AbstractMammalian oocytes present challenges for optimal study by electron microscopy (EM) due to their high level of hydration, their large size, and their relatively undifferentiated cytoplasm. This is particularly true for immunoprobe localization which has led to a dependence on light microscopic (LM) techniques, such as immunofluorescence. This study presents correlative LM and EM data to describe an example of the failure of light microscopy to correctly predict the ultrastructure of one particular organelle. Immunoprobe localization of centrosome and microtubule organizing center (MTOC) antigens in the mammalian egg was made by immunofluorescence and post-embedding immuno-EM, with best EM results achieved in Lowicryl-embedded material. Centrosome and MTOC antigens were detected by 5051 and an antibody to gamma tubulin (γtubulin). Gamma tubulin is a highly conserved element of MTOCs in many species and, thus, is highly diagnostic for them; it is also considered essential for microtubule (MT) nucleation. Results indicate that prior to nuclear breakdown, 5051 antigens and γ-tubulin are found exclusively in a type of “organelle,” the multivesicular aggregate (MVA), that bears no resemblance to MTOCs at the ultrastructural level. Until recently, the MVA was considered an organelle without a known function, while standard MTOCs were presumed to be the entities that carry the proteins recognized by centrosome antibodies. LM localization of centrosomal antigens carried the presumption that standard MTOCs were the entities labeled. Whether or not other molecules are shown to co-localize to these MVA, the presence of γ-tubulin supports the contention that MVA, or their contents, serve as centrosomal precursors with a unique ultrastructure. Thus, dependence on LM techniques alone can lead to erroneous conclusions on organelle identity and function.