, pela orientação, por todo o incentivo, disposição e amizade. Ao Prof. Ronaldo Zucchi e Dr. Sidnei Mateus, que sempre colaboraram muito, pelas dicas e discussões sobre abelhas sem ferrão. Ao Prof. Carlos A. Navas, que gentilmente disponibilizou seu laboratório e o equipamento de respirometria. Aos alunos do Prof. Navas: Renata, Gustavo, Pedro, Isabel, José, Fernando e Lucas, que pacientemente ajudaram a utilizar o equipamento de respirometria. Ap Prof. José Guilherme Chui-Berlinck e Prof. Jose Eduardo Bicudo, que forneceram bibliografia, ajudaram na formulação dos protocolos e resolveram todas as minhas dúvidas sobre respirometria. Ao Leandro Lourenção (ICB/USP), que ajudou a fazer as análises estatísticas e a utilizar o programa El Temps. Ao Prof. Jim Waterhouse, que discutiu a forma de análise dos dados e deu várias dicas importantes.

The workers of the stingless bee, Melipona quadrifasciata, assume different tasks during their adult life. Newly emerged individuals remain inside the nest, without contact with the external environment. Maturing workers go to more peripheral regions and only the oldest, the foragers, leave the nest. As this diversity of activities implies different metabolic patterns, oxygen consumption has been measured in workers of three different ages: 24-48 h (nurses), 10-15 days (builders), and older than 25 days (foragers). Oxygen consumption of individually isolated workers was determined by intermittent respirometry, under constant darkness and temperature of 25±1 °C. Sets of 24-h measurements were obtained from individuals belonging to each of the three worker groups. Rhythmicity has been assessed in the daily (24 h) and ultradian (5-14 h) domains. This experimental design allowed detection of endogenous rhythms without the influence of the social group and without inflicting stress on the individuals, as would be caused by their longer isolation from the colony. Significant 24-h rhythms in oxygen consumption were present in nurses, builders and foragers; therefore, workers are rhythmic from the age of 24-48 h. However, the amplitude of the circadian rhythm changed according to age: nurses showed the lowest values, while foragers consistently presented the largest ones, about ten times larger than the amplitude of nurses' respiratory rhythm. Ultradian frequencies were detected for all worker groups, the power and frequencies of which varied little with age. This means that the ultradian strength was relatively larger in nurses and apparently maintains some relationship with the queen's oviposition episodes.

Many factors contribute to the activity of animals in the wild. Whilst daily and seasonal rhythms are likely to be present, and to represent underlying biological functions, these will normally be modified by several factors in the environment. Important amongst these are light, temperature, humidity and whether or not it is raining. There is also the problem that the factors might interact, the effect of, say, time of day, being modified by the concomitant temperature. Separating out the effects of these different factors experimentally can be extremely arduous, if not impossible. An alternative approach is to treat the environmental factors as covariants, and then to separate out their effects from the biological ones by statistical means, using Analysis of Covariance, ANCOVA. The potential of this method is illustrated in the current report by a consideration of exits and entries of a colony of bees from their hive. Hourly measurements of this behaviour were taken during the daylight hours for three consecutive days in 11 consecutive months of the year. At the same time, ambient temperature, light intensity, humidity and whether or not it was raining were recorded. ANCOVA enabled the effects of temperature, humidity, light and rainfall upon the exits from the hive and entries back into it to be separated from the effects of time of day and time of the year. The analyses allowed those climatic variables, in addition to time-of-day and time-of-year effects, that influenced behaviour to be identified. Such climatic variables have not been previously isolated, and this might have lead to a misinterpretation of similar results in the past. Having separated out any effects of climatic variables (the covariates), the interaction between time of day and time of the year could then be investigated. Furthermore it has been possible to quantify the effects upon behaviour of each covariate. Rainfall was shown to decrease activity by more than 80%. For the other variables (temperature, humidity and light intensity), the statistical model allowed for the possibility that an increase in the variable initially produced a rise in activity but that this was followed, if the variable continued to rise, by a fall in activity. For light intensity, only a very modest increase in activity was found, and this continued throughout the range of intensities measured. However, for both temperature and humidity, the effects were more marked and showed 'turning points'. That is, activity increased as the ambient temperature rose until activity peaked at about 33-35°C; after which activity began to fall. Similarly, entries into the hive rose with increasing humidity up to a value of 48%, but fell thereafter. By contrast, exits from the hive increased with increasing humidity throughout the range measured. In the biological system tested, this form of analysis has produced valuable information about the way different factors influence activity in a field study. The results strongly suggest that the proposed methodology has a much wider and m...