George L. Streeter 85 



ganglia of the roots. The latter ganglia though not segmental more 

 closely resemble -the spinal ganglia; the attempt however to reduce the 

 cranial nerves to a spinal nerve t3'pe is deprecated. Some sucli liypo- 

 thesis as was long ago suggested by Balfour, 76, is much easier of appli- 

 cation. This investigator supposed the head and trunk to have become 

 differentiated from each other when there was only a mixed motor and 

 sensory posterior root present and no ventral root, as was then supposed 

 to be the case in tlie amphioxus. Since then it has been found (Ransom 

 and Thompson, 86, llatscheck, 92, Dogiel, 03) that the amphioxus and 

 cyclostomes have ventral, purely motor, roots. These do not arise from 

 the cord at the same level with, nor do they join, the dorsal roots. The 

 two might be spoken of as ventral and dorsal nerves. If then Balfour's 

 h3rpothesis were modified to fit with our present knowledge, it could be 

 stated as follows : The head and trunk nerves were differentiated from 

 each other at a time when there existed mixed motor and sensory dorsal 

 roots and pure motor ventral roots. These ventral and dorsal roots did 

 not then arise from the neural tube at corresponding levels, and were inde- 

 pendent of each other. In the spinal region of higher vertebrates a modi- 

 fication has occurred, by which the dorsal and ventral roots have become 

 strictly segmeutally arranged, and have joined in pairs, each pair forming 

 a common nerve, which is situated median to the myotome. In the head 

 region the nerves have retained the primitive type; the dorsal roots still 

 contain a good proportion of motor fibres, and are situated beneath the 

 epidermis and outside of the myotomes. They are not segmentally 

 arranged and do not join with the ventral roots to form common nerves, 

 but form a system of separate ventral and dorsal (lateral) nerves. 



Material and Methods. — The elements of the peripheral nervous 

 system do not reach a degree of differentiation, wliich is sufficient for 

 reconstruction, until toward the end of the third week. From then 

 changes in the form and relation continue until the tliird month, when 

 the structures have practically reached the condition found in the adult, 

 and development may be considered as completed. The various stages 

 in their growth were thus found to be covered by the embryos listed in 

 the following table (page 86). Their ages have been determined by use 

 of Mall's rule (Mall, 03), ('. c. the age in days equals tlic square root of 

 the product of the length times one hundred. 



It was found that the ganglion masses and fibre paths could be satis- 

 factorily identified and traced by means of profile reconstructions. This 

 procedure was made use of with all embryos, with the exception of one at 

 a late stage which was large enough for dissection. The details adopted 



