Why does the gas in the Bullet Cluster act like a lens? In the Bullet Cluster, we have two well defined galaxy clusters that exhibit gravity lensing.  Between them is a mass of gas (mostly hydrogen and helium) that is roughly eight to nine times the mass of the two galaxies.  The mass of gas doesn't exhibit any lensing.  My question is, why do we expect it to act like a lens?  For a lens to work, no matter how massive, it needs to focus the rays of light coming from behind it.  Since this cloud could be shaped like a box, or a cone or have any shape at all, why are we so convinced that it can focus light?  There's no question it can bend light, the question is how is it able to focus light.
A glass marble and a class cube can have the same mass. One will act like a lens, the other won't.
Edit: The answers devolved into a discussion of ΛCDM that I wanted to avoid.  Let me rephrase the question: We would expect a spherical, relaxed X-Ray cloud (without significant stellar matter) to exhibit some form of gravitational lensing.  Is there any reason that after going through a collision like the Bullet Cluster, the same cloud would display the same kind of lensing (on the scale of one or two Mpcs) or would it just scatter the light in an incoherent way?
 A: The Bullet Cluster is the after-effect of the collision of two galaxy clusters.  The gas cloud in the center of the two clusters consists of baryonic particles that interacted with each other through both electrostatic and gravity forces, slowing them down.  Most of the mass of the Bullet Cluster, however, consists of dark matter that passed through the collision unimpeded by electrostatic forces, and that now lies outside the central gas cloud.
Therefore, gravitational lensing of light that passes to us from behind the Bullet Cluster, is strongest on either side of the central gas cloud.  This can be seen in the 7th photo down as you scroll down from the top in this link: http://www.cfhtlens.org/public/what-gravitational-lensing.  Most of the lensing effect occurs when light passes through the dark matter.  There is some weak lensing from the gas cloud, but its mass is far less than the dark matter, so its lensing is weaker.
Gravitational lensing does not necessarily mean that light is focused as through a well-shaped optical lens.  It's effects can be distortion and multiple images.  All light that comes to us undergoes some gravitational lensing, but only strong gravitational lensing produces dramatic effects.
A: We don't require that a lens focus light, only that it bend light, as you said. From optics, a concave lens disperses light, sending it out at wider angles than it comes in. From astronomy, other examples of lensing include cases where a single image is broken up into multiple images, or is smeared out into a distorted version of the original. A common lensing effect is our ability to see more than 50% of the surface of a spherical body because light from just past the horizon is bent in our direction.
As far as the shape of the gas cloud plays in, the total mass and the center of gravity of the cloud are the primary concerns in determining lensing. As long as the cloud is sufficiently compact, the shape only slightly changes the otherwise spherical distortion of the lensed images.
Edit: regarding the difference between a glass marble and a glass cube: keep in mind that those objects change the path of light through refraction, while galaxy clusters change the path of light by curving space. Refraction depends on the refractive indices of two materials and an interface between them; gravitational lensing can occur in vacuum near a point mass. The two processes are very different, and we only call the latter "lensing" because the effects are like using a lens, not because the lensing body forms an optical lens.
A: 90% of the baryonic matter is in the form of gas. The majority of the gravitating matter, that is causing the strong lensing effects, is dark and appears to be associated/clumped within the two clusters and around their galaxies. 
As is completely explained in the papers you have been referred to in previous questions (e.g. Paraficz et al. 2012), the gas actually has a fairly mild effect on the lensing signatures, but it needs to be included in the modelling because the gas is offset from the dark matter. The intra-cluster gas is modelled using X-ray observations and included in the gravitational modelling along with the visible matter in the galaxies and two clumps of dark matter associated with the galaxy clusters. The latter are dominant (along with the dark matter halos around the galaxies). According to this paper the gas makes up about 9% of the gravitating matter and provides only "an external shear to the strong lensing model".
