No. I.] STUDIES ON LIMULUS. 17 



segmental cartilages found in corresponding regions in primi- 

 tive vertebrates, and which give rise to the vertebral column. 

 The cartilages serve evidently as points of attachment for seg- 

 mentally arranged muscles that meet at those points. But we 

 must not conclude that that is the only reason why the carti- 

 lages are there, because cartilages are not always formed at 

 points where muscles are attached to one another. It is obvi- 

 ously impossible to carry the comparison with vertebrate carti- 

 lages any farther. It is enough for our purpose to show that 

 the conditions in Lhmiliis are such as to produce a series of 

 segmentally arranged cartilages about the spinal cord, similar 

 to those in primitive vertebrates. 



III. There are seven pairs of branchial cartilages composed 

 of an entirely different tissue histologically (and chemically, 

 also, according to Gaskell) from that in the endosternite and 

 the segmental cartilages. 



The first pair arise from the inner surface of the chilaria, and 

 are attached to the posterior margin of the endocranium so that 

 they appear to form a part of it. The remaining six pairs arise 

 from the base of the abdominal appendages and go to the cor- 

 responding entopophyses. They develop, as will be shown in 

 one of the papers of this series, at a very early embryonic period 

 as clearly defined outgrowths of the walls of the mesoblastic 

 somites. Their union with the epidermis is secondary, and 

 they are in no wise derived from the modification of chitinous 

 ingrowths from the epidermis, as maintained by Gaskell. 



These branchial bars of the mesothorax correspond, we 

 believe, to the cartilaginous bars of the visceral clefts of verte- 

 brates, as we first indicated in '89. In '93 we called attention 

 to the surprising histological resemblance between these carti- 

 lages and those of Petromyzon, and still later ('96) we showed 

 that there was a certain number of embryos in which one or 

 more pairs of appendages were invaginated instead of evaginated. 

 Transverse slits were thus formed along the sides of the head, 

 resembling vertebrate gill slits and recalling to mind the lung 

 books of scorpions and spiders. In Limulus the appendages 

 most frequently invaginated, the thoracic ones, are not pro- 

 vided with gills. This again is suggestive since in vertebrates 



