The objectives of this work were to determine the squash entomofauna in the region of Viçosa, Minas Gerais state, to study their behavior on flowers and their importance for pollination, verifying the role of each pollinator. The most common species were Trigona spinipes (Fabricius), Trigona hyalinata (Lepeletier), Apis mellifera (L.) and Melipona quadrifasciata (Lepeletier). The visitation behavior of A. mellifera, M. quadrifasciata, and Bombus morio (Swederus) were similar. They visited flowers for nectar collection, positioning themselves vertically between the corolla and the sexual structures of the flowers, with the back directed toward the floral axis, which permitted the removal of pollen from the anthers of flowers with stamens and its deposition on the stigma of flowers with pistils, being considered therefore effective pollinators. Trigona spinipes and T. hyalinata foraged in groups, preventing other species from landing on the flowers which they occupied. Due to their small body size and only infrequent contact with the sexual structures of the flowers, these species are considered occasional pollinators. The number of fruits produced differed between freely visited flowers, those prevented from receiving visits and those visited only a single time by M. quadrifasciata, B. morio, A. mellifera, T. hyalinata or T. spinipes. Flowers prevented from receiving visits or visited only once by T. spinipes did not produce fruits. The remaining pollination systems led to fruitification, with open pollination or a single visit from either M. quadrifasciata or B. morio leading to most fruit production.

O trabalho objetivou identificar as espécies de abelhas sem ferrão (Hymenoptera, Apidae, Apini, Meliponina) presentes em três áreas de cerrado no Maranhão, nordeste do Brasil, por meio do levantamento de seus ninhos. Também foi objetivo do trabalho identificar e caracterizar os substratos vegetais utilizados como locais de nidificação. Pretendeu-se averiguar a abundância e a distribuição espacial de ninhos, bem como padrões de uso dos substratos para nidificação. Foram encontrados 73 ninhos pertencentes a 15 espécies. As espécies mais abundantes foram Partamona chapadicola Pedro & Camargo, 2003 (34,25%) e Oxytrigona sp. 2 (20,55%). Identificaramse 11 espécies vegetais utilizadas para construção dos ninhos. O substrato de nidificação mais freqüente foi Qualea parviflora (Vochysiaceae), na qual encontrou-se 38,36% do total de ninhos (n=28), seguido por Salvertia convallariodora (Vochysiaceae) (n=17; 23,29%). O intervalo de confiança de 95% para o DAP situou-se entre 36,21 cm a 41,68 cm. Este intervalo representaria árvores mais velhas que teriam mais cavidades disponíveis para nidificação, o que poderia ser o caso de S. convallariodora e Q. parviflora. O padrão de dispersão dos substratos com ninhos mostrou-se aleatório nas áreas 1 e 2 e uniforme na área 3. Padrões de distribuição aleatórios seriam um indício da ausência de competição e padrões uniformes indicariam competição.

This study aimed at identifying the stingless bees species (Hymenoptera: Apidae, Apini, Meliponina) those were found in three study sites of cerrado, northeastern Brazil. The study also aimed identifying the plant species those were used as nesting sites by the stingless bees and verifying their characteristics. We intended to verify abundance and spatial distribution of the nests and patterns of the plants used as nesting sites. Seventy-three nests of 15 species were found. The most abundant species were Partamona chapadicola Pedro & Camargo, 2003 (34,25%) and Oxytrigona sp. 2 (20,55%). Eleven tree species were used as nesting sites. The most common host tree species was Qualea parviflora (Vochysiaceae) where 38,36% of the nests (n=28) nestled, followed by Salvertia convallariodora (Vochysiaceae) (n=17, 23,29%). The 95% confidence interval of DHB was between 36,21 cm and 41,68 cm. This interval is related with older trees that should have more hollows to be used as nesting sites. This process might be occurring with S. convallariodora and Q. parviflora. The patterns of spatial distribution for study sites 1 and 2 were random and the pattern for study site 3 was uniform. Random pattern would be a sign for absence of competition. On the other hand uniform pattern would indicate competition.