Yes, because they're falsified.
Dark matter is one example. It is niether able to be shown to exist or not exist.
Dark matter could be shown false (no theory can ever be proven true). It has many testable empirical implications, and as far as I know, they have all been verified. Here's a site that discusses some of them:
http://www.astro.queensu.ca/~dursi/dm-tutorial/dm1.html
Here's a site that discusses some new proposed tests for dark matter, that would also give us a better idea of what it precisely is:
http://www.lbl.gov/Science-Articles/Archive/Phys-SNAP-dark-energy.html
String theory, on the other hand, is a very good example of modern scientists engaging in non-testable speculation.
Yes, because they're falsified."
Or because they are untestable at the time (such as when atoms were hypothesized) Your implication that a hypothesis that cannot be tested must be untrue is patently incorrect.
I assume you mean by falsified that you mean they are disproven... am I incorrect in this? Otherwise, I do not know how you can falsify a hypothesis.. DATA and EVIDENCE can be falsified, but how do you misrepresent your own hypothesis?