This can seen in the practice of science: There are more postulated entities than discovered entities. If a scientist proposes phenomenon x to explain some feature of a larger phenomenon (example: proposing that convection currents explain some feature of thunderstorms), they need to do work to show both that phenomenon x actually exists and that phenomenon x explains what is observed.
Why is evidence needed? Evidence increases the probability of something being true. This means that before evidence is found, the ontological claim that x exists is improbable. x exists in the sea of all possible ontological claims. Evidence raises claims out of this sea. The sea of all possible ontological claims is essentially infinite in size*
Because of this feature, when a scientist proposes x, the correct response is disbelief in x until evidence is brought forward to raise it out of the sea. The probability of plucking something out of the sea and being right is low - there are an infinite number of untrue ontological claims and only a finite number of true ontological claims. This position, where disbelief is the default position, is normally referred to as 'skepticism'.
Religious folk will make ontological claims as explanations for observed phenomena. Theists often argue that god exists and is an explanation for things such as
- Religious experiences
* It might be worth noting that applying known phsyical constraints and the assumption that we know our constraints may serve to prevent this being infinite in practice. Even so, the number is still high enough to warrant skepticism for any unevidenced claim. Of course, if we're even allowing gods and other supernatural/spiritual/magical beings to be considered 'possible' in our sea of possible ontological claims, I think examination will reveal that we're making the pool infinite in size again. After all, whose to say there is not a magical invisible pink unicorn with 1030 +1 eyes, right?