Calorimetry allows the direct measurement of heat production (energy turnover) of individual bees or bee groups (Fahrenholz et al., 1989, 1992). Differential calorimeters, constructed as twin-setups of 'camping box calorimeters' are the most suited variants for honey bee (and insect) calorimetry (e.g. Fahrenholz et al., 1989; Lamprecht and Schmolz, 2000). Depending on construction size, measurements are possible on individual honey bee larvae (Schmolz and Lamprecht, 2000) and adult bees (Schmolz et al., 2002), in whole colonies (Fahrenholz et al., 1992), and even in groups of tiny honey bee parasites (Varroa destructor, Garedew et al., 2004). The insects inside these instruments can be continuously supplied with fresh air if necessary, which allows long duration measurements (Schmolz and Lamprecht, 2000; Schmolz et al., 2002).
Construction of such a calorimeter is not difficult ("poor man's calorimeters"; Lamprecht and Schmolz, 2000; see also Lighton, 2008). Peltier units are attached to the bottom of camping boxes to keep the instrument at a desired temperature. It is an inherent property of Peltier elements that heating or cooling of one side can be regulated via the direction of the current applied.
To sense heat flux, such constructions use the fact that Peltier elements can also be used as thermopile heat flux sensors, because a temperature gradient along them produces a current (Seebeck effect). Such boxes are commercially available with variable volumes (~5–30 l). Heat flux sensitivity is about 10±30 mV/W, which is in the range of commercially available instruments (Lamprecht and Schmolz, 2000). To achieve an acceptable baseline stability, a differential (twin) setup is recommended, with one box serving as (empty) reference unit. An additional styrofoam insulation around both boxes compensates for an otherwise poor baseline stability.
A frequent disadvantage of calorimeters is their large time constant (see Lamprecht and Schmolz, 2000; Lighton, 2008). This means that the output signal of heat production may be heavily distorted, hiding short-time dynamic properties of events of heat production or heat release. Appropriate calibration and signal processing may help to compensate for such shortcomings and to restore the dynamic structure of the signal (see pp. 74-75 in Lighton, 2008).