There is an association apiary with up to 30 hives. Several who operate there say they get lower yields on that large site than they do at home or from smaller apiaries a couple of miles away. Very few of the hives are moved to OSR or other seasonal forage, most work the same suburban/mixed area all year and there are no significant field crops. With only the satellite images from Google/Bing to go on, there is nothing to suggest other than a fairly uniform spread of forage across the area they cover.
There are references in several of the books to foraging efficiently, that is the further bees fly the more food gets burnt travelling and less gets back to the hive. That seems a reasonable explanation of what's going on. As a simple model, the maximum honey yield would be from the shortest distance. If we assumed stocking should be maximised, then spacing the hives uniformly across the area would be the ideal case. Logistically, you don't have access to that many sites and it would be a pain to operate if they are not grouped for the keepers benefit.
Site sizes are always going to be a compromise between effort for the bees and effort for the keeper. Something like 70 years of empiricism around here has led to most shared apiary sites being around 8 to 10 hives with the garden having one or two hives. The observed effects may be self fulfilling in that the keeper places the 'best' bee stocks where they expect the best yield, that is in the smaller apiaries. That is their strategy and to them, the yields appear to confirm the value of the strategy. From a neutral observers point of view, however, there is no way of telling what proportions are due to the site size and what is due to the choice of bees placed there.
