were especially active in aiding the movement of the following seventh and eighth segments which form the oar blade. The seventh segment is a flat trapezoidal plate possessing on the posterior side a triangular lobe, that served as a guard to the eighth. The latter is oval and has a terminal notch, in which a minute claw is inserted that represents a ninth segment and was termed the palette by Hall, while later authors have more properly applied the term palette to the entire eighth segment.
The species which seem to us of especial interest in the explanation of the strange structure of the swimming leg are Eurypterus (Onychopterus) kokomoensis and Dolichopterus macrochirus. The former [pl. 26, fig. 2] shows the most primitive shape of the swimming leg known to us. This primitive character manifests itself most distinctly in the following features. Beginning at the distal end, the ninth segment is a well developed terminal claw, such as is found on the other legs, indicating that the minute ninth segment of Eurypterus is a reduced terminal claw. The seventh and eighth segments do not yet so exactly fit together into a single oar blade, the seventh being still narrower and the eighth more expanded. The triangular guard lobe of the seventh segment in Eurypterus is here represented by a long relatively narrow lobe, indicating that it originated from a broadened and flattened spine of the seventh segment. A glance at figure 26, plate 2, will show that the preceding joints also are still much more uniform in character and like those of the walking legs in Eurypterus. In Dolichopterus macrochirus [see restoration, pl. 40] on the other hand, the ninth segment has been developed into a third element of the oar blade which thereby has become still more powerful. Here we find triangular guard plates on both the preceding, the seventh and eighth segments, which by their form still distinctly indicate their origin from spines. The lobelike projecting anterior portions of the distal edges of the seventh and eighth segmentsalso suggest a similar origin, especially if we compare them with the lobelike spines of the fourth pair of postoral limbs, already noted.
The postoral limbs performed a still further function, viz, mastica-
