É relatada a mortalidade de abelhas (213 fêmeas e 60 machos) pertencentes a 20 espécies ao visitarem as inflorescências de Caesalpinia peltophoroides. A presença presumida de um composto tóxico no néctar mostrou grande variação espacial e temporal. Árvores individuais produziram ou não mortalidade de abelhas em diferentes períodos de floração e árvores vizinhas mostraram efeitos distintos em cada floração. A toxicidade sobre as abelhas foi igualmente variável. A maior parte das abelhas morria logo após visitarem as flores; algumas mostravam sinais de narcose, morrendo em seguida; poucas, após um período de narcose, deixaram este estado e voaram. Utilizando o índice de desgaste alar foram observadas abelhas jovens (29%), velhas (26%) e de idade intermediária (45%) (n = 227) entre as abelhas mortas. Se a presença do presumido composto tóxico afeta ou não o fitness da planta requer inicialmente a identificação de qual(is) fator(es) elicia(m) seu aparecimento. Uma vez que C. peltophoroides é amplamente utilizada em projetos paisagísticos urbanos, é necessário estimar o potencial efeito negativo sobre as comunidades de abelhas que vivem em áreas urbanas de diferentes regiões do Brasil. Até que este efeito seja conhecido, recomenda-se aos profissionais do planejamento urbano a não utilização paisagística da sibipiruna em nossas cidades.

It is reported the mortality of bees (213 females and 60 males) belonging to 20 species when visiting the inflorescences of Caesalpinia peltophoroides. The presumed presence of a toxic compound in the nectar showed considerable spatial and temporal variation. Sibipiruna trees may or may not kill bees during different flowering periods and neighboring trees showed distinct effects on bee mortality. The toxicity effect on bees was quite variable. Most of the bees died immediately after visiting some flowers; a few showed signals of narcosis, but after some time they came out of this state and flew away. Using a wing-wear index, 'young' (29%), 'old' (26%) and intermediate aged bees (45%) (n = 227) were observed among the dead insects. Determining whether the presence of a toxic nectar affects plant fitness requires the identification of this chemical and the knowledge of what factor or factors elicit its production. As C. peltophoroides is widely used in urban landscape projects, it is necessary to estimate the potencial negative effect on bee communities living in urban areas of Bazil. We recommend that urban planners avoid the use of sibipiruna trees for landscape purposes until the real mortality effect on bees is clarified.

São apresentados os resultados de um estudo de dois anos sobre o Symphyta neotropical Digelasinus diversipes. Esta espécie é univoltina e comum na Estação Ecológica Jataí, uma reserva de Cerrado no Estado de São Paulo. As larvas alimentam-se gregariamente em Eugenia glazioviana (Myrtaceae) de novembro a abril. Associações entre grupos de alimentação foram freqüentes. Após o período de alimentação, as larvas congregam-se e comunalmente constróem uma massa de casulos (105 casulos ± 60DP; n = 25) aderida ao tronco da planta hospedeira, permanecendo em diapausa como prepupas até o início da estação chuvosa, em outubro. O pico populacional foi observado em dezembro, quando 62% (n = 2.967) dos adultos emergiram. Em condições experimentais, foram observadas emergências das 6:30 às 15:00 h, mas 73,5% (n = 223) dos adultos emergiram entre 9:00 e 12:00 h. Não foi observado, durante a emergência, sequenciamento sexual, mas em um agregado de casulos os machos podem emergir de 20 a 40 dias antes das fêmeas. Após a emergência os machos podem (1) dispersar-se (no início e final do período de emergência; outubro e novembro, janeiro e fevereiro, respectivamente) ou (2) permanecer sobre ou próximo ao agregado de casulos e copular com as fêmeas recém-emergidas (durante o pico de emergência, em dezembro). As cópulas duraram 4,28 minutos (± 3,4DP; n = 28). Ao longo do dia, os machos podem copular com diferentes fêmeas (1-8; n = 5); contudo, as fêmeas copularam apenas uma vez. Em média, as fêmeas emergem com 76 (± 21DP; n = 19) ovos maduros e todos eles são ovipositados de uma só vez sob uma única folha da planta hospedeira. A guarda dos ovos pelas fêmeas durou apenas 2 dias (n = 12) dos 30 necessários para sua incubação. O repertório de comportamentos da fêmea contra potenciais inimigos foi menor do que o observado em outras espécies de Symphyta. Aparentemente, a fêmea induz um necrosamento do tecido foliar que cobre os ovos. Isto formaria uma proteção rígida para os ovos durante sua incubação. Em D. diversipes, adultos de ambos os sexos não se alimentaram (condição controlada) e tiveram vida curta (5,2 dias ± 1,7DP; mínimo 1, máximo 11; n = 179). A razão sexual média foi 2,83 (± 0,014EP) em favor de fêmeas. Os principais fatores de mortalidade foram falhas no desenvolvimento, falta de alimento devido à intensa herbivoria e ataque de parasitóides. Parasitóides criados, Lymeon dieloceri (Costa Lima, 1937) (Ichneumonidae), Conura (Spilochalcis) sp. (Chalcididae) e Perilampus sp. (Perilampidae).

The results of a two-year study of the Neotropical sawfly Digelasinus diversipes are presented. This species is univoltine and common in Estação Ecológica Jataí, a nature reserve of Cerrado vegetation in São Paulo State, Brazil. The larvae feed gregariously on Eugenia glazioviana (Kiaersk.) D. Legrand (Myrtaceae) from November through April. Fusions of different feeding groups were frequent. After the feeding period, larvae congregate and communally construct a cocoon mass (105 cocoons ± 60SD; n = 25) attached to the trunk of the host plant, entering in diapause as prepupae until the beginning of the rainy season in October. The population peak was in December when 62% (n = 2,967) of adult emergence occurred. In experimental conditions, emergence was observed from 6:30 a.m. until 15:00 p.m., but 73.5% (n = 223) of the adults appeared between 9:00 a.m. and 12:00 h. Sex sequence was not observed during emergence but, in a single cocoon mass, males may appear 20 to 40 days earlier than females. After emergence, males may either (1) disperse (early and late in the emergence period; October and November, January and February, respectively) or (2) remain on or near the cocoon mass and mate with newly emerged females (during the emergence peak in December). Copulation lasts 4.28 minutes (±3.4SD; n = 28), and during the day, males may mate with different females (1-8; n = 5); however, females mate only once. Newly emerged females have a mean of 76 mature eggs (± 21SD; n = 19) and oviposited all of them underneath of a single leaf of the host plant. Egg guarding by females lasted only 2 (n = 12) of the 30 days of egg incubation. The female's behavioral repertoire against potential enemies was shorter than that reported for some other sawfly species. Apparently, the female induces necrosis of the leaf tissue that covers the eggs, which forms a hard shelter for the eggs during incubation. Both sexes of D. diversipes do not feed as adults (controlled condition) and have a short life span (5.2 days ± 1.7SD; minimum 1, maximum 11; n = 179). The mean sex ratio was 2.83 (± 0.014SE) in favor of females. Main mortality factors were development defects, lack of food due to intense herbivory, and attack of parasitoids. Parasitoids reared, Lymeon dieloceri (Costa Lima, 1937) (Ichneumonidae), Conura (Spilochalcis) sp. (Chalcididae) and Perilampus sp. (Perilampidae).

Two populations of the wasp Trypoxylon rogenhoferi Kohl, 1884 from São Carlos and Luís Antônio, State of São Paulo, Brazil, were observed and sampled from May 1999 to February 2001 using trap-nests. This mass-provisioning wasp was used to test some aspects of optimal sex allocation theory. Both populations fit all the predictions of the models of Green and Brockmann and Grafen. Maternal provisions determined the size of each offspring, and females allocated well-stocked brood cells to daughters, the sex that benefits most being large. This strategy resulted in a difference in size between the sexes. In São Carlos, female weight at emergence was 1.18 times that of males, in Luís Antônio this value was 1.13. The brood cell volume was correlated with both wing length and weight at emergence in both sexes, and the chance that a given brood cell contained a male offspring decreased with increased brood cell volume. In T. rogenhoferi female body size was related to fitness. Larger females were able to collect more mass of spiders per day, the spiders they captured were heavier, and they provisioned more brood cells per day. They also produced larger daughters. For males, no relationship between body size and fitness was found, but the data were scarce. Since the patterns of provisioning were variable among different females in both study sites, it is possible that the females not follow a unique strategy for sex allocation. The sex ratio and/or investment ratio in the São Carlos population was female-biased and in Luís Antônio, male-biased. In spite of the influence of trap-nests diameters on male production in Luís Antônio, there is some evidence that in São Carlos population the local availability of prey and/or lower rate of parasitism may be major forces in determining the observed sex ratio, but further studies are necessary to verify such hypothesis.

It was compared body size (measured as intertegular span) variance of trapped-males of Eulaema nigrita Lepeletier, 1841 sampled in dry (July, August and September) and wet (December, January and February) seasons of the years 1988/89 and 1994/95 in Viçosa (Minas Gerais, Brazil). It was also compared the body size variance between males and females sampled in three nests found in São Carlos (São Paulo, Brazil) and between these males and trapped ones. The smaller male (6.4 mm) was sampled in June and the bigger (8.9 mm) in July 1994, but the majority (32.3%) showed an intertegular span ranging from 7.8 to 8.0mm. The observed variance in body size was similar between males sampled in nests and trapped-males and the variance found inside a nest was similar between the sexes. However, males sampled in 1988/89 were bigger than 1994/95 males and males sampled in nests were smaller than trapped-males. The variance of the body size of males did differ between 1988/89 and 1994/95. In the first period differences between males sampled in dry or wet season was not observed, but in 1994/95 period the males sampled in dry season showed significantly greater variance in body size than males sampled in wet season. The body size variance did not differ between the wet seasons of 1988/89 and 1994/95. The body size variation, measured as the coefficient of variation in intertegular span, did not differ between males of E. nigrita (CV = 4.3%) and ground-nesting bees. These results show that the variation in body size of males of E. nigrita reflects that one found inside the nests of this bee, being similar among males and females. This variation is expected as result of ecological factors influences the nesting females. El Niño climatic events alone or in association with the lack of local food resources due to deforestation and presence of monocultures might play a role in observed body size variation. However this hypothesis is not sufficient to explain the observed body size variation inside a single nest. The absence of pressures of selection related to the females' advantages of produce large offspring perhaps contributes to the maintenance of the observed size variation. Studies regarding maternal allocation in E. nigrita may be useful to explain either the variation in body size or sex ratio found in this Neotropical bee.

Twenty-two nests of Trypoxylon asuncicola were sampled in Viçosa, State of Minas Gerais, Brazil, in January 2000 and the occupants' behavior of three nests was registered in 2h of direct observation. 528 brood cells were excavated (24±13.84 SD cells per nest), 129 were reused cells, some of them for seven times (meconium deposit count). The mean number of total cells, mean number of open and closed cells, parasitism rate and mean number of reused cells per nest were similar between old and new nests. Parasitism rate and cell reuse were associated with the number of building cells in the nest, but nest aggregate in the sampled area may play some role in the parasitism rate. Brachymeria sp. (Chalcididae) was the most important agent of brood mortality (80%). Other parasites were Melittobia sp. (Eulophidae) (17%) and a species of Icheumonidae (3%). The number of closed cells with immature individuals per nest was 4±4.2SD (N=17) and the mean reproductivity per female was 3±2.4SD (N=5). New nests produced more offspring (0 a 35%) than old nests (0 to 11%). Females and males can be found resting in the nest but copula or guarding behavior by the male was not observed. There is some evidence that in the sampled area the switch of nests by females is great and agonistic behavior between a nest owner and a visitor was not evident. Females were larger (3.9±0.4SD mm) than males (3.1±0.3SD mm) (measured as forewing length). The secondary sex ratio was 1.26 (±0.07 SE) in favor of females, which was not different from 1:1 ratio. The majority (97%) of the sampled larvae of T. asuncicola showed diapause. Some (5.1%) 'anomalous cells' were found.

Males of Euglossa mandibularis Friese, 1899 can be categorized by (1) presence or (2) absence of scutellar tuft (a morphological feature of dense setae found in females' scutellum of the genera Euglossa Latreille, 1802 and Eulaema Lepeletier, 1841). A multivariate statistical analysis using six morphological measures (hindtibia length and width, head width, intertegular span, midbasitarsus length and scape length) show that males of the categories 1 and 2 were significantly different in relation to such measures. The width head, hindtibia length and scape length were the measures that most influenced the observed differences. Levene test suggest there is not effect of stabilizing selection on males without scutellar tuft. Since there is a fluctuation in the proportion of males with scutellar tuft occurring in a single population, the observed morphological differences might be environmentally determined.

The oviposition behavior and development time of Brachymeria villoso (Oliver, 1790) on Trypoxylon albitarse Fabricius, 1804 (Sphecidae) is reported. The parasitism occurs through a little hole made by the parasitoid female in the host nest wall. This hole is used by the parasitoid offspring after its emergence to left the host nest. The parasitoid egg is laid on the host larva before it built its cocoon, when such structure is made the parasitoid egg hatch into a larva that feeds on the living host larva and the parasitoid growth enclosed in the host cocoon without build its own. The development time of B. villosa was 65 days. From two nest of T. albitarse were reared 12 females and one male of B. villosa.

Forty six beetles Scotocryptus melitophilus Reitter, 1881 collected in a nest of Melipona quadrifasciata Lepeletier, 1836 were studied. All beetles were put together in a large petri dish with 14 males, 14 virgin queens and 14 workers of M. quadrifasciata, which was not aggressive towards the S. melitophilus and flew with 3 to 10 beetles attached to them. When groups of 20 workers, 20 males and 20 virgin queens were put separately in three petri dishes, the workers become attractive to beetles after they were seven days old, while the males after the 19 days old. Only one virgin queen five days old attracted one beetle. The anatomic parts of the workers bees more attractive to the beetles when smeared to a filter paper were the head (32%), abdomen (29%) and legs (29%). The part of the workers bees that less attracted the beetles was the thorax (10%). 5. melitophilus, both males and females, are able to distinguish caste, sex and age differences of their host M. quadrifasciata, what can increase the adaptive value of the beetles involved in phoretic behavior.

Collection data of Euglossinae males from Parque Estadual do Rio Doce (PERD) and Viçosa, both areas with remnants of Atlantic Rain Forest (Mata Atlântica) in Minas Gerais state, Brazil are presented. Comparisons made among three fragments with different sizes and states of disturbance from Viçosa showed differences in abundance of most common species and apparently, Eulaema nigrita Lepeletir, 1841 can be an useful indicator of disturbed sites. Some populations of euglossine bees seems to be restrict to a forest fragment, there being few or no flow of individuals or species of one fragment to another, even when they are only 1 km apart. 15 species of euglossines were sampled in PERD, and the most abundant was Eulaema cingulata (Fabricius, 1804). At Viçosa, 10 species were sampled, E. nigrita was the predominant one. Methyl salicylate attracted no males at both sites, in spite of large numbers of species and individuals sampled using this bait in other regions. The majority of species and individuals were collected in the rainy season. Only 0,58% of sampled males carried orchid pollinia (Catasetum Richard, Cycnoches Lindley and Coryanthes Hook) on their bodies. Emergence data of four species of Euglossa Latreille, 1802 reared from trap nests suggest that sex ratio in Euglossini is not a constant within the tribe. A list of 57 euglossine species now known to occur in Mata Atlântica are offered.

Nests of orchid bee Euplusia violacea (Blanchard, 1840) colleted in trap-nests and in the wood of a construction in Viçosa, Minas Gerais (20º45'30"S, 42º52'05"W), Brazil, were studied under laboratory condictions. Some aspects of the biology of these bees were observed, such as sex-ratio, development time, Iongevity and others. Some aspects of the behavior of E. violada were also studied.