... they don't raise queens from all larvae within the age range needed for conversion but pick a limited number often scattered across the comb.
I've been mulling over this issue, overnight. If you consider the experiments of Doc Miller etc - when trimming back the comb to get to where suitably-aged larvae are housed, in practice bees then select as many as they can physically fit alongside that cut edge for q/c development.
Admittedly, that's an artificial situation - but it does seem to suggest that physical location matters - i.e. an ease of drawing-out those cells. Perhaps it's the exact opposite observation which has more merit - that in the above scenario, they
don't draw queen cells anywhere else on that artifically-cut comb.
Another example perhaps worth considering in this context is the Hopkins method, where the bees will draw as many viable (undefined) q/cells as they are able from the under-surface of a comb held flat. But - when using the Mel Disselkoen modification to the Hopkins method, where a small number of cells - chosen randomly by the beekeeper - are protected, with the remainder being killed-off by applying a powder of some sort, then
all of those protected cells are usually successfully drawn-out into queen cells.
If there was some kind of genetic-based selection taking place, then surely one would expect only a percentage of those protected cells to be developed ?
But in practice, they draw-out as many as they are able, up to some kind of limiting number.
LJ