# 4.3.3.6. Density estimation

- Exclude colonies represented by less than a median number of drones in all density calculations in order to overcome the limitation that distant colonies will contribute fewer drones than colonies located in the vicinity of a DCA.
- Quantify the number of colonies represented by an equal or higher than median number of drones.
- Divide this number by the mean mating area of
drones (for the drone samples, 2.5 km
^{2}, Jaffé*et al*., 2009a) or queens (for the worker samples, 4.5 km^{2}, Jaffé*et al*., 2009a) to obtain an estimate of the local density of colonies at the sampling location (see Fig. 28).

**Pros:** less tedious than finding all nests in an area. Method independent
of nest spatial distribution (Arundel *et
al*., 2012).

**Cons:** Fails to detect colonies that do not produce drones. Season
dependence when based on drone trapping, and thus a relevant density figure can
only be obtained during mating season when most colonies produce drones.
Assumes a similar drone investment by all colonies. Inaccuracy due to
variable/non predictable size of mating areas of drones and queens, which can
be different between regions and honey bee populations. High costs involved in
genetic analyses, and a suitable lab space and equipment is needed.

** Fig. 28.** Schematic representation
of the approach to estimate honey bee colony densities based on the frequency
distribution of drones among the reconstructed colonies. For a given sample of
drones from a specific location, the median number of drones per colony is
first calculated. In order to estimate the local density of colonies, those
colonies represented by less than a median number of drones (red columns) need
to be discarded. The number of remaining colonies (blue columns), are then
divided by the mean mating area of drones or queens. This approach aims to
avoid the overestimation of colony densities due to the inclusion of
low-represented colonies, likely to be located beyond mean flight distances of
drones or queens.