cepheid

If you're talking about Ω_total, then it must be 1, else space would not be flat! If you're talking about Ω_matter, then it needn't be 1, and it isn't 1.

Not sure what you're getting at here when you talk about the universe not having "run away." Note that, in the absence of dark energy, models with Ω_tot < 1 still decelerate, just not as rapidly. The limit is the empty universe, which has constant expansion rate. Acceleration is only possible with dark energy.

That's the thing. The missing matter doesn't make Ω unity. Including dark matter, Ω_matter ≈ 0.25. My understanding is that, throughout the 80's and 90's, observations kept resulting in that sort of value. This was kind of a problem in the sense that theorists were hoping for Ω_tot = 1, and we assumed that the universe was matter-dominated and didn't know anything about any other sort of constituent (like dark energy). However, it was not a problem in the sense that you seem to be implying. You seem to be under the impression that the motivation for introducing dark matter was to get Ω_tot to be 1. This is not correct. The motivation for introducing dark matter has always been to explain other anomalous observations, most of which have to do with gravitationally-bound systems (on various spatial scales).

The dark matter story started as early as the 30's, with astronomer Fritz Zwicky. He observed that velocity dispersions of galaxy clusters appeared to be too high, given the amount of luminous matter that was present. In other words, the individual galaxies in the cluster were moving too fast and ought to escape the cluster. Yet, it was clear that the galaxy clusters were gravitationally-bound systems. The explanation he proposed at the time was the presence of a large amount of non-luminous matter that we could not detect. The same sorts of conclusions were drawn about individual galaxies when observations of their "rotation curves" (plots of rotational speed vs. distance from centre) showed that they were roughly flat (the speeds of the stellar orbits around the galactic centre were roughly the same at all radii). At larger radii, these orbits were much faster than Newtonian gravity would have predicted, and as a result, these galaxies ought to have been flying apart. In both of these situations, the two possible explanations were "there is extra matter present that we cannot detect, but that is providing the necessary gravity to keep the system bound," or, "there is something wrong with our theories of gravity -- maybe gravity behaves differently on large spatial scales or something." My understanding is that, although some modified gravity theories have had limited success in explaining some of the observations, none of them have been able to successfully explain all observations on all spatial scales.

So far I've talked about observational evidence for dark matter at the scales of individual galaxies, and at the scales of clusters of galaxies. What about really really large spatial scales? Well, it turns out that dark matter plays a crucial role in our models of structure formation i.e. in models that describe how tiny density fluctuations in an initially smooth/homogeneous universe grew under gravity to form the large scale structure of the universe that we see today. If you assume that there was no dark matter and only baryonic matter, you run into a problem. The basic problem has to do with the temperature fluctuations that we see in the Cosmic Microwave Background (CMB) radiation. These fluctuations in temperature are, roughly speaking, caused by (and have the same order of magnitude as) the density perturbations in the primordial plasma that existed in the very early universe (the soup of charged particles and photons). Baryonic theories of structure formation predict that, at the time, these fluctuations would have had to have been at a level of around 10^-3 (i.e. that's the factor by which there would by an overdensity or underdensity relative to the mean density level). If the fluctuations were any smaller, then under the (no dark matter) models of structure growth, there would not have been sufficient time for the overdensities to have reached a large enough level for there to be galaxies, stars etc. at the present day. The problem is, these initial perturbations are much smaller. The temperature fluctuations in the CMB are observed to be a level of ~10^-5. This is a problem! Theories of structure formation without dark matter predict that we shouldn't exist, because the matter density perturbations shouldn't have grown large enough to form the structures that we see. In other words, the universe should still be much smoother and less clumpy than it is. However, if you include non-baryonic dark matter in the models of structure formation, this problem goes away (I'm getting a bit too tired of typing to explain the details of how that works). In fact, the standard lambda-CDM (cold dark matter) cosmology has had tremendous success in explaining (with great precision) how the large scale structure that we see in the universe today developed. Furthermore, detailed observations of the fluctuations in the CMB have allowed us to figure out the "recipe" for the constituents of the universe that you hear quoted all over the place: ~73% dark energy, ~22% dark matter, and ~5% ordinary (baryonic) matter -- i.e. stuff that is made out of atoms. These values are in agreement with those obtained from other non-CMB observations.

So, we have a fair bit of observational evidence that the vast majority of the matter in the universe is some sort of non-baryonic matter that doesn't interact very strongly with anything else by means of any of the four fundamental forces except gravity. I would go even further and say that the observational evidence for dark matter is nowadays considered to be fairly conclusive. We can "see" it, in the sense that we can use gravitational lensing to map out how it is distributed spatially. This works well if you have a very dense region (such as a galaxy cluster) that lenses background objects very strongly. One of the best examples of this is the observations of the bullet cluster, a collision between two galaxy clusters that is considered to be a sort of "smoking-gun" for the presence of dark matter and nail in the coffin for alternative explanations such as modified gravity. Again, I've typed enough and would rather not explain why the bullet cluster is such strong evidence for DM. Instead, here's a link: http://apod.nasa.gov/apod/ap060824.html