I have two pieces of antenna wire which are roughly 10 m long, with banana plugs at the ends, which I bring to hang in the trees when operating portable. I also bring a 49:1 unun and hopes that my longish coax serves as enough of a counterpoise. For working 20 m, there are two mainstream ways to use these.
The first and simplest is to us one of the wire segments as an endfed halfwave. Accounting for the reflecting ground, this is fairly omnidirectional with some emphasis in the broadside direction.
The second, which I like, is to arrange the two segments in series to a longer endfed antenna and use the first harmonic. Because the two halves of such an antenna are out of phase, the broadside radiation is suppressed, and more of the radiating power will be along the axis of the antenna. (Of course, another reason I use this is to quickly be able to go on 40 m!)
Both these configurations are resonant with SWR < 1:1.5.
By deploying the second configuration, I can easily switch between the two by lowering the antenna and disconnecting the two halves at the centre. Thus, I at least imagine I get some difference in directionality, though I haven't verified this by any measurements.
There could be a third option though, which I haven't been able to find described anywhere:
What if I would leave the segments as above, but feed them both at the same point in the middle? Edit: To clarify, I electrically connect the two segments, and then connect both of them to the same (electrical) point of my transformer, pointing 180° away from each other.
Provided I adjust the segments to be resonant, would I thus get the two halves oscillating in phase, and thus amplify the broadside radiation? (Bonus question: Could I perhaps even angle the two segments somewhat to get even more directionality?)
On the one hand, I could imagine that a full-wavelength piece of wire doesn't like to resonate with two halves in phase as much as it does doing this with the halves out of phase. On the other hand, I know I can drive a skip rope holding the middle with one hand if the ends are stationary, and a skip rope is one of my naive imagined models for efhw's. On the third hand, doing that with a skip rope is a delicate manoeuvre, and you need to take care so that the sides don't interfere with each other.
I messed around with this a bit on my last activation, but I didn't manage to get the SWR low enough to really compare it to the others. This could either be because it's impossible, or just that the adjustments are more delicate than when you feed near the end. I don't have a good home qth where I can put up wires, but I intend to try again next time I'm out. That is, unless there is some good case for why it would be futile and my precious hours in the sun could instead be spent chasing DX instead :)
This is not a duplicate of Multi-wire EFHW?, as the main question there is about how parallel endfed antennas can be resonant. I'm interested in the particular case of two equal segments, and their radiation pattern.