LITERATURA CITADA

The microclimatic conditions a nesting aggregation experience have important influences on the behaviour of bees (chapters 3 and 5). There are, however, considerable difficulties in comparing microclimate measurements from sites spread across the UK. Since readings would have to be taken on different days, variation in weather conditions would confound any direct comparison of results. It was therefore decided that comparisons of prevailing weather conditions, averaged over several years, would be a much more useful data set for analysis. Microclimatic variables were, however, measured for use with behavioural studies within a site.

2.5.1 Weather data

For the 10 field sites across the UK, data about the prevailing weather conditions were compiled from the monthly journal "Weather' published by Royal Meteorological Society. The weather stations closest to each of the field sites were used (Table 2.2) and in order to establish what the prevailing weather conditions were, data from 1988-1992 were averaged. At Gibraltar Point the Nature Reserve's own weather station data were used.

Mean maximum and minimum monthly temperatures were calculated for the flight season of H. ruhicundus (May to September). These were adjusted for altitude differences between the field site and the weather station (Langmuir 1984). In addition the mean monthly hours of sunshine and mean monthly rainfall totals were also calculated for the flight season.

2.5.2 Microclimate data

Several microclimatic variables were measured throughout the activity periods of H ruhicundus. A four channel data logger (Squirrel meter/logger, Eltek Ltd, UK) was used to record ambient temperature (Tg, ground surface temperature (Tg), nest entrance temperature (Tn) and solar radiation (L). These could be automatically logged at any desired interval (usually every 30 minutes) throughout the recording session or read spontaneously from the display readout whenever required.

Temperature measurements were obtained using standard type U m ini­ thermistors (Grant Instruments Ltd, UK) located in different positions. Ta was measured 100 mm above the ground surface in the shade; this is the height at which the bees usually flew above the banking (pers. obs.). For Tg the thermistor was laid on the ground surface and for Tn the thermistor placed in the entrance of a burrow. Solar radiation was measured using a silicon cell pyranometer (Skye Instruments Ltd, UK) which was placed on the ground so that the sensor was absolutely horizontal and away from any shading.

Relative humidity (RH) measurements were obtained from a hand held humidity meter H M I 31 (Vaisala Ltd, UK), which was placed 100 mm above the ground surface in a shaded position. Wind speed (W) was obtained from a hand held anemometer (Testovent 4000, Testoterm Ltd, UK) also placed 100 mm above the ground. This device could be programmed to give a mean wind speed for a 10 minute period, thus giving a single useful value rather than a series of highly fluctuating values.

The solution of energy balance equations used to predict body temperatures of insects are generally tedious and require extensive microclimatic data, the

calculation of derived coefficients, and the making of several assumptions. The use of models of animals and standard operative temperatures (Tgo) can be used to overcome many of these problems. The models are usually dried insects with a thermocouple inserted into the thorax and possess the size, surface area and colour characteristics of a live individual. Since the models produce no heat their body temperature, Tgo, w ill be entirely determined by the environment. This technique allows the effects of radiative, convective and conductive heat exchange to be combined into a single useful measurement. Some studies (Pivnick & McNeil 1987; Corbet et al. 1993) have used hollow copper spheres painted black instead of dead insects in an attempt to standardise Tgo. The black globe temperature w ill be positively related to Tgo, and may be useful in interspecific comparisons; however the size of the globe must be appropriately chosen to be an analogue of an insect's surface area.

Differences between Tgo and the body temperature of an equivalent live animal allow estimates to be made of the amount of physiological heat exchange (TA) that is occurring (evaporation is assumed to be negligible). This w ill include any metabolic heat generated (basal or as a product of some active process) and any physiological thermoregulation such as the control of haemolymph flow around the body (Heinrich 1971). An insect's body temperature is then the sum of Tgo and TA. The Tgo measurements are best equated to a basking insect rather than a flying individual because the forced convection due to wing beating is not a factor that is incorporated into the model.

Using this technique an animal's environment can be effectively mapped by placing the Tgo model in various positions, so that the range of thermal conditions an animal is likely to experience can be characterised (Chappell 1982). The effectiveness of particular behavioural strategies for thermoregulation, such as

posturing, have been quantified (Polcyn &c Chappell 1986) and the relationship between Tgo and behaviour patterns studied (Corbet et al 1993).

TgoS were recorded using a Type K (P9005) thermometer with a modified type K thermocouple (Portée Instruments Ltd, UK). The thermocouple was inserted into the centre of the ventral surface of the thorax of a dried female H . ruhicundus so that the end was located in the centre of the thorax. The same female was used for all recordings of Tgo and had a head width of 2.70 mm (mean head width for females at Invergowrie = 2.72 ± 0.01 (159) mm). Tgo was measured at 100 mm above the ground surface for the same reason Tg was. The use of a dried specimen means that the TgoS recorded will not be the same as for a live (or freshly killed) bee; however, they can be used as an index of actual TbS experienced and so are useful when making comparisons.

For all the above measurements the sensors were set up and allowed to equilibrate for at least 10 minutes before any readings were taken. In addition to the above method of measuring microclimatic variables at a given area, all the above variables could be measured elsewhere for use with "grab and stab" measurements (see section 2.6). These would be taken immediately after catching a bee, and values recorded once the readout had settled down (usually 5 to 10 seconds). This method of recording allowed microclimatic data snapshots to be taken for a particular bee at a particular point in the habitat.