Yol. 50.] SYSTEMATIC POSITION OF THE TRILOBITES. 417 



function as jaws, a specialization further developed in the Eury- 

 pteridae, in which the most posterior of these becomes the most 

 powerful. 



The great development of the glabella in many trilobites may 

 perhaps be due in some cases to the great development of the 

 oesophagus as a ' masticatory stomach,' or, again, of the mid-gut 

 diverticula ('liver'), which almost certainly occupied this part of the 

 body (cf. Limulus and Apus). 



My theoretical deduction of all Crustacea from an annelid in 

 which the anterior end was bent round ventrally, so as to allow of 

 its appendages to function as jaws, is thus fully confirmed by these 

 early trilobites. 



2. The head-shield seems to have been a characteristic of all the 

 earliest Crustacea. I endeavoured (in ' The Apodidse ') to explain it 

 as starting from the lateral projections 



which would be necessarily caused by Yig. 6.— Sao hirsuta, after 

 the sharp bending round of the first Barrande. 



segment. A careful study of the 

 series of under-surfaces of the heads 

 (especially of Dalmanites socialis and 

 Paradoxides bohemicus) figured by 

 Barrande a has confirmed me in this 

 supposition. [Early stage, showing lateral 



Still more conclusive evidence, projeetionsas due to bending 



however, on this point is yielded by of first segment.] 



the developmental history of Sao 



hirsuta, also given in Barrande's classical work. Stages 1-8 show 

 the first segment produced on each side into points curving back- 

 wards round the outer edges of the cephalic shield (see fig. 6). 

 In stage 9 this is nearly obscured, while the head- shield of the 

 adult is very highly specialized and shows no traces of its origin. 



The head-shield thus almost certainly originated in the first 

 segment, as a pair of lateral projections due to the sharp bend in 

 that segment. The backward growth of these projections, i. e. 

 their repetition on the following segments as pleurae, was a natural 

 process. 



In Microdiscus the head-shield extends backward through three 

 segments, the fourth segment being not yet quite incorporated into it. 

 When five segments became definitely fixed as the normal number 

 of head-segments, the head-shield ran back to the posterior edge 

 of the fifth segment. Not only, however, does this fifth segment 

 often appear like a trunk-somite, but the transverse strip of the 

 head-shield belonging to it very often appears, as above noted, to 

 be a pair of pleura? belonging to the trunk-segments, fused along 

 their anterior edges with the cephalic shield. 



This fact, namely, that the comparatively recent incorporation of 

 the pleurae of the fifth head-segment is still visible, helps us to under- 

 stand the morphology of the head-shield. As above suggested, we 



1 ' Systeme silurien de la Boheme,' vol. i. (1852) Trilobites, pis. 2 a and 2 b. 



