Azaspiracids (AZAs) are a group of lipophilic polyether toxins implicated in incidents of shellfish poisoning in humans, particularly in northern Europe. In an attempt to establish the biogeographical distribution of AZA toxins, their association with plankton size-fractions, and to confirm the identity of the causative species responsible for human poisoning, a month-long oceanographic study was undertaken in coastal North Sea waters. The occurrence and abundance of AZA analogues was measured by on board triple quadrupole mass spectrometry coupled to liquid chromatography (LC-MS/MS). In size-fractionated plankton samples collected by net tows (20 μm mesh-size), by pumping from discrete depths and from Niskin entrapment bottle casts to fixed depths, AZA-1 was consistently the major azaspiracid component. In eastern Scottish coastal waters, the highest amounts of AZA-1 in net tow samples were in the 50-200 μm fractions, with lesser amounts detected in the >200 μm and 50-20 μm fractions. At these stations, the 50-200 μm fractions were rich in the ciliate Favella ehrenbergii. Cells of F. ehrenbergii isolated by microcapillary indeed contained AZA-1, but isolated cells grown and fed the non-toxic dinoflagellate Scrippsiella trochoidea for one week failed to contain any detectable AZA-1-evidence that F. ehrenbergii is merely a vector for AZA. Detailed analysis of plankton from Niskin bottle samples from around the North Sea typically showed highest amounts of AZA in the 3-20 μm fraction. From this fraction, a large number of crude cultures were established and subsequently screened for the presence of AZAs. A small photosynthetic thecate dinoflagellate, provisionally designated as strain 3D9, was isolated by microcapillary and brought into pure culture. This dinoflagellate strain was found to produce AZA-1, AZA-2 and an isomer of AZA-2. Sequence comparisons by molecular genetic techniques also indicated that this genotype was present in field samples rich in AZA. This discovery of a novel causative dinoflagellate for AZA toxicity essentially explains the lack of correlation of AZA with the abundance and distribution of the previously postulated culprit species Protoperidinium crassipes. We instead propose that such large phagotrophic dinoflagellates can act as an AZA vector following grazing upon a proximal source, such as the dinoflagellate 3D9 strain. © 2008 Elsevier B.V. All rights reserved.