To get some intuition of why quarter wavelengths seem to come up a lot here, it's helpful to imagine a long transmission line carrying your signal. What would h

To get some intuition of why quarter wavelengths seem to come up a lot here, it's helpful to imagine a long transmission line carrying your signal. What would happen if you were to add a quarter wavelength stub to this transmission line? So something like: Source ----------------------------------- Dest | | lambda/4 | stub The transmission line carries an electrical current proportional to the signal from "Source" to "Dest." When the current gets to the stub, you can think of it as going down two paths -- continuing to Dest, as well as down the stub. When the current component going down the stub gets to the end of the stub, it reflects and travels back up the stub (reflection coefficient of the "open circuit" at the end of the stub is 1). By the time the reflection gets back to the transmission line, you have a phase offset of half a wavelength relative to original signal on the transmission line. If you add two signals that are half a wavelength apart, they combine destructively (cancel each other out). So the quarter wave stub acts like a very basic filter, and you actually won't see much of your signal at "Dest" when your transmission line has a quarter-wave stub like this. Microwave engineers use all sorts of tricks involving transmission line segments that are a quarter wavelength long. So when you're trying to build an antenna (e.g. out of foil, or if you're adding metal support structures to a large antenna), it's helpful to use the heuristic: "if you need to add features that are not part of the original design, try to keep them smaller than a quarter wavelength." blackguardx is right that restricting the maximum feature size to 1/10 of the wavelength is probably a better rule of thumb if you want good performance. If you can get it down to say, 1/2 of a quarter wavelength (1/8 wavelength), there's a reasonable chance it'll still work.