Eight of the 38 EMSS clusters contain giant arcs. To extract quantitative information from cluster gravitational lenses, one needs a large, homogeneous, well-selected sample of clusters for which there exist measurements of the lengths, widths, radii of curvature, positions and orientations of any gravitational images. Arc lengths and widths are related to the size and ellipticity of the sources, while arc widths are also strongly affected by the steepness of the cluster mass density profile. The radii of curvature and orientation of the arcs depend primarily on the potential of the clusters, providing important clues about the density profile and mass substructure.

We note that making these measurements and attempting to force the often complex lensing configurations into a simplified geometry is difficult and problematic. We choose to list the arc location with respect to the brightest cluster member, which is often not the true center of mass. Measuring the arc length, width, and especially the radius of curvature is more difficult. The arcs often have structure along their length and variation in their width, and are usually too short for an accurate determination of their radius of curvature, $R_{\rm c}$ . In spite of these difficulties, we at least attempt to estimate these quantities and list them in the tables for future use. We measure the arc total magnitudes by integrating all the flux from the arc to the point where the lowest isophote blends into the local sky background. In some cases where the arc is embedded in the halo of one or more galaxies, the flux from the arc is contaminated by light from this halo, and we attempt to subtract the halo flux in order to measure the true arc magnitude.

In Table 2 we list the observed properties of all the arcs in the eight clusters that contain at least one giant arc. Refer to the diagram in Fig. 1 where we illustrate the various arc geometry terms. Columns (1), (2), and (3) contain the cluster name, the cluster (lens) redshift (z_l), and the arc name or label respectively. The length, l, seeing-deconvolved width, w, axis ratio l/w, and radius of curvature $R_{\rm c}$ , are listed in Cols. (4), (5), (6) and (7). The radius of curvature can be determined by measuring the chord length across the two ends of the arc, and the sagittal depth from the arc to this chord, assuming the arc is circular. Column (8) contains the distance, $d_{\rm c}$ , from the (approximate) center of the arc to the center of the cluster. Often the true cluster center is ill defined, so for our purposes, we take the position of the optically dominant galaxy as the cluster center unless otherwise specified. The angle $\theta$ listed in Col. (9) refers to the angular position of the center of the arc with respect to the center of the cluster, where 0 $^{\rm o}$ is defined as north and positive rotations are measured in the counterclockwise direction, while the angle $\varphi$ , listed in Col. (10), is a measure of the orientation of the arc with respect to the cluster radius vector (again the ccw direction is positive). Column (11) contains the (approximate) position of the center of curvature ( $x_{\rm c}$ , $y_{\rm c}$ ) with respect to the cluster center. These numbers are given in arcsec with E and N being the positive directions. Columns (12), (13) and (14) list the B and R magnitudes and (B-R) colors of the arcs (note, V and I magnitudes for some of the arcs can also be found in LeFèvre et al.1994). Finally, Col. (15) contains references to more detailed studies for each cluster.

**Table 2:** Properties of the large arcs
$\begin{table} \includegraphics [angle=180,width=12cm,clip]{ds8209_table2.ps}\end{table}$

5.1 MS0302.7+1658

MS0302.7+1658 at z=0.426 was the first EMSS cluster found to contain a lensed arc (see Fig. 2). The arc is first mentioned by Giraud (1991). Mathez et al.(1992) presented detailed observations of the cluster and 3 arcs: the long ( $l\simeq 10''$ ), so-called "straight'' arc A1 located between the two bright elliptical galaxies G1 and G2 (in the Mathez et al. labelling); a shorter arc A2 (l $\,\sim\,$ 8'') to the south of the brightest cluster galaxy G1; and a faint arc A1W attached to the western end of A1. Giraud (1992) also showed limited evidence for variability in two additional arclets seen in images taken in 1989 and 1991. Models by Mathez et al.reproduce the geometry and shape of the arcs for assumed redshifts of $z\,\sim\,0.8$ and $z\,\sim\,0.6$ for A1 and A2 respectively.

In our R-band CCD images taken in $0\hbox{$.\!\!^{\prime\prime}$}6$ seeing we can easily see the 3 large arcs. We do not see the arclets mentioned by Giraud (1992). In Fig. 3, we also show a true color image of the cluster core where we used the I, R and V frames for the rgb colors. The giant arc between galaxies G1 and G2 (a combination of A1 and A1W) is over 20'' long and appears to be circular centered on G1. However upon closer inspection, we verify that the western and eastern components of A1 are both more-or-less straight and are folded at the point where they join to appear concave. Mathez et al.modeled the cluster potential as a bimodal configuration with the main mass deflector located near the position of the brightest cluster galaxy (BCG) and the secondary deflector located near the second ranked galaxy to the NW. This model is the simplest way to produce "straight'' arc images as in the example of A2390 (Pello et al.1991; Pierre et al.1996).

$\begin{figure} \includegraphics [width=18cm,clip]{fig2.eps}\end{figure}$

Figure 2: R-band images for seven of the eight EMSS clusters with giant arcs

Fabricant et al.(1994) have obtained spectra of a number of galaxies in MS0302+1658 (as well as the neighboring cluster MS0302+1717) and find the cluster has a velocity dispersion of 921⁺¹⁹²_-123 kms^-1. They also observe photometric and spectroscopic evidence of galaxy evolution (i.e. the Butcher-Oemler effect) with $20-30\%$ of the galaxies having strong Balmer absorption and emission lines indicative of a recent episode of star formation. Carlberg et al.(1996) measure a lower value for the velocity dispersion of 646 $\pm$ 93 km s^-1 using 27 galaxies. Ellingson et al.(1997) have produced a photometric and spectroscopic catalog for the galaxies in MS0302+1658 bringing the total number of cluster members with measured redshifts to 94, but a new value for the velocity dispersion has not been published yet.

Kaiser et al.(1998) have performed a weak lensing study of the supercluster at 0.4 using deep I and V band images taken with the UH8K mosaic camera at the CFHT. The supercluster is composed by the three clusters MS0302.7+1658 at z=0.426, MS0302.5+1717 at z=0.425 (both X-ray selected) and of the optically selected cluster CL0303+17, at z=0.418, discovered by Dressler & Gunn in 1992. All of the major concentrations apparent in the X-ray and optical images are detected in the mass reconstructions, and indicate that most of the super-cluster mass, like the early type galaxies, is concentrated in the three X-ray clusters. A mean mass to light ratio for the clusters of M/L_B $\simeq$ 260 h is obtained. The implication of the results for the cosmological density parameter is also discussed.

$\begin{figure} \includegraphics [width=18cm,clip]{fig3.ps}\end{figure}$

Figure 3: True color image of MS0302+1658 showing the blue giant arc system. The image is a 512 $\times$ 512 pixel subarray (750 kpc $\times$ 750 kpc at z=0.426; H₀=50, $q_0\,$ = $\,{1\over 2}$ )that was produced using the V (5400s), R (6000s), and I (3000s) CCD frames. North is up and East is to the left

5.2 MS0440.5+0204

MS0440+0204 at z=0.190 is an optically poor cluster with a compact, multiple-nucleus cD galaxy surrounded by a large halo in which are embedded a number of blue, circular structures that appear to be lensed arcs and arclets (Luppino et al.1993). The largest arc has a length $l\simeq 10''$ and remains unresolved in $0\hbox{$.\!\!^{\prime\prime}$}5$ seeing. Most of the arcs and arclets lie on or near a 24'' (100h₅₀^-1 kpc) radius critical line. Luppino et al.computed an enclosed mass and central mass-to-light ratio of $1.0\ 10^{14} M_{\hbox{$\odot$}}$ and 110 $M_{\hbox{$\odot$}}/L_{\hbox{$\odot$}}$ respectively assuming a source redshift of $z_s\simeq 0.4$ . This cluster has been studied in detail by Gioia et al.(1998a) who present a combined analysis of X-ray imaging and spectroscopic data and HST data. From possible multiple images formed by gravitational lensing of five background sources, Gioia et al.derive limits to the mass distribution in the range 50-100 h₅₀^-1 kpc. For the central 100h₅₀^-1 kpc region, the possible range in projected mass is $6.6-9.5\ 10^{13}$ h₅₀^-1 $M_{\hbox{$\odot$}}$ . At about 600 kpc from the center of the cluster a simple $\beta$ model fit to the X-ray data yields a mass of (1.3 $\pm$ 0.2) 10¹⁴ $M_{\hbox{$\odot$}}$ .At 100 kpc, the lower limit mass from lensing is a factor 2 greater than the X-ray determined mass. In order to reconcile the different mass estimates, Gioia et al.tentatively explore a model where the mass profile increases more rapidly than the X-ray $\beta$ model at large radii.

5.3 MS0451.6-0305

The high X-ray luminosity is confirmed by Donahue & Stocke (1995) who obtained a 16 ksec ROSAT PSPC image of MS0451-0305. They also see a slight elongation of the X-ray isophotes with roughly the same EW orientation as the optical galaxies. Donahue (1996) has also obtained ASCA data that yield a high value for the temperature of 10.4 $\pm$ 1.2 keV. Combining the temperature of the gas with the image parameters of the ROSAT PSPC, Donahue obtains a total mass within a radius of 1 h^-1₅₀ Mpc of 9.7 $^{+3.8}_{-2.2} \ 10^{14}$ h^-1₅₀ $M_{\hbox{$\odot$}}$ . This value is in agreement with the masses implied by both the virial estimates using the velocity dispersion of 1371 km s^-1 reported by the CNOC group (Carlberg et al.1996, see also Ellingson et al.1998, for a spectrophotometric catalog of the cluster galaxies) and with the weak lensing results by Clowe (1998).

Our deep CCD images reveal a rich cluster with a giant arc and at least one smaller arclet (see Figs. 2, 4 and 5). A true color image is shown in Fig. 4. This image is a 750 $\times$ 750 pixel subarray extracted from the much larger 2048² CCD frames, and was produced using the I, R, and V CCD images (UH 88-inch) for the rgb colors. The integration times and seeing (FWHM) in the 3 colors were 4200s and $0\hbox{$.\!\!^{\prime\prime}$}7$ in I, 7200s and $0\hbox{$.\!\!^{\prime\prime}$}6$ in R, and 8400s and $0\hbox{$.\!\!^{\prime\prime}$}7$ in V. The color image measures $2\hbox{$.\mkern-4mu^\prime$}75 \times 2\hbox{$.\mkern-4mu^\prime$}75$ corresponding to 1.2 Mpc $\times$ 1.2 Mpc in the cluster frame (7.39 kpc/arcsec at z=0.55 for H₀=50, $q_0=\,\frac 1 2$ ). The cluster has an obviously flattened (EW) morphology with red galaxies that are easily distinguished from the blue field population. Note that the core of the cluster appears to contain 2 bright galaxies, but the true color image reveals that the southern bright galaxy is quite blue and appears to be a foreground, face-on spiral.

$\begin{figure} \includegraphics [width=18cm,clip]{fig4.ps}\end{figure}$

Figure 4: True color image of MS0451-03 produced using the V (8400s), R (7200s), and I (4200s) CCD frames. The image is a 750 $\times$ 750 pixel subarray and measures $2\hbox{$.\mkern-4mu^\prime$}75 \times 2\hbox{$.\mkern-4mu^\prime$}75$ (1.2 Mpc $\times$ 1.2 Mpc at z=0.55 for H₀ $\,=\,$ 50, $q_0={\frac 1 2}$ ). North is up and East is to the left

$\begin{figure} \includegraphics [width=18cm,clip]{fig5.eps}\end{figure}$

Figure 5: A high contrast image of the core of MS0451.6-0305 at z=0.55. The giant arc A1 and bright arclet A2 are clearly visible. Note that arc A1 has obvious structure along its length, and little if any curvature

The high contrast image in Fig. 5 is the sum of the V, R and I frames and is displayed to reveal the details of the giant arc, A1 and the bright arclet A2. Arc A1 is $\,\sim\,9''$ long and is located $d_{\rm c}=32''$ E of the brightest cluster galaxy. It appears to be marginally resolved, has variable surface brightness, and at least 1 break along its length. Upon close inspection, A1 appears to be another example of a "straight'' arc, but precise measurement of the radius of curvature is difficult. We find a center-to-edge deviation from linearity of $$\mathrel{\mathchoice {\vcenter{\offinterlineskip\halign{\hfil $\displaystyle ...$ 1 pixel over the $\,\sim\,$ 41 pixel length of the arc, allowing us to place a lower limit on the radius of curvature of $$R_{\rm c}\mathrel{\mathchoice {\vcenter{\offinterlineskip\halign{\hfil $\displa... ...nterlineskip\halign{\hfil$\scriptscriptstyle ...$ --only slightly larger than the distance to the BCG. Arclet A2 is $\,\sim\,$ $3\hbox{$.\!\!^{\prime\prime}$}5$ long, and is located 22'' NW of the BCG. Lines drawn normal to this arclet and normal to arc A1 intersect at a point $\,\sim\,$ $7\hbox{$.\!\!^{\prime\prime}$}5$ W of the BCG. This point is $\,\sim\,$ 40'' from A1 and is more consistent with the $R_{\rm c}$ measured for A1. A2 also appears to be curved and its approximate center is consistent with this location.

In many ways, arc A1 resembles the original "straight'' arc in the z=0.231 cluster A2390 (Pello et al.1991; Pierre et al.1996). The A2390 arc is resolved, has breaks along its length, and has a spectroscopically-measured redshift of z=0.913. Both the MS0451-0305 arc A1 and the A2390 arc are orthogonal to the optical and X-ray major axes of their respective clusters (i.e. the arcs trace ellipses aligned with the optical and/or X-ray major axes). As we mentioned above, the lower limit on $R_{\rm c}$ for A1 in MS0451-03 is not significantly larger than the distance to the BCG and is reasonably consistent with the intersection point of the normals of A1 and A2. However, if $R_{\rm c}\gt\gt d_{\rm c}$ , then, as in the case of A2390, we require a secondary mass concentration to the East of the cluster where there is no excess of optical galaxies, suggesting the presence of clumped dark matter.

However, a possible alternative way to produce nearly straight, elongated images is with a "marginal'' lens (Kovner 1987b), i.e. a cluster with a relatively large core radius. In such cases, we would expect to see larger arc widths than with the small core radii clusters, whose arcs would be thin. Since the large arc in MS0451-03 is marginally resolved, we must consider this situation seriously. Note, however, that the orientation of the arc, with respect to both the optical and X-ray major axes, supports our initial statement that the mass density profile is compact, and thus our interpretation that the shape of arc A1 implies mass substructure is the preferred one.

MS0451.6-0305 is in the list of clusters detected by Doug Clowe (1998) using weak gravitational lensing. The mass distribution created with the Kaiser & Squires algorithm (1993) is centered on the bright central galaxy, but the broad EW structure evident in the galaxy distribution is absent. Instead a moderately large northern extension from the central peak present in the mass distribution is consistent with a (much weaker) structure seen in the cluster galaxy distribution.

5.4 MS1006.0+1202

MS1006+1202 is a moderately luminous ( $L_{\rm x}=4.8$ 10⁴⁴ ergs^-1) cluster at z=0.221 with a flattened (NS), irregular morphology and an optical core containing a number of overlapped galaxy images, but with a central galaxy that has neither the size nor brightness to be called a cD. DSG have detected H $_\alpha$ and [NII] emission suggesting the presence of a cooling flow. Carlberg et al.(1996), measure a velocity dispersion of 906 km s^-1 from 26 cluster members.

This cluster contains a remarkable assortment of faint, blue, linear structures that appear to be lensed images (LeFèvre et al.1994). There are three arclets, A1, A2, and A3, located $\,\sim\,22''-30''$ to the North of the BCG. All three arclets are oriented roughly perpendicular to the major axis of the optical galaxy distribution (see the image in Fig. 2, this paper, and in Fig. 1a in LeFèvre etal.). The arclets have lengths ranging from $l\,\sim\,3\hbox{$.\!\!^{\prime\prime}$}7$ for A3 to $l\,\sim\,7\hbox{$.\!\!^{\prime\prime}$}5$ for A1. All three appear to be marginally resolved and have no apparent curvature, but they are so short that any slight curvature would be difficult to recognize. Furthermore, some portions of these arclets may be obscured by a bright ( $R\,\sim\,15$ ) star located near the arclet positions. Presumably arclets A2 and A3 could be 2 pieces of the same large arc since lines extending perpendicular from the centers of these arclets intersect at the same point $\,\sim\,58''$ to the SW, and the break between the two arclets occurs close to the foreground star. The length of A2 and A3 combined is $\,\sim\,27''$ .The longest single arc-like structure, A4, is located 67'' NW of the BCG and is $\,\sim\,10''$ long. There is some question whether this particular object is truly a gravitational lens image. The arc appears to be linear, with galaxies apparently at both ends, and might possibly be some kind of tidal tail from an interacting system. If A4 is a gravitational image, then this cluster may have a peculiar underlying mass distribution. The absence of curvature for A4 suggests the presence of a secondary deflector to the West, where there is no obvious excess of optical galaxies (see the wide field CCD image in Gioia & Luppino 1994). Furthermore, the orientation of A4 is odd when compared to the more-or-less orthogonal orientation of A1-3, although without detailed modelling, it is difficult to say anything quantitative at this time. Again, we note that an alternative interpretation is to consider this cluster an example of a marginal lens with a large core radius, rather than considering the absence of curvature in the arcs as evidence for invisible mass substructure.

5.5 MS1358.4+6245

$\begin{figure} \includegraphics [width=12cm,clip]{fig6.eps}\end{figure}$

Figure 6: R-band image of MS1358.4+6245. The lower panel reveals the faint, curved arc to the SW of the central galaxies, corresponding to the 4.92 red arc in the HST images of Franx et al.(1997). The axes are in arcsec, and the scale at the cluster is 5.8 $\,h_{50}^{-1}$ kpc/arcsec

MS1358+6245 is a very rich cluster at z=0.328 and is extremely X-ray luminous with $L_{\rm x}=1.1$ 10⁴⁵ ergs^-1. This cluster has been the focus of a number of investigations of galaxy evolution in distant, X-ray selected clusters (Luppino et al. 1991; Fabricant et al.1991; Annis 1994; Kelson et al.1997), and displays the Butcher-Oemler effect with a photometrically and spectroscopically determined blue galaxy fraction of $f_{\rm b}\sim20\%$ . During the course of these various investigations, lensed arcs were looked for but not seen, probably because the typical seeing in these data was of order 1.3''-1.5''. From WFPC2 images obtained with HST, Franx et al.(1997) discovered serendipitously an extremely red arc. Keck II spectra of the arc revealed to be the lensed image of a galaxy at z=4.92, among the most distant galaxies known. Both Yee et al.(1998) and Fisher et al.(1998), have performed a spectroscopic survey of this rich cluster. Yee et al.find evidence for a second cluster at the same redshift as MS1358+6245, located at the southern edge of the central cluster field. From their catalog of 232 cluster members Fisher et al.(1998) derive a mean redshift of z=0.3283 $\pm$ 0.0003 and a velocity dispersion of 1027⁺⁵¹_-45 km s^-1 in fair agreement the value of 987 $\pm$ 54 determined by Carlberg et al. (1997). Fisher et al.(1998) show that there is significant evidence for substructure in the central part of the cluster and that the distribution of line-of-sight velocities departs significantly from a Gaussian. They identify two subgroups with at least 10-20 members and dispersions of $\,<\,$ 400 km s^-1. Note that they concentrate on the main body of the cluster and that the second concentration found by Yee et al.south of the cluster center is outside the Fisher et al.field limits. The presence of two subclumps in the cluster, implies that MS1358+6245 has not yet reached equilibrium and is still in the process of virialization and accretion of new members. van Dokkum et al.(1998) investigate the color-magnitude relation for MS1358+6245 using a wide-field mosaic of multicolor WFPC2 images. A weak lensing study by Hoekstra et al.(1998), finds a total projected mass within 1 Mpc h^-1₅₀ of (4.4 $\pm$ 0.6) $10^{14}~M_{\hbox{$\odot$}}$ and a M/L ratio = (90 $\pm$ 13) h₅₀ $M_{\hbox{$\odot$}}/L_{V{\hbox{$\odot$}}}$ consistent with being constant with radius. These authors use the maximum probability extension of the original Kaiser & Squires (1993) algorithm and compare the resulting mass map to the result from a finite field construction algorithm developed by Seitz & Schneider (1996 and 1998). MS1358+6245 was also studied with ROSAT and ASCA by Bautz et al.(1997), and by Allen (1998). The results by Allen, who takes into account the effects of cooling flows on the X-ray images and spectra, imply a projected mass of 4.2^+4.1_-0.8 $10^{14}~M_{\hbox{$\odot$}}$ in excellent agreement with the weak lensing mass by Hoekstra et al.(1998).

As part of this survey, we obtained R-band images in $0\hbox{$.\!\!^{\prime\prime}$}8$ seeing with the UH 2.2m telescope and V-band images in $0\hbox{$.\!\!^{\prime\prime}$}6$ seeing with HRCam on the CFHT. Although there are no bright arcs visible, we did find a large, very faint, curved arc located to the SW of the central galaxies, which corresponds to the red arc serendipitously discovered by Franx et al.(1997). The arc is barely discernable when the R-band image is displayed with "normal'' contrast at the top of Fig. 6, but can be seen clearly in the high contrast image in the lower panel. The same faint, curved structure is also present in our V-band CFHT images. The arc is located 23'' (133 $\,h_{50}^{-1}$ kpc) from the BCG and is 17'' long with a small radius of curvature of $R_{\rm c}\simeq 9''$ (75 $\,h_{50}^{-1}$ kpc). The center of curvature is displaced 13'' SW of the BCG.

5.6 MS1621.5+2640

MS1621+2640 is an optically rich cluster at z=0.426 with one large, faint arc located close to the second brightest cluster galaxy (Luppino & Gioia 1992). The cluster X-ray luminosity is $L_{\rm x}=4.6\ 10^{44}$ ergs^-1. Note the $L_{\rm x}$ listed in the original Luppino & Gioia (1992) paper assumed q₀=0 and used a point source correction rather than the cluster extended emission correction as applied later by Henry et al.(1992) and Gioia & Luppino (1994), and as used in this survey. Ellingson et al.(1997) published photometric and redshift catalogs of galaxies in the field of the cluster as part of the CNOC cluster redshift survey.

The arc is $\,\sim\,9''$ long, has a patchy surface brightness, and has a very small radius of curvature ( $R_{\rm c}\simeq 7''$ ). The arc is located $\,\sim\,14''$ from the BCG and the center of curvature appears to be located $\,\sim\,3''$ S of galaxy 2, between it and the BCG.

5.7 MS2053.7-0449

MS2053-0449 has a redshift of z=0.583 and an X-ray luminosity 5.78 $\ 10^{44}$ ergs^-1. It is not very optically rich, especially when compared with its other z $\,\gt\,$ 0.5 EMSS counterparts like MS0016+1609 or MS0451-0305. Luppino & Gioia (1992) presented high resolution ( $0\hbox{$.\!\!^{\prime\prime}$}6$ seeing) V band images of MS2053-0449 taken with HRCam on the CFHT that revealed a large arc-like structure located $\,\sim\,$ 16'' from the BCG. The arc is $\,\sim\,$ 11'' long and breaks into 2 distinct clumps, labeled A and B, with a center of curvature that appears to be close to the position of the BCG. Luppino & Gioia also noted a very faint arclet, labeled C, located closer in to the BCG.

Clowe (1998) presents a weak lensing study for this cluster which is the least massive of the $z\,\sim\,0.55$ clusters in the EMSS. The mass profile, generated by aperture densitometry is well fit by a "universal'' CDM profile (Navarro et al.1996) with parameters r₂₀₀=520 h^-1 kpc and c=2 assuming a background galaxy redshift $z_{\rm bg}=1.75$ . MS2053-0449 is also well fit by an isothermal sphere model with a velocity dispersion of $\sigma=700$ km s^-1 for $z_{\rm bg}=1.75$ , indicating that the cluster is close to virialization.

Kelson et al.(1997), measure structural parameters and central velocity dispersions for the galaxies in MS2053-0449 to define the fundamental plane relation in a cluster at intermediate redshift. They find that the fundamental plane relation of galaxies is very similar to that of Coma, suggesting that the structure of the early-type galaxies has changed little since z=0.58.

5.8 MS2137.3-2353

MS2137-2353 is a very X-ray luminous cluster ( $L_{\rm x}=1.56\ 10^{45}$ ergs^-1) at z=0.313 with a large, curved arc and the first reported case of a radial gravitational image (Fort et al.1992). The radial image is embedded in the halo of the optically dominant galaxy and is best seen in the HST image of Hammer et al.(1997).

The unique lensing configuration, with both a tangential and radial image and 3 arclets, allowed Mellier, Fort et al.(1993) to develop a tightly constrained mass distribution model with only 2 free parameters (core radius and velocity dispersion) that fit the large arcs and predicted the positions of the 3 arclets. Their model assumed an elliptical cluster potential (circular being ruled out due to the absence of a counter-arc). They found a centrally-peaked mass distribution with a finite core radius that has a value of $r_{\rm c}\simeq50\,h_{50}^{-1}$ kpc; considerably smaller than the core radii derived from X-ray measurements, and in agreement with the general trend that lensing clusters have centrally peaked mass density profiles. Allen (1998) shows that the strong gravitational lensing and X-ray mass measurements for this cooling-flow cluster are in excellent agreement, implying that the thermal pressure dominates over non-thermal processes in the support of the X-ray gas against gravity in the central regions of the cooling-flow clusters, and also validating the hydrostatic assumptions used in the X-ray mass determination.

Hammer et al.(1997) studied in detail the core of this cluster using WFPC2 images. An analysis of the lensing properties of the dark matter component indicates that within 30 to 150 h₅₀^-1 kpc from the mass center, the major axis and ellipticity of the dark matter component are in rather good agreement with those derived from X-ray and visible light, while the dark matter profile has a slope much flatter than that of the visible light (0.875 for the dark matter profile vs. 1.35 for the visible light profile). MS2137-2353 is a good example of an essentially relaxed cluster, as is the case for cooling-flow clusters, with an increasing mass-to-light ratio from the very center to $\,\sim\,$ 150 h₅₀^-1 kpc.