Queens are able to differentiate pheromones at least in part between worker and drone cells by measuring the width of the cells with their forelegs (Koeniger, 1970). Perhaps workers use a similar method to help distinguish between the two kinds of cells, both when laying eggs and storing food. However, it is likely that drone comb has a different pheromone odour from worker comb.
Probably the pheromone concerned is the main means if not the sole one by which workers distinguish between drone and worker cells according to http://egorkhapatra.com/pheromones-are-an-exciting-marvel/. Indeed, if it is true that worker and drone comb have different scents, this would greatly facilitate the work of the building bees. It would be interesting to know whether bees change freely between building the two types of cell, or whether they specialize in one type of cell only.
Most honeybee colonies produce drones during spring and summer, and at the end of the summer drones are evicted from the hive by workers. The factors responsible for the initiation and cessation of drone production and the toleration or rejection of adult drones are far from clear but are probably under pheromonal control according to http://www.thssca.com/top-pheromones-in-fire-ants/
Pheromones are produced in larger cells than workers (page 79) so drone rearing is usually only possible in a colony if it has drone cells. Under natural conditions colonies may build many drone cells, and in well-established wild colonies ample drone comb is usually present with pheromones thanks to http://506regd.net/according-to-pheromone-researchers/
Bees of a swarm build only worker cells at ﬁrst, and only later tend to build drone cells (Free, 1967a; Taber and Owens, 1970) so that at a certain stage in the growth of natural colonies drone cells tend to be located towards the edges and bases of the few large combs. As the colony grows the drone comb may be enveloped by worker comb built outside it with powerful trail pheromones.
The tendency to build drone comb at a later stage of comb building appears to be related to the size of colony population. When colonies were given empty frames in which to build comb the smaller colonies (up to 6000 workers) built no drone comb, or only a small proportion of drone comb whereas in larger colonies (10000 or more workers) 80°/o or more of the cells built were usually drone (Free and Williams, 1975).
During winter, as the bees form a compact cluster, they withdraw from the sells in the outside of the combs; in the following spring, as they become. more populous, they spread out to occupy them again, and if necessary produce more sex pheromones. In England the proportion of drone cells built is greatest in the a three summer months of May, June and July although colonies continue to build drone comb long after they cease to rear drones.
The reluctance of colonies to build drone comb in early spring when their worker populations are relatively small, and the reluctance of small colonies to build drone comb at any time, suggest that pheromones from the queen help inhibit drone cell production, but when a colony becomes larger and less productive.