THE CHESHIRE CAT GRAVITATIONAL LENS: THE FORMATION OF A MASSIVE FOSSIL GROUP

Strong gravitational lensing of background galaxies by foreground groups or clusters of galaxies is a powerful tool for determining the mass of the intervening system. Unlike other mass determination methods that require the assumption of hydrostatic equilibrium of the cluster gas or dynamical assumptions about the constituent galaxies, the lensing method is insensitive to the dynamical state of the lensing matter, and is dependent only on the amount of mass between the source and the observer. In general, lensing mass estimates agree reasonably well with the hydrostatic equilibrium method, at least for relaxed clusters (e.g., Wu 2000; Rzepecki et al. 2007; Halkola et al. 2008). Although X-ray/lensing joint analyses have been used mostly for massive clusters, for which the lensing signal is stronger, its application for groups of galaxies has become more robust (e.g., Grant et al. 2004; Fassnacht et al. 2008), providing a new window to study group dynamics.

As the strength of lensing depends on the mass concentration of the lens, lens searches should be most adept at finding lensing systems which, at a given mass, have the highest mass concentrations, such as fossil groups of galaxies. Fossil groups are classically defined as systems dominated by a single giant elliptical galaxy for which there is at least a two magnitude difference between that galaxy and the second rank galaxy (in r-band) within 0.5 r₂₀₀ (where r₂₀₀ approximates the virial radius where the average mass density inside this radius is 200 times the critical density of the universe), and are bright sources of extended X-ray emission (${L}_{X,\mathrm{bol}}\gt {10}^{42}\;{h}_{50}^{-2}$ erg s⁻¹; Jones et al. 2003). Given the lack of bright galaxies other than the central brightest group galaxy (BGG), the correlation found between concentration and formation epoch in N-body simulations (Wechsler et al. 2002), and the fact that fossil groups seem to have higher central mass concentrations than non-fossil groups (e.g., Khosroshahi et al. 2007; c.f., Sun et al. 2009), it was suggested that fossil groups formed earlier than normal groups, with most of the larger galaxies merging into a single central dominant elliptical galaxy, with the system then remaining undisturbed for a very long time until the present day (Ponman et al. 1994; Vikhlinin et al. 1999; Jones et al. 2000). In this picture, age and a lack of subsequent interaction with another group would be the key differences between a fossil group and a normal group.

However, more recently, X-ray and optical measurements of fossil groups have shown several potential inconsistencies with this formation mechanism. Fossil groups appear to resemble clusters of galaxies more closely than groups of galaxies despite their low galaxy count (e.g., Proctor et al. 2011). The most massive fossil groups have a hot intergalactic medium similar to that of clusters, sometimes in excess of 4 keV (e.g., Khosroshahi et al. 2006). Measurements of galaxy velocity dispersions in fossil groups (Cypriano et al. 2006; Mendes de Oliveira et al. 2006; Miller et al. 2012) seem to be consistent with the measured T_X, at least for the few fossil groups with relatively good X-ray data. This is further suggested by their (not atypical) location in the ${L}_{X}-{T}_{X}$ relation (e.g., Khosroshahi et al. 2007; Miller et al. 2012). This would suggest that they have relatively deep gravitational potential wells typical of clusters, and not normal groups. Other studies, however, have not found substantial differences in the X-ray and optical scaling relations of fossil groups and normal groups (Sun et al. 2009; Voevodkin et al. 2010; Harrison et al. 2012; Girardi et al. 2014). A further inconsistency is that despite having high measured mass concentrations which would indicate early formation epochs, fossil groups in general do not have well-developed cool cores as would be expected from an old system with a cooling time less than the Hubble time (Khosroshahi et al. 2004, 2006; Sun et al. 2004). Conversely, non-fossil groups often do exhibit cool cores (Finoguenov & Ponman 1999).

Cosmological simulations indicate that the fossil stage of a group might be transitory in nature over the lifetime of the group. N-body simulations of a cube 80 Mpc on a side by von Benda-Beckmann et al. (2008) showed that nearly half of the simulated fossil groups identified at z = 0 would not be identified as such at z = 0.3, and virtually none would be identified as fossil groups at z = 0.93. Conversely, few of the fossil groups identified at z = 0.93 remained fossil groups until z = 0, with subsequent infall of one or more luminous galaxies after z = 0.93 removing the group from fossil group status by dropping the magnitude gap between the first and second rank galaxies below 2.0. In addition, Dariush et al. (2010) find that 90% of fossil groups identified in simulations by magnitude gap arguments become non-fossil groups after ∼4 Gyr as a result of infall of luminous galaxies. Thus, it is not clear whether fossil groups truly have a different formation mechanism from normal groups. On one hand, their observed galaxy luminosity distributions are distinct from non-fossil groups; on the other hand the fossil group status of a group seems rather easy to alter with simply the infall of a single luminous galaxy, a process that apparently occurs frequently according to simulations.

In this paper, we use the strongly lensed Cheshire Cat galaxy group to better understand how groups pass in and out of the fossil group stage via merging. We assume a cosmology of H₀ = 70 km s⁻¹ Mpc⁻¹, ${{{\rm \Omega }}}_{M}=0.27$, and ${{{\rm \Omega }}}_{{{\rm \Lambda }}}=0.73$, so that at the group redshift of z = 0.431, $1\prime\prime$ = 5.67 kpc. All our masses and radii are given in units of ${(h{/h}_{70})}^{-1}$, while all our equations specifically state the ${h}_{70}^{-1}$ dependence. We discuss the properties of the group in Section 2. We describe the Gemini GMOS, Chandra, and Hubble Space Telescope (HST) data preparation and analysis in Sections 3, 4, and 5, respectively. We discuss our results in Section 6, and summarize our findings in Section 7.

As mentioned above, groups with high central mass concentrations should make them more efficient strong gravitational lenses, capable of being detected by the Sloan Digital Sky Survey (SDSS). One such very peculiar strong lens system is SDSS J103842.59+484917.7, informally known as the Cheshire Cat (Carroll 1865) Lens. The Cheshire Cat Lens was discovered by the Cambridge Sloan Survey of Wide Arcs in the Sky (Belokurov et al. 2009), and independently by Kubo et al. (2009). The unusual visual nature of the system (Figures 1 and 4) results from four separate background galaxies with redshifts ranging from 0.80 to 2.78 (Bayliss et al. 2011) being lensed by a foreground galaxy group dominated by two closely spaced elliptical galaxies. These two bright elliptical galaxies representing the "eyes" of the Cheshire Cat are SDSS J103843.58+484917.7 and SDSS J103842.68+484920.2 at redshifts z = 0.426 and z = 0.433, respectively, surrounded by a collection of much fainter galaxies. Hereafter, we will refer to SDSS J103843.58+484917.7 as the eastern eye and SDSS J103842.68+484920.2 as the western eye. The Einstein radii of the four lensed galaxies range from 9 0–125 (51–71 kpc). The total projected mass enclosed within the outermost ring is $3.3\times {10}^{13}$ ${M}_{\odot }$ (Belokurov et al. 2009), about a factor of 10 greater than the mass of the Local Group in a volume with a radius only slightly greater than the Milky Way–Large Magellanic Cloud distance. The redshift difference between the two bright eye galaxies corresponds to an 1800 km s⁻¹ rest-frame velocity difference. We argue below that the two galaxies are at the same distance but with high relative velocity, rather than the redshift difference being due to the Hubble flow. The high relative velocity combined with other merging indicators presented below suggests that the eye galaxies are BGGs of two groups (which we designate "G1" for the eastern eye group, and "G2" for the western eye group) that are in the process of merging in a line of sight collision. Furthermore, our analysis indicates that one group was most likely a fossil group prior to the merger, while the other group was a near-fossil group, and that once these eye galaxies merge, the system will become a massive fossil group. Thus, the Cheshire Cat appears to be a prime example of fossil groups leaving, and eventually re-establishing their fossil group status. Figure 1 shows the spatial distribution of the spectroscopically confirmed galaxies members of the Cheshire Cat group. Open (blue) circles are galaxy members of the G1 group and open (red) squares are the galaxy members of the G2 group. The diamonds (yellow) are the spectroscopically confirmed background and foreground galaxies. Galaxies with no redshift information are marked with a cross (yellow).

Zoom In Zoom Out Reset image size

Previous work has studied the galaxy population and strong/weak lensing of the Cheshire Cat and established a virial mass (M₂₀₀) in the range of 1.0–2.9 $\times \;{10}^{14}$ ${M}_{\odot }$, r₂₀₀ of 1.5 Mpc, a concentration parameter (c₂₀₀) of 17–34, and a galaxy richness parameter $({N}_{200})$ of 25 (all values have been scaled to our adopted cosmology; Bayliss et al. 2011; Oguri et al. 2012; Wiesner et al. 2012). In particular, c₂₀₀ is quite high compared to other lensed systems studied in Oguri et al. (2012) and Wiesner et al. (2012).

The optical observations (imaging and spectroscopy) were performed with the Gemini Multi-Object Spectrograph (hereafter GMOS; Hook et al. 2004) at the Gemini North Telescope in Hawaii, in queue mode, as part of the program GN-2011A-Q-25.

The direct images were recorded through the r' and i' filters during the night of 2011 January 4, in dark time, with seeing median values of $0\buildrel{\prime\prime}\over{.} 8$ and $0\buildrel{\prime\prime}\over{.} 9$ for the r' and i' filters, respectively. The night was not photometric. Three 300 s exposures (binned by two in both axes, with pixel scale of 0146) were observed in each filter. Offsets between exposures were used to take into account the gaps between the CCDs (37 un-binned pixels) and for cosmic ray removal. The images were processed with the Gemini GMOS package version 1.8 inside IRAF⁶ . The images were bias/overscan-subtracted, trimmed and flat-fielded. The final processed images were registered to a common pixel position and then combined. Because the group was observed under non-photometric conditions, we used the values listed in the Gemini Performance Monitoring webpage⁷ of 28.22 and 28.15 mag for r' and i' filters, respectively (values closest to the date of the observations), as an approximate zero point for the photometry.

Object detection and extraction of the photometric parameters were obtained with the program Source Extractor (Bertin & Arnouts 1996). The parameter $\mathrm{MAG}\_\mathrm{AUTO}$ was adopted as the value for the total magnitude of the objects. Colors were derived by measuring fluxes inside a fixed circular aperture of 1 2 in diameter (∼8 pixels) in both filters. Because the images were observed under non-photometric conditions, we have to determine the correction (offset) necessary to apply to the magnitudes due to the effect of the extinction produced by the clouds during the observations. The correction was determined by comparing the magnitudes of well isolated and non-saturated field stars, measured at a fixed circular aperture of 3'', with the magnitudes of the same stars, measured inside the same aperture (the fiber magnitude), from the Sloan Digital Sky Survey Data Release 8. (SDSS DR8)⁸ The comparison yielded an offset between the SExtractor magnitudes and the SDSS DR8 magnitudes of −0.41 ± 0.07 mag and −0.41 ± 0.12 mag for the r' and i' filters, respectively. The magnitudes in the final catalog were corrected by these offsets. The final galaxy catalog contains the total magnitudes in r' and i', colors and other structural parameters for 595 objects classified as galaxies, inside an area of $\sim 5\times 5$ arcmin². The galaxies were defined as objects with SExtractor stellarity flag $\leqslant 0.6$ and brighter than r' = 24.5 mag (${M}_{{r}^{\prime }}=-17.7$ mag), which is the magnitude where the number counts start to turn over.

All galaxies brighter than r' = 22 mag and with r' i < 0.8 mag were selected for spectroscopic follow-up (151 galaxies). Two masks were designed for the observations, one with 36 galaxies and another with 39 galaxies (∼50% of the selected sample). The two masks were observed between 2011 May 23 and 2011 July 01. All spectra were observed during dark time, with median seeing values of 0 8 and under photometric conditions. The spectra were acquired using the R400 grating centered at 7000 Å, in order to maximize the wavelength coverage for galaxies at the cluster distance. For each mask, a total exposure time of 2.5 hr (5 × 1800 s) was used. Small offsets of ∼50 Å in spectral direction toward the blue and/or the red were applied between exposures to avoid any lost or important emission/absorption lines presented in the spectra that could lie in the gaps between the GMOS CCDs. Spectroscopic dome flats and CuAr comparison lamp spectra were taken before or after each science exposure.

The spectra were reduced with the Gemini GMOS package, following the standard procedure. The science spectra, dome flats, and comparison lamps were bias-subtracted and trimmed. Spectroscopic dome flats were processed by removing the calibration unit plus GMOS spectral response and the calibration unit uneven illumination, normalizing and leaving only the pixel-to-pixel variations and the fringing. The two-dimensional spectra were then flat-fielded, wavelength calibrated (rms ∼ 0.15–0.20 Å), corrected by S-shape distortions, sky-subtracted and extracted to a one-dimensional format using a fixed aperture of 12 (1.5 × FWHM). The final extracted spectra have a resolution of ∼7.5 Å at ∼7000 Å (measured from the arc lines FWHM) with a dispersion of ∼1.37 Å pixel⁻¹, covering a wavelength interval of ∼4000–8400 Å (the wavelength coverage depends on the position of the slit in the GMOS field of view).

The redshifts of the galaxies were determined using the cross-correlation technique or emission line fitting. For galaxies with clear emission lines, the routine RVIDLINE in the IRAF RV package was used employing a line-by-line Gaussian fit to measure the radial velocity. The residual of the average velocity shifts of all measurements were used to estimate the errors. For early-type galaxies, the observed spectra were cross-correlated with high signal-to-noise templates using the FXCOR program in the RV package inside IRAF. The errors given by FXCOR were estimated using the R statistic of Tonry & Davis (1979).

We were able to measure the redshift for 75 galaxies (100% success rate) of which 34 had redshifts consistent with the Cheshire Cat. The calculated redshift, corrected to the heliocentric reference frame, the corresponding errors, the magnitudes and other relevant information are presented in Table 1. To supplement our sample, spectroscopic redshifts of galaxies in the area of the Cheshire Cat from the work of Bayliss et al. (2014) were added to our sample. Based on a spectroscopic sample of 168 galaxies, Bayliss et al. (2014) obtained 23 redshifts within a physical radius of 2.0 Mpc of the center of the group that were not already in our GMOS catalog, of which 14 were identified as Cheshire Cat galaxies. The coordinates, magnitude, and redshift of these 14 additional Cheshire Cat galaxies and nine foreground/background galaxies are shown in Table 1, and labeled as B1–B23.

Table 1.  Redshifts and Magnitudes of the Observed Galaxies

ID	R.A.	Decl.	r'	i'	(r'−i)	z	${{\rm \Delta }}(z)$	R/#em. lines	Member?
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
387	10 38 31.884	+48 49 38.75	20.76	20.42	0.40	0.246811	0.000120	... / 7	N
405	10 38 33.036	+48 48 45.68	20.01	19.72	0.33	0.246617	0.000267	3.83/...	N
403	10 38 33.072	+48 49 32.38	21.19	20.51	0.67	0.432056	0.000133	4.79/...	Y/G2
446	10 38 33.648	+48 49 19.02	22.33	21.70	0.59	0.426248	0.000087	7.99/...	Y/G1
532	10 38 34.728	+48 48 13.10	22.54	21.74	0.78	0.802925	0.000110	... / 2	N
535	10 38 34.836	+48 49 54.95	21.60	20.90	0.66	0.429267	0.000177	4.52/...	Y/G1
590	10 38 35.700	+48 48 20.09	21.77	21.36	0.38	0.470159	0.000083	... / 6	N
611	10 38 36.060	+48 48 33.55	22.45	21.75	0.73	0.512468	0.000077	8.97/...	N
654	10 38 36.708	+48 50 15.54	22.83	22.39	0.40	0.601393	0.000267	3.77/...	N
679	10 38 37.320	+48 50 33.11	20.37	20.05	0.39	0.247048	0.000267	4.89/...	N
702	10 38 37.464	+48 49 08.69	21.45	20.80	0.64	0.430621	0.000100	9.56/...	Y/G2
694	10 38 38.220	+48 47 43.33	20.36	19.81	0.57	0.441562	0.000267	5.16/...	N
764	10 38 38.544	+48 50 45.20	20.37	20.04	0.34	0.321436	0.000267	4.95/...	N
788	10 38 39.120	+48 49 02.96	20.15	19.73	0.42	0.239652	0.000043	... / 8	N
830	10 38 39.804	+48 49 44.22	22.76	22.12	0.58	0.433140	0.000153	5.20/...	Y/G2
893	10 38 40.596	+48 48 20.95	22.63	22.03	0.59	0.661674	0.000254	... / 4	N
905	10 38 40.956	+48 49 35.26	21.92	21.30	0.64	0.431102	0.000103	11.46/...	Y/G2
894	10 38 41.100	+48 48 20.81	22.19	21.55	0.60	0.426418	0.000267	8.50 4	Y/G1
952	10 38 41.712	+48 48 50.72	21.92	21.26	0.66	0.426842	0.000063	11.40/...	Y/G1
970	10 38 41.820	+48 49 29.53	22.30	21.67	0.59	0.426719	0.000143	8.23/...	Y/G1
1006	10 38 42.432	+48 48 36.36	21.57	20.90	0.66	0.433954	0.000087	12.14/...	Y/G2
1092	10 38 42.720	+48 49 20.28	19.15	18.49	0.68	0.432519	0.000157	8.45/...	Y/G2 (W)
1012	10 38 43.188	+48 49 37.45	21.27	20.64	0.65	0.428253	0.000160	9.02/...	Y/G1
983	10 38 43.332	+48 49 05.12	20.98	20.25	0.71	0.433964	0.000130	8.15/...	Y/G2
972	10 38 43.584	+48 49 17.83	18.86	18.21	0.67	0.425845	0.000140	8.56/...	Y/G1 (E)
1090	10 38 43.836	+48 50 27.13	22.04	21.34	0.72	0.552115	0.000267	4.84/...	N
1224 (B1)	10 38 44.068	+48 49 06.11	23.28	22.67	0.61	0.433700	0.000600	...	Y/G2
1124	10 38 44.736	+48 49 54.70	21.49	20.87	0.64	0.428783	0.000093	12.40/...	Y/G1
1164	10 38 44.916	+48 49 18.12	22.53	21.94	0.62	0.430498	0.000093	8.92/...	Y/G2
1195	10 38 45.456	+48 49 12.79	21.96	21.30	0.67	0.433063	0.000100	9.66/...	Y/G2
1222	10 38 45.816	+48 49 24.96	22.56	21.94	0.66	0.433743	0.000127	7.33/...	Y/G2
1230	10 38 45.996	+48 49 02.64	20.24	19.94	0.33	0.378415	0.000050	... / 7	N
1265	10 38 46.428	+48 49 16.64	21.51	20.82	0.66	0.431549	0.000110	8.06/...	Y/G2
1305	10 38 46.860	+48 50 14.68	22.01	21.38	0.59	0.429240	0.000173	5.60/...	Y/G1
1329	10 38 47.148	+48 49 28.09	22.07	21.38	0.66	0.434998	0.000147	7.09/...	Y/G2
1331	10 38 47.184	+48 49 18.91	21.50	20.85	0.61	0.427146	0.000093	9.43/...	Y/G1
1340	10 38 47.508	+48 50 19.14	20.59	20.12	0.57	0.461146	0.000180	7.29/...	N
1362	10 38 47.760	+48 50 39.52	19.26	18.68	0.59	0.471193	0.000140	6.88/...	N
1393	10 38 48.372	+48 47 46.46	20.86	20.27	0.61	0.429957	0.000123	8.83/...	Y/G1
1406	10 38 48.552	+48 48 30.85	22.21	21.59	0.68	0.835114	0.000190	... / 2	N
1438	10 38 49.020	+48 50 47.62	22.27	21.86	0.47	0.445005	0.000163	5.90/...	N
1435 (B2)	10 38 49.469	+48 48 51.24	21.89	21.43	0.56	0.426500	0.000600	...	Y/G1
1454	10 38 49.740	+48 48 18.65	20.21	19.82	0.37	0.472110	0.000193	8.09/...	N
1576	10 38 51.432	+48 48 12.78	21.57	21.00	0.59	0.433917	0.000270	2.30/...	Y/G2
1605	10 38 51.432	+48 48 48.74	22.66	22.01	0.61	0.212771	0.000113	6.34/...	N
1544	10 38 51.972	+48 48 00.94	20.94	20.67	0.39	0.434270	0.000063	... / 4	Y/G2
1696	10 38 53.088	+48 48 25.81	22.89	22.37	0.33	0.791988	0.000140	... / 4	Nd
1792	10 38 54.168	+48 48 51.84	22.27	21.66	0.60	0.432626	0.000150	5.88/...	Y/G2
B3	10 38 24.822	+48 49 56.384	...	...	...	0.43048	0.00016	...	Y/G2
2043	10 38 28.248	+48 48 51.98	22.70	22.21	0.40	0.260293	0.000157	... / 2	N
2049	10 38 28.320	+48 48 56.23	21.83	21.12	0.70	0.470996	0.000103	9.13/...	N
2080	10 38 29.436	+48 50 28.32	22.57	21.68	0.69	0.890203	0.000077	... / 4	N
2107	10 38 29.508	+48 48 59.90	21.57	21.19	0.35	0.420291	0.000200	... / 6	N
1356	10 38 31.272	+48 49 48.32	21.51	20.95	0.59	0.430565	0.000123	6.64/...	Y/G2
2184	10 38 31.308	+48 50 06.68	19.18	18.86	0.40	0.208317	0.000267	7.29/...	N
381	10 38 32.424	+48 47 43.87	22.34	21.80	0.54	0.431942	0.000137	5.63/...	Y/G2
450	10 38 33.396	+48 47 45.42	22.81	22.34	0.49	0.419921	0.000203	2.84/...	N
517	10 38 34.260	+48 47 40.49	22.37	21.86	0.47	0.472610	0.000107	... / 6	N
644	10 38 36.744	+48 51 27.58	22.47	21.95	0.55	0.433904	0.000160	4.31/...	Y/G2
778	10 38 38.292	+48 51 35.71	20.24	19.76	0.49	0.360463	0.000060	... / 7	N
1096	10 38 44.268	+48 51 19.51	20.19	19.69	0.61	0.431825	0.000140	6.30/...	Y/G2
B4	10 38 44.357	+48 52 13.602	...	...	...	0.11716	0.00050	...	N
1112 (B5)	10 38 44.990	+48 47 15.43	17.47	17.01	0.48	0.176120	0.000500	...	N
1181	10 38 45.420	+48 46 59.34	22.84	22.08	0.34	0.493398	0.000050	... / 4	N
B6	10 38 47.589	+48 45 52.831	...	...	...	0.43289	0.00034	...	Y/G2
1385	10 38 49.020	+48 47 17.56	21.82	21.28	0.67	0.451399	0.000167	6.49/...	N
B7	10 38 49.644	+48 52 20.771	...	...	...	0.44164	0.00006	...	N
1674	10 38 52.764	+48 48 10.04	22.34	21.68	0.58	0.994848	0.000180	... / 2	N
1681	10 38 52.764	+48 47 33.43	21.23	20.59	0.62	0.433710	0.000083	11.96/...	Y/G2
1711	10 38 53.484	+48 50 07.44	21.60	21.06	0.52	0.530524	0.000267	4.68/...	N
1787	10 38 54.420	+48 50 40.09	21.29	20.63	0.68	0.434060	0.000077	12.58/...	Y/G2
1782	10 38 54.708	+48 50 13.63	22.69	21.90	0.72	0.891784	0.000127	... / 2	N
1847	10 38 55.068	+48 50 37.82	22.66	22.24	0.63	0.892748	0.000077	... / 4	N
253	10 38 55.536	+48 49 56.10	21.61	20.76	0.80	0.542742	0.000093	12.37/...	N
228	10 38 55.716	+48 49 49.40	20.89	20.40	0.42	0.239949	0.000017	... / 8	N
208	10 38 55.932	+48 47 18.71	19.85	19.39	0.49	0.190752	0.000294	2.75/...	N
100	10 38 56.940	+48 50 04.63	20.30	19.99	0.34	0.296108	0.000083	... / 8	N
63	10 38 57.660	+48 47 33.32	21.10	20.43	0.75	0.416682	0.000130	4.56/...	N
1	10 38 57.948	+48 50 13.99	21.42	20.70	0.78	0.502117	0.000287	7.91/...	N
102	10 38 58.092	+48 50 16.62	22.35	21.59	0.75	0.502978	0.000247	3.06/...	N
72	10 38 58.488	+48 46 56.06	20.46	19.79	0.73	0.435771	0.000103	9.05/...	Y/G2
2	10 38 58.668	+48 47 06.58	20.89	20.22	0.66	0.435345	0.000223	5.02/...	Y/G2
B8	10 39 00.638	+48 48 34.769	...	...	...	0.20626	0.00050	...	N
B9	10 38 14.912	+48 52 04.017	...	...	...	0.42786	0.00013	...	Y/G1
B10	10 38 15.809	+48 52 59.636	...	...	...	0.43042	0.00013	...	Y/G2
B11	10 38 25.802	+48 44 42.486	...	...	...	0.08975	0.00050	...	N
B12	10 38 37.863	+48 53 25.550	...	...	...	0.42949	0.00009	...	Y/G1
B13	10 38 42.254	+48 53 14.399	...	...	...	0.11697	0.00050	...	N
B14	10 38 44.615	+48 54 15.524	...	...	...	0.42546	0.00022	...	Y/G1
B15	10 38 57.378	+48 45 40.897	...	...	...	0.37074	0.00010	...	N
B16	10 39 02.732	+48 46 35.045	...	...	...	0.42704	0.00016	...	Y/G1
B17	10 39 05.797	+48 53 24.039	...	...	...	0.43186	0.00016	...	Y/G2
B18	10 39 05.940	+48 46 45.029	...	...	...	0.43386	0.00011	...	Y/G2
B19	10 39 07.866	+48 48 09.185	...	...	...	0.40277	0.00009	...	N
B20	10 39 09.609	+48 49 59.268	...	...	...	0.42759	0.00011	...	Y/G1
B21	10 39 10.594	+48 50 23.643	...	...	...	0.42996	0.00012	...	Y/G1
B22	10 39 14.033	+48 51 28.229	...	...	...	0.44482	0.00014	...	N
B23	10 39 18.069	+48 48 05.024	...	...	...	0.42894	0.00009	...	Y/G1

Note. Galaxies within 0.5 r₂₀₀, 0.5–1.0 r₂₀₀, and 1.0–1.5 r₂₀₀ of the X-ray centroid of the Cheshire Cat shown in bold, regular font, and italics, respectively. Column (1): source Extractor Galaxy ID, with B1–B23 from Bayliss et al. (2014); Columns (2), (3): right Ascension and Declination (J2000.0). The units of Right Ascension are hours, minutes and seconds, and the units of Declination are degrees, arcminutes and arcseconds; Columns (4) and (5): total magnitudes in r' and i'; Column (6): Sloan r'−i color measured inside a fixed aperture of 1  2; Columns (7) and (8): redshift and the associated error, corrected to the heliocentric system; Column (9): the R value (Tonry & Davis 1979) and/or the emission lines used to calculate the redshift; Column (10): membership flag: Y/G1—member galaxy of G1, Y/G2—member galaxy of G2, N—background/foreground galaxy.

Download table as:  ASCIITypeset images: 1 2

Combining our group members with those in the Bayliss et al. (2014) sample, there are 48 galaxies in the redshift interval between $0.424\lt z\lt 0.437$ that we identify as being part of the Cheshire Cat system. Thirty-six of the galaxies are located inside the GMOS field of view, which corresponds to the virial radius of the group of 1.3 Mpc (see below), and 12 are located between 1.0–1.5 virial radii (1.3–2.0 Mpc). We use only the galaxies inside the virial radius to calculate the virial mass and group richness values, while we use all 48 galaxies to derive the redshift distribution. The redshift distribution of these galaxies is shown in Figure 2. Two prominent peaks are visible in the redshift distribution, suggesting that there could be more than one structure along the line of sight of the group. In order to investigate its structure, we used the KMM test (Ashman et al. 1994), which is appropriate to detect the presence of two or more components in an observational data set. We have considered whether the data is consistent with a single component or not. The results of applying the test in the homoscedastic mode (common covariance) yields strong evidence that the redshift distribution of member galaxies is at least bimodal, rejecting a single Gaussian model at a confidence level of 98.10% (P-value of 0.019). The P-value is another way to express the statistical significance of the test, and is the probability that a likelihood test statistic would be at least as large as the observed value if the null hypothesis (one component in this case) were true. Assuming two components, the procedure assigns a mean value of $z\sim 0.428$ and $z\sim 0.433$ with 19 (41%) and 29 (59%) galaxies for the structures, respectively (groups we designate G1 and G2 in Figure 2). We used the posterior probability of group membership given by KMM to assign the galaxies to each group. Seventeen out of 19 galaxies in group G1 and 24 out of 29 galaxies in group G2 have a probability >80% of being members of these groups. Two of the galaxies in group G1 have a probability ∼63% of being members of this group. In the case of group G2, there are five galaxies with a probability between 58% and 63% of being members of this group. It cannot be ruled out the possibility that these galaxies are part of a smaller group located between G1 and G2 groups or are part of a filament connecting the two structures. The membership flag and galaxy group assignments are given in the last column of Table 1. The seven galaxies with lower probability of being members of groups G1 and G2 are represented by the shaded histogram in Figure 2.

Zoom In Zoom Out Reset image size

We have used the bi-weight estimator for location and scale (Beers et al. 1990) to calculate the average redshift and the line of sight velocity dispersion of the Cheshire Cat group as as whole. We used an iterative procedure by calculating the location and scale using the Robust Statistics program (Beers et al. 1990) and applying a 3σ clipping algorithm to the results. We repeated this procedure until the velocity dispersion converged to a constant value. We calculate an average redshift for the group of 0.430877 ± 0.000307 and a line of sight velocity dispersion of ${\sigma }_{\mathrm{los}}=659\pm 69$ km s⁻¹. Given this velocity dispersion and using the prescription of Heisler & Bahcall (1985), the virial mass is $(1.5\pm 0.2)\times {10}^{14}$ ${M}_{\odot }$, and the virial radius is 1.3 ± 0.1 Mpc, with uncertainties at the 68% confidence intervals. We do the same thing for G1 and G2 separately, and find average redshifts of $0.427751\;\pm \;0.000250$ and $0.432867\;\pm \;0.000213$, respectively, with with line of sight velocity dispersions of ${\sigma }_{\mathrm{los}}=318\pm 51\;\mathrm{km}\;{{{\rm s}}}^{-1}$ and ${\sigma }_{\mathrm{los}}=337\pm 42\;\mathrm{km}\;{{{\rm s}}}^{-1}$, respectively (see Figure 2). The two groups are separated by ∼1350 km s⁻¹ in the group rest frame. The virial masses are $(0.33\pm 0.07)\times {10}^{14}$ ${M}_{\odot }$and $(0.39\pm 0.06)\times {10}^{14}$ ${M}_{\odot }$, while the virial radii are 0.64 ± 0.07 Mpc and 0.69 ± 0.08 Mpc for groups G1 and G2, respectively. The velocities of the eastern and western eye galaxies are shown with an "E" and "W," respectively. Curiously, neither eye is at the center of its respective group in velocity space. The eastern eye is offset to a lower velocity by $\approx 570$ km s⁻¹, while the western eye is offset by only $\approx 100$ km s⁻¹.

Figure 3 shows the color–magnitude diagram (CMD) of the 595 galaxies brighter than r' = 24.5 mag (black dots). The blue circles and the red squares represent the spectroscopically confirmed members of the G1 and G2 groups, respectively. Filled symbols represent galaxies within 0.5 r₂₀₀ of the combined system, while open symbols represent galaxies outside 0.5 r₂₀₀. In the figure we can see that the red sequence for members of the group shows a moderate scatter in color. We also can see that members of the G1 group are slightly bluer than the members of the G2 group. There is a two magnitude gap in r' between the G1 central galaxy (galaxy "E"; ID 972) and the next brightest confirmed G1 galaxy (ID 1393) within 0.5 r₂₀₀ (0.32 Mpc) of the G1 center. We note that there are two galaxies within this radius that are within two magnitudes in r' of galaxy "E" for which we do not have a spectroscopic redshift. However, one is blue in color (r'−i' < 0.5). Of the 24 galaxies with spectroscopic redshifts that have colors this blue, only one galaxy is a member of either G1 or G2. Thus, it is statistically unlikely that this galaxy actually belongs to the Cheshire Cat. The other galaxy is red, but has an SDSS-determined photometric measurement that is >3σ larger than the redshift of the Cheshire Cat. So G1 was likely a fossil group before its collision with G2, although further optical spectroscopy will be needed for these galaxies. Similar arguments lead to a gap of 1.83 magnitude for the G2 central galaxy ("W"; ID 1092) and the second rank galaxy within 0.5 r₂₀₀ (0.35 Mpc) of the G2 center (ID 983), indicating G2 might have been a near-fossil group before collision with G1. The two magnitude gap between the first- and second-ranked galaxies in a group (${{\rm \Delta }}{m}_{12}$) is not the only criterion that has been used to identify fossil groups. Dariush et al. (2010) suggest a 2.5 magnitude gap between the first- and fourth-ranked galaxies (${{\rm \Delta }}{m}_{14}\gt 2.5$) as an alternative means of identifying early-formed systems. With this definition, G1 was a fossil group unless both galaxies without spectroscopic redshifts mentioned above are G1 members. G2 would once again be a near-fossil group with ${{\rm \Delta }}{m}_{14}=1.83$ at worst, or ${{\rm \Delta }}{m}_{14}=2.36$ if the galaxies without spectroscopic redshifts are ignored⁹ This will become relevant in Section 6.3 when we discuss the past and future of the Cheshire Cat system once this suspected merger is complete.

Zoom In Zoom Out Reset image size

We can use our CMD to estimate N₂₀₀, the number of red ridge galaxies brighter than 0.4 ${L}^{\star }$ that are within a radius of r₂₀₀. For the Cheshire Cat, 0.4 ${L}^{\star }$ corresponds to an apparent i' magnitude of 20.84 (Wiesner et al. 2012). Using Sloan and WIYN data, Wiesner et al. (2012) estimate that ${N}_{200}=25$ for the Cheshire Cat based on broadband optical colors. Our Gemini GMOS imaging covers most of a 1.3 Mpc radius circle around the group, except for small regions to the extreme north and south of the 1.3 Mpc circle. Thus, we are able to image nearly every red ridge galaxy within r₂₀₀. From our spectroscopic data, we confirm 20 red ridge galaxies in the Cheshire Cat within 1.3 Mpc whose i' magnitude in the Sloan Catalog (not in our catalog, since N₂₀₀ was calibrated for SDSS magnitudes) is brighter than 20.84. From Figure 3, there are a few galaxies that lie in the domain of the brighter red ridge group members that do not have redshift measurements. Given the purity of the red ridge sample in this area, it is likely that 2–3 of these galaxies belong to the Cheshire Cat and should be counted in the N₂₀₀ determination. An additional bright red ridge galaxy might be expected statistically in the small portion of the 1.3 Mpc circle that is not covered by our observations to the north and to the south of the group. Thus, there are most likely 23–24 red ridge galaxies brighter than i' = 20.84 within 1.3 Mpc, consistent with Wiesner et al. (2012). Separating the bright red ridge galaxies into either the G1 or G2 groups within their respective r₂₀₀ determined above, we estimate ${N}_{200,G1}=7$ and ${N}_{200,G2}=13$.

Previously unobserved in X-rays, Chandra ACIS-S observations (ObsID 11756 and 12098) for a combined 70.2 ks exposure were obtained on 2010 January 24 and 2010 February 1 for the Cheshire Cat. The data sets were processed beginning with the level 1 event file in a uniform manner following the Chandra data reduction threads employing CIAO 4.5 coupled with CALDB 4.5.7. Events with grades of 0, 2, 3, 4, and 6 were selected for all subsequent processing and analysis. The observation specific bad pixel files were applied from the standard calibration library included with CALDB 4.5.7.

There were no times of excessively high background, so the entire 70.2 ks observation time was utilized. The two individual observations were merged using the CIAO tool reproject_obs for the spatial analysis. X-ray point sources were detected using wavdetect and verified by eye. For spatial work, the surface brightness was extracted using SHERPA in annular rings centered on the peak of X-ray emission. Background was selected from a hot gas-free and point source-subtracted annular region centered on the peak of emission with an extent of 120''–150'' . For spectral work, we used specextract to extract spectra and generate RMF and ARF files for each observation separately, and then combine them into one spectrum by appropriately weighting the two RMFs and ARFs. Spectra were fit within XSPECv12.8 using an apec thermal model with Galactic absorption (using tbabs and a value of ${N}_{H}=1.35\times {10}^{20}$ cm⁻²; Dickey & Lockman 1990) using ${\chi }^{2}$ statistics with energy channels grouped such that each energy bin contained at least 25 photons. The abundance table from Grevesse & Sauval (1998) was adopted. Background was selected from the same region used in the spatial analysis. Spectra were fit in the 0.5–7.0 keV energy range.

The adaptively smoothed 0.3–7.0 keV X-ray contours are shown plotted over the HST image in Figure 4. Diffuse X-ray emission is detected coincident with the group with a centroid of $(\alpha ,\delta )=({10}^{{{\rm h}}}{38}^{{{\rm m}}}43\buildrel{{\rm s}}\over{.} 196$, $+48^\circ 49\prime 19\buildrel{\prime\prime}\over{.} 09)$. By computing the X-ray centroid on a variety of spatial scales from $5\prime\prime$ to $60\prime\prime$ and noting the spread in center values, we estimate the uncertainty of the X-ray centroid to be $2\prime\prime$. The centroid is not located at the position of either eye, but approximately half-way between the eyes (see Figure 5). A bright X-ray point source is coincident with the western eye with a luminosity of ${L}_{X}=1\times {10}^{43}$ erg s⁻¹ (0.3–10 keV) assuming it is at the same redshift as the eye. The background-subtracted X-ray surface brightness profile is shown in Figure 6, with emission detected reliably out to a radius of $120\prime\prime$ (570 kpc) with a total of about 1160 net counts. The profile was fit within SHERPA with a standard β-model with best-fit parameters $\beta =0.60\pm 0.03$ and a core-radius ${r}_{c}={64.7}_{-7.4}^{+8.5}$ kpc (1σ uncertainties), both typical values for groups of galaxies.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The global temperature within a radius of $60\prime\prime$ (where with signal-to-noise begins to decline significantly) is ${4.3}_{-0.7}^{+0.9}$ keV (1σ uncertainties). No useful constraints could be placed on the metallicity of the gas, so it was set to 40% of the solar value for all subsequent work. The inner $60\prime\prime$ was broken into three equally spaced annular regions to derive a temperature profile, shown in Figure 7. The central $20\prime\prime$ shows a strong peak (${5.4}_{-0.8}^{+1.2}$ keV) before dropping below 3 keV at larger radii. The total 0.3–10 keV (0.1–2.4 keV) luminosity within 120'' is $1.1(0.72)\times {10}^{44}$ erg s⁻¹. The central cooling time can be estimated from Voigt & Fabian (2004) given as ${t}_{\mathrm{cool}}=20{({n}_{e}{/10}^{-3}\;{\mathrm{cm}}^{-3})}^{-1}{(T/{10}^{7}\;{{\rm K}})}^{1/2}\;\mathrm{Gyr}$. Given a central temperature of 5.4 keV and a derived central density of $1.5\times {10}^{-2}$ cm⁻³, the central cooling time is ∼3 Gyr.

Zoom In Zoom Out Reset image size

In order to study the eye galaxies and gravitational lenses of the Cheshire Cat in more detail, we examined archival observations taken with HST WFC3 using the F390W, F110W and F160W filters (program ID 13003). HST observed the target for 2400 s with F390W, 1100 s with F110W, and 1200 s with F160W on 2013 March 16. The pivot wavelengths of these filters respectively correspond to 2741, 8060, and 10740 Å in the rest frame. We initially screened for cosmic rays using lacosmic (van Dokkum 2001). We aligned and co-added the filter sub-exposures using the astrodrizzle package (Gonzaga et al. 2012), with a scale of 0 0588 per pixel and a pixel drop fraction of 0.6.

In order to estimate the center-of-mass for the system for comparison with results from the X-ray data, we examined the arc system of the Cheshire Cat. Belokurov et al. (2009) and Bayliss et al. (2011) identified four major ring systems associated with background galaxies at redshifts 0.80, 0.97, 2.20, and 2.78. The arc system displayed in HST imaging, however, is considerably more complex than indicated in either previous study, with multiple overlapping arcs and knots where previously only single knots were assumed. To first order, we assume that the center of mass for the lens is at the center of a circle which fits the surrounding system of arcs and knots, as for a simple "Einstein ring." We then fit circles by eye which correspond approximately to a given arc system. We begin with the assumption that Belokurov et al. (2009) and Bayliss et al. (2011) have recovered the correct arc systems, and then modify the circles as needed, associating knots by color.

Each of these circles is centered within ∼15 of $(\alpha ,\delta )=({10}^{{{\rm h}}}{38}^{{{\rm m}}}43\buildrel{{\rm s}}\over{.} 079$, $+48^\circ 49\prime 18\buildrel{\prime\prime}\over{.} 39)$. The X-ray centroid (Section 4) falls very near the lens arc centroids (Figure 5). This result is consistent with the bulk of lensing matter being located at approximately the midpoint between the eastern and western eyes. If, alternatively, we assumed that either eye is the center of a lens, we would expect to see additional knots at other points around the eye, but we find no evidence for such counterpart knots. New and deeper integral field unit spectroscopy could confirm the configuration of these lensing systems, which in turn would allow a more comprehensive mass distribution reconstruction. Such spectroscopy might be possible from the ground with optimal seeing, given a typical knot thickness of $\sim 0\buildrel{\prime\prime}\over{.} 7$. We note that a weak lensing study of this system by Oguri et al. (2012) found an offset between the center of the inferred mass distribution and the BGG of the group, presumably the eastern eye.

Inspection of the western eye reveals it to be complex, with two distinct nuclei separated by 0 37 (2.1 kpc) embedded off-center in an extended optical halo (Figure 5). We label the nucleus to the northeast WA and the nucleus to the southwest WB. Careful alignment of the Chandra and HST images indicates that the bright X-ray point source mentioned in (Section 4) is associated with the nucleus WA. This X-ray source is coincident with a 5.96 mJy VLA FIRST source. WA also appears significantly brighter in the HST F390W band image than WB, while the nuclei are of nearly equal brightness in the two long wavelength bands. Taken together this indicates that WA harbors an AGN.

6.1. The Mass and Size of the Cheshire Cat Group

Our combined optical and X-ray study of the Cheshire Cat system gives us multiple means of determining the mass of the group at various radii as well as characteristic size estimates of the group. Beginning with the galaxy richness parameter ${N}_{200}=23$, ${N}_{200,G1}=7$, and ${N}_{200,G2}=13$, that we determined in Section 3, we can use scaling relations from the Northern Sky Optical Cluster Survey (NoSOCS; Lopes et al. 2009) to find the predicted r₂₀₀ and M₂₀₀, and compare those values to scaling relations involving measured X-ray quantities. Lopes et al. (2009) find

$$\begin{eqnarray}&&\mathrm{ln}({r}_{200})=0.05+0.39\mathrm{ln}(\displaystyle \frac{{N}_{200}}{25}){h}_{70}^{-1}\;\mathrm{Mpc},\end{eqnarray} \tag{ 1 }$$

which gives ${r}_{200}=1.02$ Mpc, ${r}_{200,G1}=0.64$ Mpc, and ${r}_{200,G2}=0.81$ Mpc. The Lopes et al. (2009) relation for mass M₂₀₀ is

$$\begin{eqnarray}&&\mathrm{ln}({M}_{200})=0.21+0.83\mathrm{ln}(\displaystyle \frac{{N}_{200}}{25})\times {10}^{14}{h}_{70}^{-1}\;{M}_{\odot }\end{eqnarray} \tag{ 2 }$$

to give masses of ${M}_{200}=1.2\times {10}^{14}$ ${M}_{\odot }$, ${M}_{200,G1}=4.3\times {10}^{13}$ ${M}_{\odot }$, and ${M}_{200,G2}=7.2\times {10}^{13}$ ${M}_{\odot }$. Alternatively, the M₂₀₀–N₂₀₀ relation for maxBCG clusters with a mean redshift of z = 0.25 from Johnston et al. (2007)

$$\begin{eqnarray}&&{M}_{200}=1.26\;{(\displaystyle \frac{{N}_{200}}{20})}^{1.28}\times {10}^{14}{h}_{70}^{-1}\;{M}_{\odot },\end{eqnarray} \tag{ 3 }$$

gives masses of ${M}_{200}=1.5\times {10}^{14}$ ${M}_{\odot }$, ${M}_{200,G1}=3.3\times {10}^{13}$ ${M}_{\odot }$, and ${M}_{200,G2}=7.3\times {10}^{13}$ ${M}_{\odot }$. The mass for G1 is consistent with the mass determined from the velocity dispersion (Section 3), while the mass for G2 is about a factor of two larger than the mass determined from the velocity dispersion.

The galaxy richness parameter can also predict the X-ray properties of the Cheshire Cat group. Once again from Lopes et al. (2009):

$$\begin{eqnarray}&&\mathrm{ln}({L}_{X})=-1.85+1.59\mathrm{ln}(\displaystyle \frac{{N}_{200}}{25})\times {10}^{44}{h}_{70}^{-1}\;\mathrm{erg}\;{{{\rm s}}}^{-1},\end{eqnarray} \tag{ 4 }$$

where L_X is measured in the ROSAT 0.1–2.4 keV band, and

$$\begin{eqnarray}&&\mathrm{ln}({T}_{X})=1.05+0.56\mathrm{ln}(\displaystyle \frac{{N}_{200}}{25}){h}_{70}^{-1}\;\mathrm{keV},\end{eqnarray} \tag{ 5 }$$

yielding estimates of ${L}_{X}=1.4\times {10}^{43}$ erg s⁻¹ and T_X = 2.7 keV (if applied to each sub-group separately, we obtain ${L}_{X,G1}=2.1\times {10}^{42}$ erg s⁻¹, ${L}_{X,G2}=5.6\times {10}^{42}$ erg s⁻¹, ${T}_{X,G1}$ = 1.4 keV, and ${T}_{X,G2}$ = 2.0 keV). Compared to the measured values of ${L}_{X}(0.1{{\rm -}}2.4\;\mathrm{keV})=7.2\times {10}^{43}$ erg s⁻¹ and T_X = 4.3 keV, the Cheshire Cat is hotter and much more luminous than predicted. We believe the elevated X-ray luminosity and temperature are the result of shock heating from a merger of the two groups, and we will return to this issue in Section 6.2.

We can use the best-fit hot gas temperature along with the best-fit r_c and β values from the X-ray surface brightness distribution to estimate r₂₀₀ and total gravitational mass via Helsdon & Ponman (2003) and Arnaud & Evrard (1999)

$$\begin{eqnarray}&&{r}_{200}=0.81{(\displaystyle \frac{{kT}}{\mathrm{keV}})}^{1/2}{h}_{70}^{-1}E{(z)}^{-1}\;\mathrm{Mpc},\end{eqnarray} \tag{ 6 }$$

where $E(z)={[{{{\rm \Omega }}}_{m}{(1+z)}^{3}+{{{\rm \Omega }}}_{{{\rm \Lambda }}}]}^{1/2}$ and

$$\begin{eqnarray}M(r={r}_{200}) & = & {8.0810}^{14}\beta \\ & & \times \displaystyle \frac{{kT}}{10\;\mathrm{keV}}\displaystyle \frac{{r}_{200}}{\mathrm{Mpc}}\displaystyle \frac{{({r}_{200}/{r}_{c})}^{2}}{1+{({r}_{200}/{r}_{c})}^{2}}{h}_{70}^{-1}\;{M}_{\odot },\end{eqnarray} \tag{ 7 }$$

respectively. We calculate each quantity using the measured global temperature of 4.3 keV, and the temperature outside the central, hot core beyond 110 kpc of 2.7 keV which might more accurately represent the unshocked virial temperature of the group. For kT = 4.3 keV, we obtain ${r}_{200}=1.36$ Mpc and ${M}_{200}=2.8\times {10}^{14}$ ${M}_{\odot }$, while for kT = 2.7 keV, we obtain ${r}_{200}=1.08$ Mpc and ${M}_{200}=1.3\times {10}^{14}$ ${M}_{\odot }$. Using a lower temperature of kT = 2.7 keV gives an estimate of M₂₀₀ more in line with the estimate derived from the galaxy richness and our estimate from the galaxy velocities, lending further credence to the notion that the hot $\gt 5$ keV gas in the center and the high X-ray luminosity result from shock heating.

We must compare the properties of the Cheshire Cat to cluster samples in the literature rather than groups, since the global temperature of 4.3 keV exceeds that of the hottest groups in group samples (and hotter than all 11 fossil groups in the Miller et al. (2012) sample). With a bolometric X-ray luminosity of ${L}_{X,\mathrm{bol}}=1.1\times {10}^{44}$ erg s⁻¹, the Cheshire Cat lies below the best-fit ${L}_{X,\mathrm{bol}}-{kT}$ relation of Vikhlinin et al. (2009) and Wu et al. (1999), indicating that it is hotter than expected for its X-ray luminosity (or alternatively stated, X-ray underluminous for its temperature). This is not unexpected given that the system has likely undergone a recent collision that has shock heated the gas to higher temperature. The Cheshire Cat lies slightly above the ${L}_{X}-{M}_{200}$ relation of Rykoff et al. (2008) and is consistent with the fossil group sample of (Miller et al. 2012). The summed r' magnitudes of the confirmed galaxies is $6\times {10}^{11}$ ${L}_{\odot }$, putting the Cheshire Cat somewhat above the ${L}_{\mathrm{opt}}-{L}_{X,\mathrm{bol}}$ relation of Girardi et al. (2014) for non-fossil groups, but within the scatter for fossil groups. Given that the Cheshire Cat is believed to be a merger, it is useful to compare its properties to the scaling relations of merging groups and clusters. Kettula et al. (2014) recently determined scaling relations for a sample of clusters from weak lensing and X-ray analyses, and found no difference between their merging and non-merging sub-samples within the uncertainties. Thus, even among merging systems the Cheshire Cat appears to be hotter than expected based on scaling relations.

Using the four gravitational arcs that have measured redshifts and assuming the radius of the arc is equal to the Einstein radius, we can estimate the mass inside the radius following, e.g., Narayan & Bartelmann (1996)

$$\begin{eqnarray}&&M={\theta }_{e}^{2}\displaystyle \frac{{c}^{2}}{4G}\displaystyle \frac{{D}_{s}{D}_{l}}{{D}_{{ls}}},\end{eqnarray} \tag{ 8 }$$

where ${\theta }_{e}$ is the angular Einstein radius, c is the speed of light, G is the gravitational constant, D_s and D_l are the angular diameter distances to the lensed galaxy and the lens, respectively, and D_ls is the distance between the lensed galaxy and the lens. The four arcs have angular radii of 9 0, 9 7, 12 3, and 125 (51, 55, 70, and 71 kpc, respectively) corresponding to lensed galaxies at redshifts 0.97, 2.78, 0.80, and 2.20, leading to mass estimates of $2.3\times {10}^{13}$ ${M}_{\odot }$, $1.8\times {10}^{13}$ ${M}_{\odot }$, $5.3\times {10}^{13}$ ${M}_{\odot }$, and $3.3\times {10}^{13}$ ${M}_{\odot }$, respectively. These are comparable to the values obtained by Belokurov et al. (2009) and Kubo et al. (2009) for the arc redshifts that were known at the time of those studies. The uncertainties of the positions of the centers of the arcs and therefore the angular radii of the arcs of $\sim 1\buildrel{\prime\prime}\over{.} 5$ lead to uncertainties on the mass estimates of 20%–30%, which may explain why the masses do not increase monotonically with radius. The above formula is also only strictly accurate for a point source mass rather than an extended mass distribution. Still, it is clear that the large lensing masses within a small volume demonstrate that this is a very mass-concentrated group, in agreement with the very high c₂₀₀ parameter of 17–34 as measured by Bayliss et al. (2011) and Wiesner et al. (2012).

6.2. A Nearly Certain Line of Sight Merger

6.3. The Past and Future of the Cheshire Cat and Its Implications on Fossil Groups

6.4. Similar Systems to the Cheshire Cat Group

Our optical and X-ray study of the Cheshire Cat gravitational lens system indicates that this system is one of the most likely (and potentially the first) example of a fossil group progenitor. Our primary findings are as follows.

(1) We obtained Gemini GMOS-determined spectroscopic redshifts for 34 group galaxies, and added 14 more from the literature. The group shows a bimodal velocity distribution, with peaks separated by 1350 km s⁻¹. The two eye galaxies lie in different peaks, although neither eye is at the center of its respective velocity peak. Using our imaging and velocity information along with published galaxy richness and X-ray scaling relations, we find that for the Cheshire Cat ${N}_{200}=23$, ${\sigma }_{\mathrm{los}}=659\pm 69$ km s⁻¹, ${r}_{200}$ = 1.0–1.3 Mpc, and ${M}_{200}\;=(1.2{{\rm -}}1.5)\times {10}^{14}$ ${M}_{\odot }$. Dividing the galaxies into two separate sub-group distributions gave velocity dispersions of ${\sigma }_{\mathrm{los}}=318\pm 51$ km s⁻¹ and ${\sigma }_{\mathrm{los}}=337\pm 42$ km s⁻¹, and virial masses of $(0.33\pm 0.07)\times {10}^{14}$ ${M}_{\odot }$ and $(0.39\pm 0.06)\times {10}^{14}$ ${M}_{\odot }$, respectively, for the eastern eye group (G1) and the western eye group (G2). On the other hand, group scaling relations predict that G2 is approximately twice as massive as G1. Four separate gravitational arcs give mass estimates of ${few}\times {10}^{13}$ ${M}_{\odot }$inside a radius of ∼60 kpc.

(2) The Cheshire Cat shows abundant evidence for a merging of two groups of approximately equal mass along the line of sight. In addition to a bimodal velocity distribution, the central X-ray temperature and X-ray luminosity are much higher than expected for its galaxy richness even in comparison to fossil groups, indicating heating from an $M\sim 1.9$–2.4 shock. Also, the center of the X-ray halo and of the four gravitational arcs are not located at the position of either eye galaxy, but midway between them, indicating that neither of the two dominant galaxies in the system reside at the center of the mass distribution of the group. Before the collision, the eastern eye was mostly likely the BGG of a small fossil group, and the western eye was most likely the BGG of a near-fossil group.

(3) HST imaging reveals that the western eye is composed of two nuclei of similar brightness embedded off-center in an extended stellar halo. The nuclei are separated by a projected distance of only 2.1 kpc, and should merge together in a few tens of Myr. The two eye galaxies should merge in less than a Gyr to form a massive fossil group with a ${M}_{r}=-24.0$ BGG at its center, after which the system will be a fossil group. Thus, the Cheshire Cat is a prime candidate for a fossil group progenitor. Observers currently within $z\sim 0.3$ most likely already see the Cheshire Cat as a massive fossil group.

(4) The timescale for the eyes to merge and for the system to satisfy the criteria of a fossil group is shorter than the estimated cooling time of the hot gas in the core of the group. This means there will be a period of time where the system will be a fossil group without a cool core. If the merging of fossil (or near-fossil) groups is a viable mechanism for creating a substantial fraction of present-day fossil groups, the observed lack of cool cores in some fossil groups is understandable.

We thank Rob Proctor, Ka-Wah Wong, Dacheng Lin, and Li Ji for useful conversations and feedback, and the anonymous referee for a thorough reading of the manuscript. We also thank Matthew Bayliss for assisting us in obtaining galaxy redshifts from his sample. This work was supported by Chandra Grant GO0-11148X, and based partially on observations obtained at the Gemini Observatory, which is operated by the Association of Universities for Research in Astronomy, Inc., under a cooperative agreement with the NSF on behalf of the Gemini partnership: the National Science Foundation (United States), the National Research Council (Canada), CONICYT (Chile), the Australian Research Council (Australia), Ministério da Ciência, Tecnologia e Inovação (Brazil) and Ministerio de Ciencia, Tecnología e Innovación Productiva (Argentina).