418 Charles Emerson Beecher. 



the legs had been determined by slicing enrolled individuals. 

 Antennae, however, had not been clearly made out until 1893, 

 when their presence was announced in the August number of 

 this Journal. This discovery was of great value and promised 

 much toward a better understanding of the ventral anatomy of 

 trilobites and their systematic position among the Crustacea. 

 This led to Beecher's visiting the locality in 1893 to take out 

 several tons of the shale. Even as late as last fall he developed 

 from this material specimens of Trinucleus showing the ven- 

 tral appendages in the greatest detail. Since 1893 Beecher 

 has published fifteen papers on the trilobites. Of these three are 

 devoted to the larval stages, seven to the ventral anatomy, and 

 five to classification and the systematic position of these forms. 



The ventral anatomy is most completely known in Triarthrus, 

 "an active creature " belonging to an ancient Cambrian family. 

 Beecher showed that in this genus the entire series of thoracic 

 legs are biramous, one of them setae-bearing and used for 

 swimming (expodite), and the other without setae and used for 

 crawling (endopodite). The limbs of the pygidium overlap 

 each other, are much crowded, and are adapted for swimming 

 or guiding the animal, although they may also have served as 

 egg carriers. The individual segments " are considerably 

 expanded transversely, thus making a paddle-like organ." The 

 head has five pairs of appendages as follows : Anterior anten- 

 nae or uniramous antennules attached at the side of the hypos- 

 toma, followed by four pairs of biramous appendages closely 

 resembling the thoracic legs. These are (1) posterior antennae, 

 (2) mandibles, (3 and 4) maxillae. The ventral membrane of 

 Triarthrus "is of extreme tenuity" and is an " uncalcified, chi- 

 tinous, flexible pellicle, and thus was in strong contrast with 

 the much thicker and calcified dorsal test." 



The larval stages he studied in nine genera ranging from 

 the Cambrian to the Lower Devonian. He concluded that 

 " all the facts in the ontogeny of the trilobites point to one 

 type of larval structure." This larva, not more than one milli- 

 meter in length, is " characteristic of all trilobites, and among 

 different genera, varying only in features of secondary impor- 

 tance. This stage may therefore be called the jprotasjpisP 

 He found that Barrande's four orders of trilobite development 

 are but stages of his first order, and that Agnostus is " neither 



