=0\degr$  and perpendicular  if  $\theta_3 =90\degr$.  We find  large
1022: fluctuations in the distribution, and there is no significant alignment signal
1023: between the  spin axes of  the late-type central galaxies,  regardless whether
1024: the HG is more massive or less than the NNG.
1025: 
1026: 
1027: 
1028: 
1029: 
1030: \subsection{The position angle of the satellite galaxy relative to the
1031:   direction of the NNG}
1032: 
1033: 
1034: Many  studies  have attempted  to  detect  the  alignment signal
1035: between  the orientations of  central galaxies  and satellite
1036: galaxies  in the  clusters of galaxies  (e.g., Plionis et  al. 2003;
1037: Strazzullo et  al.  2005;  Torlina, De Propris \& West 2007), most
1038: of them found only null or weak signal. Using a sample of 4289
1039: host-satellites pairs from the SDSS  DR4, Agustsson \& Brainerd
1040: (2006) found a weak signal of  the alignment. Adopting the same
1041: group  catalogue as that used here,  Faltenbacher et  al.  (2007)
1042: searched  for the alignment between  the orientations of the BGG and
1043: the satellites. They considered the total sample and submaples with
1044: different color of  the satellite galaxy, and found a small but
1045: definite  alignment signal between the  major axes of the  central
1046: and the satellite galaxies  of   the  host   group,  especially at
1047: small  scales $r_p<0.1r_{\rm vir,h}$, where  one expects a strong
1048: tidal  force at such small separation.
1049: 
1050: 
1051: \begin{figure}[tb]
1052: %  \resizebox{\hsize}{!}{\includegraphics{figure16.ps}}
1053: \includegraphics[width=0.5\textwidth]{figure14.ps}
1054: \caption{The  normalized probability distribution,  $f_{\rm pairs}(\theta_4)$,
1055:   of the angle $\theta_4$ between  the orientations of the satellite galaxy in
1056:   HG and the  direction linking the host and NNG.  In  the left panel, results
1057:   are measured  for the whole sample  (solid line) and the  subsample with the
1058:   distance  limit,  3  times the  virial  radius  of  the host  group  (dotted
1059:   line). In the right panel, we show results for the subsamples where the mass
1060:   of the HG are large (smaller)  than that of the NNG with dashed (dot-dashed)
1061:   line.} \label{fp_sat1}
1062: \end{figure}
1063: 
1064: \begin{figure}[htb]
1065: %  \resizebox{\hsize}{!}{\includegraphics{figure17.ps}}
1066: \includegraphics[width=0.5\textwidth]{figure15.ps}
1067: \caption{Same  as  Figure~\ref{fp_sat1},   but  for  subsamples  of  different
1068:   morphological  types   of  central  galaxies   of  the  host   and  neighbor
1069:   groups. } \label{fp_sat2}
1070: \end{figure}
1071: 
1072: 
1073: As we have already noticed, the NNG and the large scale environment can affect
1074: the distribution of  satellite galaxies in the HG.  Here, we  check if the NNG
1075: (and  the large  scale environment)  can also  impact the  orientation  of the
1076: satellite galaxies.   The method to  obtain the alignment angle  $\theta_4$ is
1077: similar to that for the angle  $\theta_3$, the only difference is that here we
1078: use  the major axes  of the  satellites to  substitute the  major axis  of the
1079: central galaxy of the HG.   Fig.~\ref{fp_sat1} shows the alignment between the
1080: major axes of the satellites in the HGs and the direction linking the host and
1081: NNGs. There is apparently no alignment between the major axes of the satellite
1082: and the  direction linking the  host and the NNG.   Fig~\ref{fp_sat2} displays
1083: the dependence on  the $f_{\rm pairs}(\theta_4)$ on the  morphological type of
1084: the central galaxies  of the HGs and NNGs. Again we  do not find statistically
1085: significant alignment signal for the morphology subsamples.  Moreover, we also
1086: checked the  distribution $f_{\rm  pairs}(\theta_4)$ separately for  the early
1087: and late type satellite galaxies, but there is no significant alignment signal
1088: either.
1089: 
1090: 
1091: \section{Summary and discussion}
1092: 
1093: In the cold dark matter scenario,  small dark matter halos form first and grow
1094: subsequently  to larger  structures via  accretion and  merger  processes. The
1095: accretion  of  material  might  preferentially  occur  along  the  filamentary
1096: structure  (West  1994), which  leads  to  correlations  between the  internal
1097: structures of  the neighboring groups.  Binggeli (1982)  pioneered the studies
1098: of the alignment between neighboring  clusters of galaxies, and found that the
1099: host clusters tend to be  aligned with their nearest neighbors. The subsequent
1100: studies  confirmed  this   tendency,  although  conflicting  results  appeared
1101: occasionally in the  literature (West 1989; Plionis 1994).   In this paper, we
1102: used  the host-neighboring  group systems  extracted from  the SDSS  DR4 group
1103: catalogue  (Yang  et  al.  2007)  to  probe  the  impact  of the  large  scale
1104: environment  (as  represented  by   the  nearest  neighboring  group)  on  the
1105: distribution of the  satellite galaxies and on the  orientation of the central
1106: and satellite galaxies.  For this purpose, four types of alignment signals are
1107: measured and the main results are summarized as follows.
1108: 
1109: \begin{enumerate}
1110: \item  There is  a strong  alignment signal  between the  distribution  of the
1111:   satellites  relative to  the direction  of the  NNG. This  signal  is rather
1112:   insensitive  to  the separation  between  the HG  and  NNG,  and extends  to
1113:   separation beyond  $12r_{vir}$ of  the HG.  For  the system with  both early
1114:   central galaxies of  the HG and NNG, the alignment  signal is the strongest,
1115:   for the  system with both  late central galaxies,  the signal is  weaker but
1116:   still quite significant.
1117: 
1118: 
1119: \item  The  satellite  galaxies  in  the  HG have  a  weak  preference  to  be
1120:   distributed at the  near side of the NNG with an  early type central galaxy,
1121:   and at the far side of the NNG with a late type central galaxy.
1122: 
1123: \item The  major axis  of the  central galaxy of  the HG  is aligned  with the
1124:   direction  of  the NNG,  especially  in the  massive  HGs.   This effect  is
1125:   stronger for  the systems when the central  galaxies of HG and  NNG are both
1126:   early types.  And  we find this alignment signal only  exists between HG and
1127:   NNG pairs at separation $\la 5r_{vir}$ of the HG.
1128: 
1129: \item The  distribution of the satellite  galaxies and the  orientation of the
1130:   central galaxy  of the HG  show stronger alignment  signals with the  NNG in
1131:   systems with more massive HGs and NNGs.
1132: 
1133: \item There is a preference for the  two major axes of the central galaxies of
1134:   the HG  and NNG to be  parallel for the  system with the both  early central
1135:   galaxies, while there  is no evident correlation between  the two major axes
1136:   of the central galaxies for the systems with both late type cental galaxies.
1137: 
1138: \item Although the distribution of the satellites of the HG is correlated with
1139:   the  direction of  the NNG,  their  orientations (position  angles) are  not
1140:   correlated with the direction of the NNG.
1141: \end{enumerate}
1142: 
1143: 
1144: According to our various alignment measurements, we find that the distribution
1145: of  satellite galaxies and  the orientation  of the  central galaxy  both show
1146: strong alignment signals with respect  to the direction of the NNG ($\theta_1$
1147: and $\theta_2$).  Such  alignment signals are stronger in  massive halos where
1148: central  galaxies  are  early  type  ones.   Because of  these  two  kinds  of
1149: alignments, the alignment  between the orientations of the  cetral galaxies of
1150: the HG and  the NNG ($\theta_3$) is naturally  expected.  For the orientations
1151: of satellite  galaxies, however, we  do not find significant  alignment signal
1152: relative  to the  direction of  NNG  ($\theta_4$), while  Faltenbacher et  al.
1153: (2007) have found prominent alignment  signals between the orientations of the
1154: central and  satellite galaxies at  very small separation.  This  may indicate
1155: that the orientations of satellite  galaxies are only strongly affected by the
1156: tidal force of the central galaxy  and the host halo, but not significantly by
1157: the NNG or the large scale environment.
1158: 
1159: Possible explanations for the strong alignment signals in the first and second
1160: types of the alignments  are: (i) the NNG affects the shape  of the host halo,
1161: while  the distribution of  satellite galaxies  and the  shape of  the central
1162: galaxy are mainly affected by the  host halo; (ii) the large scale environment
1163: directly affects the  distribution of satellite galaxies and  the shape of the
1164: central galaxy.  Judging from the first alignment signal, it seems likely that
1165: the impact of large scale environment is  mainly on the shape of the host halo
1166: instead of  directly on  the distribution of  the satellite  galaxies, because
1167: otherwise we should expect to find  a strong dependence on the distance to the
1168: NNG, rather than a mass dependence of  the HG and NNG.  The fact that stronger
1169: alignment signals is  found for the subsamples with more  massive HGs and NNGs
1170: is likely  correlated with the  fact that the  more massive halos  have larger
1171: triaxialities (e.g, Jing \& Suto 2002; Wang \& Fan 2004; W08).
1172: 
1173: \begin{figure}[tbp]
1174: %  \resizebox{\hsize}{!}{\includegraphics{figure18.ps}}
1175: \includegraphics[width=0.5\textwidth]{figure16.ps}
1176: \caption{The  normalized probability distribution,  $f_{\rm pairs}(\theta_1)$,
1177:   of the angle $\theta_1$ for  the systems with early-type host and early-type
1178:   neighbor.  The asterisk  and triangle symbols represent the  results for the
1179:   submaples  with $0^{\degr}<\theta_2  < 45^{\degr}$  and $45^{\degr}<\theta_2
1180:   <90^{\degr}$, respectively.  The triangle symbols have been shifted slightly
1181:   along the horizontal axis for clarity.  } \label{fp_alig1}
1182: \end{figure}
1183: 
1184: 
1185: In  order  to  check  whether  the  shape of  the  central  galaxy
1186: is  mainly determined/affected by its  own host halo, or by the  NNG
1187: at small separation, we  carried out  an  additional  test.  In
1188: Fig.~\ref{fp_alig1},  we show  the alignment signal $f_{\rm
1189: pairs}(\theta_1)$ for the systems with the early-type host and
1190: early-type neighbor, where  the asterisk and triangle symbols show
1191: the results for  the system with  $0^{\degr}<\theta_2 < 45^{\degr}$
1192: and $45^{\degr}<\theta_2  <90^{\degr}$, respectively. When
1193: $45^{\degr}<\theta_2 <90^{\degr}$,  the major axis  of the central
1194: galaxy of  the  HG is  rather perpendicular to the direction of the
1195: NNG. If the shape  of central galaxy is not much affected by the
1196: host  halo, but only  affected by the NNG,  we would expect that the
1197: alignment  signals of  the two  subsamples are similar. The results,
1198: however,  show that  the alignment signals are opposite for the two
1199: different subsamples (the difference level is $7.5\sigma$). On the
1200: other hand, the signals are precisely what one would expect if the
1201: mass and light are reasonably  well-aligned (i.e., the  image of
1202: early type  central galaxy and the surrounding mass  of the HG) and
1203: the satellites in  the HG trace the surrounding mass. That is, one
1204: would get a  peak at $\theta_1 = 90^{\degr}$ if $45^{\degr} <
1205: \theta_2 < 90^{\degr}$ and a valley at $\theta_1 = 90^{\degr}$ if
1206: $0^{\degr} < \theta_2 < 45^{\degr}$.   Hence, the shape and
1207: orientation of  the central galaxy  is more likely  to  be
1208: determined  by  its  own host  halo.   While  the  different
1209: separation   dependencies   of   $\langle   \theta_2   \rangle$
1210: shown   in Figure~\ref{fp_r2}  and $\langle \theta_1  \rangle$ shown
1211: in  Figure~\ref{fp_r1} indicate  that  NNG at  small  separation may
1212: also  play,  however, only  a secondary  impact on  the orientation
1213: of the  central galaxy  of the  HG at particular conditions, e.g,
1214: interaction with the host halo (e.g., Lin et al. 2003; Ludlow et al.
1215: 2009). This conclusion is also strengthened by the recent
1216: measurements which  show that the orientations of  the central
1217: galaxies are preferentially  aligned with the distribution of
1218: satellite galaxies (e.g. Y06 and references therein).
1219: 
1220: Finally, we draw our conclusion that the large scale environment traced by the
1221: NNG have  impacts on the shape (orientation)  of the host halo.   On the other
1222: hand, the  distribution of the satellite  galaxies, the shapes  of the central
1223: galaxies, and the shape of the satellite galaxies at small radii, however, are
1224: mainly affected by their own host  halos.  Apart from these, the NNG also have
1225: direct impacts  on the  distribution of satellite  galaxies, which  produce an
1226: asymmetric alignment signal  with respective to the near or  far away sides of
1227: the NNG.  And NNG at small separations may also have small secondary impact on
1228: the orientation of the central galaxy of the HG.
1229: 
1230: 
1231: 
1232: 
1233: 
1234: 
1235: 
1236: %%%%%%%%%%%%%%%%%
1237: % Ackowledgements
1238: %%%%%%%%%%%%%%%%%
1239: 
1240: \section*{Acknowledgments}
1241: 
1242: We  thank  the  referee  for   the  constructive  and  detailed
1243: comments  and suggestions. This  work has started during  YGW's
1244: visit to KIAS,  and he would like to express  his gratitude for
1245: KIAS.  CBP and  YYC acknowledge the support of  the   Korea  Science
1246: and  Engineering  Foundation   (KOSEF)  through  the Astrophysical
1247: Research  Center for the  Structure and Evolution of  the Cosmos
1248: (ARCSEC).  XY  acknowledges the support  by the Shanghai Pujiang
1249: Program (No. 07pj14102),  973 Program  (No.   2007CB815402), the CAS
1250: Knowledge  Innovation Program  (Grant  No.  KJCX2-YW-T05)  and
1251: grants  from  NSFC (Nos.   10533030, 10673023, 10821302). YGW  and
1252: XLC acknowledges the support  of the 973 program (No.2007CB815401),
1253: the   CAS  Knowledge   Innovation   Program  (Grant   No.
1254: KJCX3-SYW-N2), and the NSFC grant 10503010. YGW is also supported by
1255: the Young Researcher Grant of National Astronomical Observatories,
1256: Chinese Academy of Sciences. XLC is also supported by the NSFC
1257: Distinguished Young Scholar Grant No.10525314.
1258: 
1259: 
1260: Funding for  the SDSS  and SDSS-II has  been provided  by the Alfred  P. Sloan
1261: Foundation, the  Participating Institutions, the  National Science Foundation,
1262: the   U.S.  Department  of   Energy,  the   National  Aeronautics   and  Space
1263: Administration, the  Japanese Monbukagakusho, the Max Planck  Society, and the
1264: Higher  Education  Funding  Council  for   England.   The  SDSS  Web  Site  is
1265: http://www.sdss.org/.
1266: 
1267: The  SDSS  is  managed  by  the  Astrophysical  Research  Consortium  for  the
1268: Participating  Institutions. The Participating  Institutions are  the American
1269: Museum  of Natural  History,  Astrophysical Institute  Potsdam, University  of
1270: Basel, Cambridge  University, Case  Western Reserve University,  University of
1271: Chicago, Drexel  University, Fermilab, the  Institute for Advanced  Study, the
1272: Japan Participation  Group, Johns Hopkins University, the  Joint Institute for
1273: Nuclear  Astrophysics,  the  Kavli  Institute for  Particle  Astrophysics  and
1274: Cosmology,  the  Korean  Scientist  Group,  the Chinese  Academy  of  Sciences
1275: (LAMOST),  Los  Alamos   National  Laboratory,  the  Max-Planck-Institute  for
1276: Astronomy (MPIA), the Max-Planck-Institute  for Astrophysics (MPA), New Mexico
1277: State University, Ohio State  University, University of Pittsburgh, University
1278: of Portsmouth, Princeton University,  the United States Naval Observatory, and
1279: the University of Washington.
1280: 
1281: 
1282: \begin{thebibliography}{}
1283: \bibitem[]{}Adelman-McCarthy, J. K., et al. 2006, \apjs, 162, 38
1284: \bibitem[]{}Agustsson, I., \& Brainerd, T. G. 2006, \apj, 644, L25
1285: \bibitem[]{}Agustsson, I., \& Brainerd, T. G. 2007, preprint (arXiv:0704.3441)
1286: \bibitem[]{}Altay, G., Colberg, J. M., \& Croft, R. A. C. 2006, \mnras, 370, 1422
1287: \bibitem[]{}Aragon-Calvo, M., van de Weygaert, R., \& van der Hulst, T. 2006, Proc.
1288:             Bernard¡¯s Cosmic Stories: From Primordial Fluctuations to the Birth
1289:             of Stars and Galaxies, Meeting Abstracts. Uimp, Valencia, Spain, p. 85.1
1290: \bibitem[]{}Avila-Reese, V., {et al.} 2005, \apj, 634, 51
1291: \bibitem[]{}Azzaro, M., Patiri, S. G., Prada, F., \& Zentner, A. R.
1292:             2007, \mnras, 376, L43
1293: \bibitem[]{}Bailin, J., \& Steinmetz, M. 2005, \apj, 627, 647
1294: \bibitem[]{}Bailin, J., Power, C., Norberg, P., Zaritsky, D., \& Gibson, B. K.
1295:             2008, \mnras, 390, 1133
1296: \bibitem[]{}Basilakos, S., Plionis, M., \& Maddox, S. J. 2000, \mnras, 316, 779
1297: \bibitem[]{}Basilakos, S., Plionis, M., Yepes, G., Gottl$\rm\ddot{o}$ber, S., \& Turchaninov, V.
1298:             2006, \mnras, 365, 539
1299: \bibitem[]{}Binggeli, B. 1982, \aap, 107, 338
1300: \bibitem[]{}Blanton, M. R., et al. 2005, \aj, 129, 2562
1301: \bibitem[]{}Brainerd, T. G. 2005, \apj, 628, L101
1302: \bibitem[]{}Brainerd, T. G., \& Specian, M. A. 2003, \apj, 593, L7
1303: \bibitem[]{}Carter, D., \& Metcalfe, N. 1980, \mnras, 191, 325
1304: %\bibitem[]{}Ciotti, L., \& Dutta, S. N. 1994, \mnras, 270, 390
1305: \bibitem[]{}Colless, M., et al. (The 2dFGRS Team), 2001, \mnras, 328, 1039
1306: \bibitem[]{}Dekel, A. 1985, \apj, 298, 461
1307: \bibitem[]{}Donoso, E., O'Mill, A., \& Lambas, D. G. 2006, \mnras, 369, 479
1308: \bibitem[]{}Einasto, M., Einasto, J., M$\rm\ddot{u}$ller, V., Hein$\rm\ddot{a}$m$\rm\ddot{a}$ki, P., \&
1309:             Tucker, D. L. 2003, \aap, 401, 851
1310: \bibitem[]{}Faltenbacher, A., Gottl$ \rm\ddot{o}$ber, S.,  Kerscher, M., \&
1311:             M$\rm\ddot{u}$ller, V. 2002, \aap, 395, 1
1312: \bibitem[]{}Faltenbacher, A., Li, C., Mao, S., van der Bosch, F. C., Yang, X.,
1313:             Jing, Y. P., Pasquali, A., \& Mo, H. J. 2007, \apj, 662, L71
1314: \bibitem[]{}Faltenbacher, A., Jing, Y. P., Li, C., Mao, S., Mo, H. J., Pasquali, A., \& van den Bosch, F. C.
1315:             2008, \apj, 675, 146
1316: \bibitem[]{}Faltenbacher, A., Li, C., White, S. D. M., Jing, Y. P., Mao, S., \& Wang, J.
1317:             2009, RAA, 9, 41
1318: \bibitem[]{}Fasano, G., Pisani, A., Vio, R., \& Girardi, M. 1993, \apj, 416, 546
1319: %\bibitem[]{}Fleck, J. -J., \& Kuhn, J. R. 2003, \apj, 592, 147
1320: \bibitem[]{}Hahn, O., Porciani, C., Carollo, C. M., \& Dekel, A. 2007a, \mnras, 375, 489
1321: \bibitem[]{}Hahn, O., Carollo, C. M., Porciani, C., \& Dekel, A. 2007b, \mnras, 381, 41
1322: \bibitem[]{}Hartwick, F. D. A. 1996, in Morrison H., Sarajedini A., eds, ASP
1323:             Conf. Ser. Vol. 92, Formation of the Galactic Halo . . . Inside and
1324:             Out. Astron. Soc. Pac., San Franciso, p. 444
1325: \bibitem[]{}Hartwick, F. D. A. 2000, \aj, 119, 2248
1326: \bibitem[]{}Hawley, D. L., \& Peebles, P. J. E. 1975, \aj, 80, 477
1327: \bibitem[]{}Holmberg, E. 1969, Ark. Astron., 5, 305
1328: \bibitem[]{}Hopkins, P. F., Bahcall, N. A., \& Bode, P. 2005, \apj, 618, 1
1329: \bibitem[]{}Jing, Y. P., \& Suto, Y. 2002, \apj, 574, 538
1330: \bibitem[]{}Kang, X., van den Bosch, F. C., Yang, X., Mao, S., Mo, H. J., Li, C., \& Jing, Y.
1331:             P. 2007, \mnras, 378, 1531
1332: \bibitem[]{}Katgert, P., Biviano, A., \& Mazure, A. 2004, \apj, 600, 657
1333: \bibitem[]{}Kasun, S. F., \& Evrard, A. E. 2005, \apj, 629, 781
1334: \bibitem[]{}Knebe, A., Gill, S. P. D., Gibson, B. K., Lewis, G. F., Ibata, R. A.,
1335:             \& Dopita, M. A. 2004, \apj, 603, 7
1336: \bibitem[]{}Knebe, A., Draganova, N., Power, C., Yepes, G., Hoffman, Y., Gottl$\ddot{o}$ber, S., \& Gibson, B. K.
1337:             2008a, \mnras, 386, L52
1338: \bibitem[]{}Knebe, A., Yahagi, H., Kase, H., Lewis, G., \& Gibson, B. K.
1339:             2008b, \mnras, 388, L34
1340: \bibitem[]{}Koch, A., \& Grebel, E. K. 2006, \aj, 131, 1405
1341: \bibitem[]{}Kroupa, P., Theis, C., \& Boily, C. M. 2005, \aap, 431, 517
1342: \bibitem[]{}Libeskind, N. I., Frenk, C. S., Cole, S., Helly, J. C., Jenkins, A.,
1343:            Navarro, J. F., \& Power, C. 2005, \mnras , 363, 146
1344: \bibitem[]{}Libeskind, N. I., Cole, S., Frenk, C. S., Okamoto, T., \& Jenkins,
1345:   A. 2007, \mnras, 374, 16
1346: \bibitem[]{}Lin, W. P., Jing, Y. P., \& Lin, L. 2003, \mnras, 344, 1327
1347: \bibitem[]{}Ludlow, A. D., Navarro, J. F., Springel, V., Jenkins,
1348:   A., Frenk, C. S., \& Helmi, A. 2009, ApJ, 692, 931L
1349: \bibitem[]{}Lynden-Bell, D. 1976, \mnras, 174, 695
1350: \bibitem[]{}Lynden-Bell, D. 1982, Observatory, 102, 202
1351: \bibitem[]{}MacGillivray, H. T., Dodd, R. J., McNally, B. V., \& Corwin H. G.
1352:             Jr. 1982, \mnras, 198, 605
1353: \bibitem[]{}Majewski, S. R. 1994, \apj, 431, L17
1354: \bibitem[]{}Maulbetsch, C., Avila-Reese, V., Colin, P., Gottl$\rm\ddot{o}$ber, S.,
1355:             Khalatyan, A., \& Steinmetz, M. 2007, \apj, 654, 53
1356: \bibitem[]{}McConnachie, A. W., \& Irwin, M. J. 2006, \mnras, 365, 902
1357: \bibitem[]{}McKay, T., et al. 2002, \apj, 571, L85
1358: \bibitem[]{}Metz, M., Kroupa, P., \& Jerjen, H. 2007, \mnras, 374, 1125
1359: \bibitem[]{}More, S., van den Bosch, F. C., \& Cacciato, M. 2009a,
1360:             \mnras, 392, 917
1361: \bibitem[]{}More, S., van den Bosch, F. C., Cacciato, M., Mo, H. J., Yang, X., \& Li,
1362:             R. 2009b, \mnras, 392, 801
1363: \bibitem[]{}Okumura, T., Jing, Y. P., \& Li, C. 2009, \apj, 694, 214
1364: \bibitem[]{}Orlov, V. V., Petrova, A. V., \& Tarantaev, V. G. 2001, \mnras, 325, 133
1365: \bibitem[]{}Park, C., \& Choi, Y. -Y. 2005, \apj, 635, L29
1366: \bibitem[]{}Park, C., \& Choi, Y. -Y. 2009, \apj, 691, 1828
1367: \bibitem[]{}Paz, D. J., Stasyszyn, F., \& Padilla, N. D. 2008,
1368:             \mnras, 389, 1127
1369: \bibitem[]{}Plionis, M. 1994, \apjs, 95, 401
1370: \bibitem[]{}Plionis, M., Barrow, J. D., \& Frenk, C. S. 1991, \mnras, 249, 662
1371: \bibitem[]{}Plionis, M., Basilakos, S., \& Tovmassian, H. M. 2004, \mnras, 352, 1323
1372: \bibitem[]{}Plionis, M., Benoist, C., Maurogordato, S., Ferrari, C., \& Basilakos, S. 2003,
1373:             \apj, 594, 144
1374: \bibitem[]{}Plionis, M., Basilakos, S., \& Ragone-Figueroa, C. 2006, \apj, 650, 770
1375: \bibitem[]{}Ragone-Figueroa, C., \& Plionis, M. 2007, \mnras, 377, 1785
1376: \bibitem[]{}Sales, L., \& Lambas, D. G. 2004, \mnras, 348, 1236
1377: \bibitem[]{}Sales, L., \& Lambas, D. G. 2009, \mnras, 395, 1184
1378: \bibitem[]{}Sharp, N. A., Lin, D. N. C., \& White, S. D. M. 1979, \mnras, 187, 287
1379: \bibitem[]{}Steffen, J. H., \& Valenzuela, O. 2008, \mnras, 387, 1199
1380: \bibitem[]{}Strazzullo, V., Paolillo, M., Longo, G., Puddu, E., Djorgovski, S. G., De Carvalho, R. R.,
1381:             \& Gal, R. R. 2005, \mnras, 359, 191
1382: \bibitem[]{}Struble, M. F. 1990, \aj, 99, 743
1383: \bibitem[]{}Torlina, L., De Propris, R., \& West, M. J. 2007, \apj,
1384:             660, L97
1385: %\bibitem[]{}Usami, M., \& Fujimoto, M. 1997, \apj, 487, 489
1386: \bibitem[]{}van den Bosch, F. C., Norberg, P., Mo, H. J., \& Yang,
1387:             X. 2004, \mnras, 352, 1302
1388: \bibitem[]{}Wang, H. Y., Jing, Y. P., Mao, S., \& Kang, X. 2005,
1389:             \mnras, 364, 424
1390: \bibitem[]{}Wang, Y. -G., \& Fan, Z. -H. 2004, \apj, 617, 847
1391: \bibitem[]{}Wang, Y., Yang, X., Mo, H. J., Li, C., van den Bosch, F. C., Fan, Z., \& Chen,
1392:             X. 2008, \mnras, 385, 1511 (W08)
1393: \bibitem[]{}West, M. J. 1989, \apj, 344, 535
1394: \bibitem[]{}West, M. J. 1994, \mnras, 268, 79
1395: \bibitem[]{}Yang, X., Mo, H. J., van den Bosch, F. C., \& Jing, Y. P. 2005, \mnras,
1396:             356, 1293
1397: \bibitem[]{}Yang, X., van der Bosch, F. C., Mo, H. J., Mao, S., Kang, X., Weinmann, S. M.,
1398:             Guo, Y., \& Jing, Y. P. 2006, \mnras, 369, 1293 (Y06)
1399: 
1400: \bibitem[]{}Yang, X., Mo, H. J., van den Bosch, F. C., Pasquali, A., Li, C., \&
1401:             Barden, M. 2007, \apj, 671, 153
1402: \bibitem[]{}Yang, X., Mo, H. J., \& van den Bosch, F. C. 2009, \apj, 695, 900
1403: \bibitem[]{}York, D. G., et al. 2000, \aj, 120, 1579
1404: \bibitem[]{}Zaritsky, D., Smith, R., Frenk, C. S., \& White, S. D. M. 1993, \apj, 405, 464
1405: \bibitem[]{}Zaritsky, D., Smith, R., Frenk, C. S., \& White, S. D. M. 1997, \apj, 478, 39
1406: \bibitem[]{}Zehavi, I., et al. 2002, \apj, 571, 172
1407: \bibitem[]{}Zentner, A. R., Kravtsov, A. V., Gnedin, O. Y., \& Klypin, A. A. 2005,
1408:             \apj, 629, 219
1409: 
1410: \end{thebibliography}
1411: 
1412: \appendix
1413: %\section{Appendix}
1414: 
1415: In this appendix, we explain how  to measure the angle between the
1416: orientation of the central galaxy of the host group and NN, taking
1417: into account the effect of  curvature  of  the  sky.   We  take  one
1418: pair  of  the  host  $(\alpha_h, \delta_h,\Phi_h)$  and neighbor
1419: $(\alpha_n,  \delta_n,\Phi_n)$ as  an example, where  $\alpha_{h}$,
1420: $\delta_h$  and  $\Phi_h$  are   the  right  ascension, declination,
1421: and  the position  angle of the  host centrals  respectively, and
1422: $\alpha_{n}$, $\delta_n$  and $\delta_n$  are the corresponding
1423: parameters of the NN. Here we assume that $\alpha_n>\alpha_h$
1424: $\delta_h>0$, and $\delta_n>0$ (See the  right panel of
1425: Figure~\ref{fig:ang}  ). Extending the  major axis of the central
1426: galaxy of  the NN, it will cross the longitude  of the host at the
1427: point  B. Now  in the  spherical triangle  $\bigtriangleup ABC$,  we
1428: have two angles $\angle BAC$  and $\angle ACB$ and one  side
1429: $\widehat{AC}$. Therefore, we  can  get  the  angle  $\angle  CBA$
1430: by  solving  the  spherical  triangle $\bigtriangleup ABC$.  It  is
1431: clear that the angle  between the orientation of the  central galaxy
1432: of the  host group  and NN  can be  written as  $\Phi_h -
1433: (180^{\degr}  - \angle CBA)$,  which is  equivalent to  the angle
1434: $\theta_3 = \theta_{3a}  - \theta_{3b}$  (see  the right  panel of
1435: Figure~\ref{fig:ang}). Position angles of the major axes are with
1436: respect to the east direction.
1437: 
1438: 
1439: \end{document}
1440: