Relationships between colony size and biological parameters
The number of brood cells was the biological parameter that best explained colony size, followed by the external activity and egg-laying rate (Table 2 and Figure 1). As brood counts involved highly invasive sampling (complete removal of brood comb), which would likely lead to eventual death of the colony, we also considered the second candidate model as the most viable proxy for colony size.
The number of brood cells was the biological parameter that best explained colony size, followed by the external activity and egg-laying rate (Table 2 and Figure 1). The linear regression of each biological parameter tested, and the number of adult bees showed a strong relationship with the external activity for two of the species evaluated (M. flavolineata : R2 = 0.193; M. fasciculata : R2 = 0.752; S.aff. postica : R2 = 0.313; F. longipes : R2 = 0.221; P. minima : R2 = 0.559), while there was a strong relationship with the laying rate in four species (M. flavolineata : R2 = 0.654; M. fasciculata : R2 = 0.879; S.aff. postica : R2 = 0.055; F. longipes : R2 = 0.575; P. minima : R2 = 0.726). In only one species there was a strong relationship with the food stocks (M. flavolineata : R2 = 0.0003;M. fasciculata : R2 = 0.872; S. aff. postica : R2 = 0.014; F. longipes : R2 = 0.109; P. minima : R2 = 0.104), and there was a strong relationship with the number of brood cells in all species evaluated (M. flavolineata : R2 = 0.596; M. fasciculata : R2 = 0.908; S. aff. postica : R2 = 0.562; F. longipes : R2 = 0.784; P. minima : R2 = 0.752) (Figure 2).
The formulae that provided an estimation of population size for all species, by using a simple regression, are: (1) External activity: Y=45.9*X + 448.9; p<0.001; R2=0.913; (2) Brood cells: Y=0.905*X - 21.6; p<0.001, R2=0.964; (3) Egg-laying rate: Y=25.9*X + 372.5; p<0.001; R2=0.906. The number of food stocks was not significant in the simple regression (p<0.108; R2=0.052).