In preparing an outline of Lucretius Book One it appears to me that the following is a reasonable summary of Book One line 503:
Since we have determined that everything is composed of only two things, atoms and void, and that nothing else can exist, we conclude that wherever there is empty space there is no body there, and where any body exists, there is no void, and from this we conclude that the atoms are solid bodies free from any void.
Here is Bailey:
 First, since we have found existing a twofold nature of things far differing, the nature of body and of space, in which all things take place, it must needs be that each exists alone by itself and unmixed. For wherever space lies empty, which we call the void, body is not there; moreover, wherever body has its station, there is by no means empty void. Therefore the first bodies are solid and free from void.
I would like to compare Munro and others on this point, but presuming that Bailey has it correct, it seems that this might be an example of reasoning similar to the distinctions that Epicurus draws between pleasure and pain and that where one exists the other is absence.
I make note of this because I would expect that if reasoning like this is embedded so closely into the Physics as to the nature of atoms, it is easy to suspect that the Epicureans became comfortable with such "black and white" logical division, and that this attitude of reasoning carries over from "bodies and void" into "pleasure and pain."
The parallel is pretty clear:
We are not able to observe the atoms or the void directly, but we are confident that they are there based on the impact that their combinations make on our senses. We are not able to observe the ultimate mechanisms of pleasure or pain either, but we are confident of our conclusions about them based on their impact on our feelings.
This method of argument is not going to impress a skeptic who argues that nothing can be known, but it works great for those who are willing to take confidence in reasoning based on repeated evidence, and who are willing to conclude that the results of repeated experience are reliable as a basis for knowledge.