Most tintinnid species have a shortest linear dimension < 50 μm. Hence, a priori, nets of mesh sizes ≥ 50 μm will likely under‐sample most tintinnid species. However, studies often appear (23 since 2015) using sampling with nets of meshes sizes ≥ 50 μm, reporting both tintinnid concentrations, and community composition. How biased are results from using coarse mesh nets? We provide a comparison of whole water vs. net sampling based on fortuitous, that is, unplanned, parallel sampling. Pairs of samples from a standard monitoring station in the Bay of Villefranche (N.W. Mediterranean Sea) taken on 44 dates from 2013 to 2018 were compared. Tintinnids were enumerated in settled material from a water column sample, an integration of six discrete depth samples between 5 and 70 m, prepared for analysis of phytoplankton composition and in material from a plankton net (52 μm mesh) tow from 70 to 0 m, taken the same day. Despite the large confidence limits due to low raw cell counts from whole water samples, cell concentration estimates were about an order of magnitude higher than those from plankton net samples and frequently biomass estimates as well. Community composition also differed. The most common species in whole water samples were small (diameter ≤ 20 μm), and some common forms were absent, or nearly, from the net samples. We show that, while valuable for collecting larger and rarer species, coarse net samples do not yield robust estimations of overall concentrations, nor allow identification of the dominant tintinnid species.