George L. Streeter 109 



might account for the {liffereut arrangements found by Holl. A further 

 source of variation is presented by slight differences in the division line 

 between the rootlets of the hypoglossus and Hie first cervical; the fibres 

 destined for the r. desccndens, for instance, may be either picked up with 

 the more caudal rootlets of the hypoglossal, when there will be little or no 

 communication between the hypoglossal and the first cervical, or on the 

 other hand may be picked up with the first cervical and reach their 

 destination through anastomosis witli the hypoglossal. Thus in embryo 

 No. 144 of the Mall collection (Plate II) on the right side the first cer- 

 vical contributes no fibres to the hypoglossal and descendens, while ou the 

 left side a large communicating bundle exists between them. 



In the early stages the rootlets of the hypoglossal present a close simi- 

 larity to the ventral roots of the spinal nerves, and now are generally 

 considered as a cranial continuation of them ; the nerve being thus derived 

 from the fusion of three or four segmental spinal nerves, which in the 

 course of phylogenesis have become enclosed in the cranium. In the 

 hypothetical ancestor the segments of the nerve belonged to the trunk, 

 and possessed, in addition to the ventral roots, both dorsal roots and 

 ganglia, the latter becoming subsequently reduced coincidently with the 

 invasion of the vagus group into this region. Strong support to this 

 view was given by Froriep, 83, who in the hoofed animals found persist- 

 ent dorsal roots and ganglia belonging to one or two of the more caudal 

 divisions of the nerve. Similar precervical ganglion masses and rootlets 

 Avere found in the rabbit, cat, and mouse by Martin, 91, and Robinson, 93. 

 The former describes five hypoglossal ganglia in cat embryos, of which 

 he finds only the most caudal one to persist. He thus apparently 

 includes those that in our series of reconstructions are considered as 

 accessory root ganglia, which we think have a different phylogenetic signi- 

 ficance. In the human embryo His. 88, describes an abortive precervical 

 ganglion, and names it after Froriep. Inasmuch as he considers the 

 hypoglossus to belong phylogenetically to the vagus rather than to the 

 spinal nerves, he is inclined to doubt a relation between the Froriep 

 ganglion and the hypoglossus. In our reconstructions a typical ganglion 

 may be seen in Figs. 7 and H. The former is the same embryo 

 pictured by His, and does not essentially differ. On the other side of this 

 embryo. Fig. 8, the first cervical ganglion creeps forward a short distance 

 along the accessorius tract, and thus represents what may be styled as a 

 precervical tendency. An interesting case is shown in Fig. 8, where the 

 first cervical ganglion is divided in two equal parts, each having its own 

 ventral root. With further growth they would have become separated, 

 as the spinal ganglia do, and then we should have in the more oral one 



