FWIW - I ran an experiment a few years ago to try and put a figure on just how close entrances could be
on the same side of a hive array without drifting occurring. In order to establish this, I set up two 'quad-nuc' (4x5-frame) hives next to each other thusly:
The box entrances were 9" or so from each other, with the two entrances 'across the gap' being around 12-14". (Can't remember that exact measurement offhand - it's written down somewhere).
During Season 1, the colonies in the yellow and green slots at the furthest ends of the array became very strong; the colonies in the yellow and green slots 'across the gap' were of medium strength; whereas the colonies in the blue and white slots in both boxes had become depleted and were struggling. Very clearly, drifting had been occurring in favour of colonies within end slots, despite the bees being Carniolans which are generally considered to be the least likely of all bees to engage in drifting.
Towards the end of Season 1, and in view of the above, colonies in the blue and white slots were removed, extra brood given to them, and re-housed. Those entrances were then sealed, the outer dividers in both boxes removed leaving just the central divider in place, thus creating 2x 10-frame cavities within each box, with their entrances then being some 24" apart. The separation distance between the hives was increased such that the distance between the adjacent end entrances was 18".
Throughout Season 2, all four colonies developed without any obvious drifting having taken place. An 18" entrance separation was thus established as being the minimum for these bees, at this location, for hives with entrances located on the same side.
Following this experiment, those entrance arrangements were stripped away, and four fresh entrances made, one on each face of the box - following which no obvious drifting has been observed.
Hives within this apiary are placed randomly - some in rows, some not - all are distanced from each other by at least 18", usually by several feet. The majority of entrances face South or South-East, with a view to benefitting from the early morning sun, but not all - indeed, this is a feature which doesn't appear to make a huge difference.
LJ