Direct vs. Delegated - jQuery .on()

I am trying to understand this particular difference between the direct and delegated event handlers using the jQuery .on() method. Specifically, the last sentence in this paragraph:

When a selector is provided, the event handler is referred to as delegated. The handler is not called when the event occurs directly on the bound element, but only for descendants (inner elements) that match the selector. jQuery bubbles the event from the event target up to the element where the handler is attached (i.e., innermost to outermost element) and runs the handler for any elements along that path matching the selector.

What does it mean by "runs the handler for any elements"? I made a test page to experiment with the concept. But both following constructs lead to the same behavior:

$("div#target span.green").on("click", function() {
   alert($(this).attr("class") + " is clicked");
});

or,

$("div#target").on("click", "span.green", function() {
   alert($(this).attr("class") + " is clicked");
});

Maybe someone could refer to a different example to clarify this point? Thanks.

Case 1 (direct):

$("div#target span.green").on("click", function() {...});

== Hey! I want every span.green inside div#target to listen up: when you get clicked on, do X.

Case 2 (delegated):

$("div#target").on("click", "span.green", function() {...});

== Hey, div#target! When any of your child elements which are "span.green" get clicked, do X with them.

In other words...

In case 1, each of those spans has been individually given instructions. If new spans get created, they won't have heard the instruction and won't respond to clicks. Each span is directly responsible for its own events.

In case 2, only the container has been given the instruction; it is responsible for noticing clicks on behalf of its child elements. The work of catching events has been delegated. This also means that the instruction will be carried out for child elements that are created in future.

The first way, $("div#target span.green").on(), binds a click handler directly to the span(s) that match the selector at the moment that code is executed. This means if other spans are added later (or have their class changed to match) they have missed out and will not have a click handler. It also means if you later remove the "green" class from one of the spans its click handler will continue to run - jQuery doesn't keep track of how the handler was assigned and check to see if the selector still matches.

The second way, $("div#target").on(), binds a click handler to the div(s) that match (again, this is against those that match at that moment), but when a click occurs somewhere in the div the handler function will only be run if the click occurred not just in the div but in a child element matching the selector in the second parameter to .on(), "span.green". Done this way it doesn't matter when those child spans were created, clicking upon them will still run the handler.

So for a page that isn't dynamically adding or changing its contents you won't notice a difference between the two methods. If you are dynamically adding extra child elements the second syntax means you don't have to worry about assigning click handlers to them because you've already done it once on the parent.