They are principally sulfates, as the analysis indicates. They do share many morphological features with limestone, though don't they. The dissolution features seem more prominant. I have been curious about the spheres, which almost seem to be in nests here. Many times, there is a short gap all around the sphere. This would fit with my original scenario thatthey are internal mold fossils, where the outer shell is not preserved (organic?). They would fill with fine sediments like those around them - this certainly seems to be the case; the same types of small colored particles seems to compose them as in the surrounding layers. They have lacked layering in all cases where the interior is observed. The texture is identical to the surrounding rock. Only the color and hardness is different, so there must be some compositional difference. However, I doubt the Nasa claim that they are composed almost exclusively of Hematite. I believe that the Hematite is only a fine cement, say 30% of the total, but that since it is on the exterior of all the grains, it has predominated the analysis. Visible and especially x-rays have very little penetration. I view the Hematite as a cement only. Its origin may well be due to former decaying organic material that composed the spheres. Another well supported conclusion (counter to Nasa claims) is that the spheres (and the solid pre-void materials) formed contemporaneously with the sediments. The sediment layers may be seen wrapping around the spheres in many, if not most cases. This does not exclude the possibility that they are concretions (or accretions, precipitates, bio-accretions, etc) that gently dropped into these sediments after forming elsewhere, such as at the surface. Their differential composition could have caused hematite to precipitate later, along with dissolution of the void material(s).