raymomd: new and old Silurian trilobites. 11 



of the species in this family, the other genera having, as a usual thing, 

 only from one to two or three species each. Bumastus comes next to 

 lUaenus in the number of species, and has its greatest development in 

 America where there are at least six species in the Middle Ordovician 

 and about fifteen in the Middle Silurian, as contrasted with three or 

 four species in the Silurian of Great Britain, about the same number in 

 Scandinavia and Russia, and two in Bohemia. 



Considering the great abundance of the illaenids, we have sur- 

 prisingly little information as to their ancestry or relationships. 

 Following the usual theory, which seems to be borne out by the facts 

 in most cases, one would expect these smooth forms to be the descend- 

 ants of more normal trilobites with glabellar furrows and with ribs 

 on the pygidium. But among all the illaenids there does not seem to 

 be one which shows any trace of ribs on the pygidium, while only a 

 few show indications of glabellar furrows. And such indications of 

 furrows as exist are merely spots or slight depressions on the smooth 

 glabella. The nearest relatives of the Illaenidae are undoubtedly the 

 Goldiidae (Bronteidae) not the Asaphidae, with which family they 

 have usually been classed. The presence in both the Illaenidae and 

 the Goldiidae of an epistoma, similar hypostomas, forward expanding 

 glabella, large eyes which are placed far back, unfurrowed pleura in 

 the thorax, and short axial lobe on the pygidium, indicate a very close 

 relationship, some of these characteristics being apparently too funda- 

 mental to admit of explanation on the ground of parallelism. The 

 Goldiidae, in spite of their specialization, are more like the typical 

 trilobite than the Illaenidae, and it would be natural to place them in 

 the ancestral position. The geological range at once negatives this 

 attempt, for the Goldiidae did not appear until the Middle Ordovi- 

 cian, are very rare in the Ordovician and reached their greatest de- 

 velopment in the Silurian and Devonian. The illaenids, on the other 

 hand, appeared in the basal Ordovician, possibly even in the Cambrian, 

 reached their greatest development in the Middle Silurian and did 

 not survive that period. That the Goldiidae should have been de- 

 rived from the Illaenidae, however, seems highly improbable, for the 

 phylogeny of the former family pursues a normal course, the oldest 

 members of the family being most highly segmented, and the usual 

 ''smoothing out" process producing such (relatively) lUaenus-like 

 species as Goldius dormitzeri, G. campanifer, and G. hrongniarti in the 

 Devonian of Bohemia. If these species, without glabellar furrows 

 and with highly convex almost ribless pygidia occurred in the Ordo- 

 vician, and forms like Goldius hmatvs (Billings) in the Devonian, we 



