Packages created in different source directories are the same package, as far as the classloader is concerned. It also doesn't matter if the class files are in the same jar or different jars. The JVM does not discriminate based on where the source code came from.
(Of course if you have two jars loaded by different classloaders those are going to be treated differently.)
One case where you frequently have different source trees with the same package is when you have tests in a different directory (using the usual Maven convention where the code is under src/main/java and the tests are in src/test/java) but with the same package as the code that they exercise. These tests are able to exercise protected and package-private parts of the code under test, because they're in the same package as that code.
The path of directories inside the jar should start at the root of the package. (The topmost directory should be /, then one called com or org or whatever, etc.) Packages do form a tree-like structure, and when you put your code in a filesystem you end up having a hierarchy of packages, but the language itself doesn't recognize a concept of "subpackage" (except that packages that start with java are special and get special treatment by the classloader).
Organizing code into packages is done differently by different people. Some people like to organize their code by layer (putting all controllers in one package, all services in another package, all daos in still another package), some like to organize their code by feature.
Package-by-layer is the conventional way of organizing code, it seems to be the preferred practice in the Java community. One consequence of this is that when code implements a feature as a vertical slice at right angles to the package structure (as it may require a new controller endpoint, maybe a new service method, etc.), so closely-related bits of code for the same feature end up scattered across different directories. The Java Practices website makes an interesting case for package-by-feature:
Package By Feature Package-by-feature uses packages to reflect
the feature set. It tries to place all items related to a single
feature (and only that feature) into a single directory/package. This
results in packages with high cohesion and high modularity, and with
minimal coupling between packages. Items that work closely together
are placed next to each other. They aren't spread out all over the
application. It's also interesting to note that, in some cases,
deleting a feature can reduce to a single operation - deleting a
directory. (Deletion operations might be thought of as a good test for
maximum modularity: an item has maximum modularity only if it can be
deleted in a single operation.)
Here's an SO question asking about package by feature or layer.
