Diets of pterosaurs have mainly been inferred from indirect evidence such as comparative anatomy, associations of co-occurring fossils, and functional morphology. Gut contents are rare, and until now there is only a single coprolite (fossil dropping), with unidentified inclusions, known. Here we describe three coprolites collected from a palaeosurface with numerous pterosaur tracks found in early Kimmeridgian (Hypselocyclum Zone) intertidal deposits of the Wierzbica Quarry, Poland. The specimens’ morphology and association to the tracks suggest a pterosaur producer. Synchrotron scans reveal numerous small inclusions, with foraminifera making up the majority of the identifiable ones. Other small remains include shells/carapaces (of bivalves, ostracods, and other crustaceans/arthropods) and bristles (some possibly of polychaete worms). The high density of the small shelly inclusions suggest that they were not accidently ingested, but constituted an important food source for the pterosaur(s), perhaps together with unpreserved soft-bodied animals. The combined evidence from the tracks and coprolites suggest a filter-feeding ctenochasmatid as the most likely tracemaker. If true, this significantly expands the bromalite record for this pterosaur group, which was previously only known from gastroliths. Moreover, this study also provides the first direct evidence of filter feeding in Jurassic pterosaurs and shows that they had a similar diet to the recent Chilean flamingo (Phoenicopterus chilensis).

No abstract

The goals of a program for computational historical linguistics Jäger (p. 156) sets up four key principles for a computational historical linguistics: replicability, rigorous evaluation, separation of training and test data, and only