Nanoviruses are multipartite ssDNA viruses transmitted by aphid vectors. Their mode of transmission is believed to be of the circulative non-propagative type [1, 2], but a recent analysis of the relative frequency of the distinct genome segments in aphids versus in plants indicated a more complex relationship [3]. Previous microscopy studies using immunofluorescence against the coat protein of Banana bunshy top virus (BBTV) confirmed that the virus is detected solely in gut and principal salivary gland cells of its aphid vector Pentalonia nigronervosa [4]. However, the subcellular localization and the cell compartment(s) with which the virus associates during transcytosis could not be determined [5]. Another unresolved question is whether, as demonstrated in monopartite viruses, nanoviruses are submitted to strong bottleneck during aphid transmission (see the abstract by Romain Gallet). In that case, when the virus particles traverse cellular barrier, different particles containing different segments may be separated, increasing the chance of loosing genomic information (loosing segments) and thus aborting transmission. We have revisited the cycle of a nanovirus, the Faba bean necrotic stunt virus, within the body of its aphid vector, Acyrthosiphon pisum, using a combination of FISH (Fluorescent In Situ Hybridization), immunolabelling, confocal and electron microscopy. FISH labeling of viral DNA indicates that the virus massively accumulates in all cells of the midgut but is virtually absent in all other regions of the gut. At the salivary glands, the virus is very much restricted not only to the principal salivary glands but also to a single cell type within these glands. No signal could ever be detected in any other organ. When using DNA-FISH probes specific to distinct segments, we could demonstrate that all segments are always colocalized within the midgut, and to a lesser extent also within the salivary gland cells. This observation contrasts with the situation