One of the most common problems is the space required to properly separate antennas. This is true with both transmitting and receiving antennas. While this article specifically deals with receiving antennas, much of it applies to transmitting systems.
I find it useful to develop a feel for what actually happens in systems, rather than just blindly following advice. One of the largest areas of misunderstanding is in fields, and how antennas generate or receive "signals". Despite what you might hear in folklore, always remember noise has exactly the same radiation characteristics ( so far as antennas or propagation is concerned) as signals from intentional transmitters. There simply isn't an antenna made that sorts "good signals" from "bad noise", except through the directional characteristics of the antenna. Despite rumors and folklore, noise is no more electric field dominant than desired signals. Near an antenna, coupling to multiple unknown sources is largely unpredictable. For an in depth explanation of noise, read the "NOISE" article on this site. For more information on how antennas radiate (and receive), read the "Radiation Myths" article.
Susceptibility to unwanted near-field and induction field coupling between receiving antennas and large transmitting antennas or noise sources is largely unpredictable, al though many problems can be corrected through experimentation with antenna placement and/or detuning or electronically canceling radiation from surrounding structures. for an in-depth explanation of unwanted coupling, see the "Antenna Placement" article on this site.
One of my best arrays on Europe is only a few hundred feet from a transmitting four-square, clearly in the near field of the four-square. There is no detectable influence on this array when it "looks away" from the four-square, although there is a quite noticeable effect in F/B ratio while beaming back into the four-square. The null to the SW is very deep, in excess of 35dB, regardless of four-square tuning. The converse is not true, the null NE when looking SW is only 10-15dB deep. When dominant noise or QRM arrives from the NE, this array is almost useless (compared to others with deeper nulls in the direction of noise). Detuning the four-square completely restores southwest performance, even though spacing is close.