In wiki syntax ou can only add a caption to an image in case you are using "thumb" mode, eg. [[Image:example.jpeg|thumb|some useful text]]. In all other modes but thumb the explanatory text is just shown as tooltip.
However especially with templates (such as town boxes) you want an image without borders but with a caption (as the frame of the town box table already provides a frame).
As of now people are using horrible worarounds such as extra tables and two variables for one image in templates:
Or in templates they are using things like that
|Image = Image.city.jpeg|
|Caption = city center|
The first workaround example has the disadvantage of very ugly and complicated syntax and of course tables make a lot of troubles on converting texts to other formats (for example a book on printing or the Wikipedia DVD).
The second workaround within templates has the huge disadvantage on removing and replacing images. You need to edit to things and need to know that both belong together. If you're and admin cleaning up a template you haven't encountered so far this isn't trivial. Now imagine a Wikimedia Commons admin trying to do this in a wiki he can't speak the language... As well the CommonsDelinker bot has no chance delinking such images properly without leaving rubbish behind. So local wikis do run into technical problems there.
So please enable caption for "border" and "frameless" mode in order to resolve the two decribed problems.