Figure 8. Parargornis messelensis, holotype (HLMD Be 193), uncoated specimen to show the excellently preserved feathering; scale bar equals 10 mm.
protrudes ventrally.The caput humeri directs caudally; like in Argornis it lacks the distal enlargement which is characteristic for Jungornis. The crista deltopectoralis is large and tapering (visible on the left side of HLMD Be 163). As in Jungornis and Argornis, the processus supracondylaris dorsalis is weakly developed (Fig. 5). It is situated in a similar relative position to that of Argornis, about one-quarter of the length of the humerus from the distal end. Whether there was a second, distal process as in Argornis and Cypselavus cannot be said with certaint , since the area directly below the (?proximal) processus supracondylaris dorsalis is broken (the corresponding
part of the humerus is only visible on the right side of HLMD Be 193). However, if a second process was present, it must have been situated directly next to the one that is preserved (in Argornis the two processus are also situated close together, whereas they are more distant in Cypselavus). The condylus ventralis is elongated, the tuberculum supracondylare ventrale is in a similar relative position to Argornis. The sulcus humerotricipitalis is not as deeply excavated as in recent Apodidae.
As in recent swifts, the ulna greatly exceeds the humerus in length (Fig. 6), whereas in humming- birds it is as short as the humerus. As far as is