ID, I presume.
Perhaps you'd extend your question a bit. What do you think is not testable or observable, but natural?
Also, do you mean observable in principle (given appropriate sensory or instrumental enhancements) or merely in practice by humans (given limitations on human reason and perception) ?
The frontiers of the latter are subject to change. E.g. bacteria existed before Leeuwenhoek and affected human life, but were not observable in practice before the microscope.
There do appear to be phenomenological limits inherent in nature on (at least direct) observability in principle. Someone mentioned the Casimir effect. The Godel incompleteness theorem and the Heisenberge principle come to mind as well.
Why do you ask ?
™ Version 4.2