PCECILOPODA. 537 



central lobe forms a well-marked keel, and at the line of insertion of the 

 rim-like edge of the lateral lobes are placed the two pairs of eyes (st and 

 ft). The abdominal region is also distinctly trilobed and divided into nine 

 segments ; the last, which is merely formed of a median process, being the 

 rudiment of the caudal spine. The edges of the second to the seventh arc 

 armed with a spine. The changes in the appendages are not very con- 

 siderable. The anterior pair nearly meet in the middle line in front 01 

 the mouth ; and the latter structure is completely covered by an upper 

 lip. Each abdominal appendage of the second pair is provided with four 

 gill-lamella;, attached close to its base. 



Three weeks after hatching an ecdysis takes place, and the larva passes 

 from a trilobitc into a limuloid form. The segmentation of the abdomen 

 has become much less obvious, and this part of the embryo closely resem- 

 bles its permanent form. The caudal spine is longer, but is still relatively 

 short. A fourth pair of abdominal appendages is established, and the first 

 pair have partially coalesced, while the second and third pairs have become 

 jointed, their outer ramus containing four and their inner three joints. 

 Additional gill-lamelku attached to the two basal joints of the second and 

 third abdominal appendages have appeared. 



The further changes are not of great importance. They are effected in 

 a series of successive moults. The young larvae swim actively at the 

 surface. 



Our, in many respects, imperfect knowledge of the development of 

 Linuilus is not sufficient to shew whether it is more closely related to the 

 Crustacea or to the Arachnida, or is an independent phylum. 



The somewhat Crustacean character of biramous abdominal feet, etc. 

 is not to be denied, but at the same time the characters of the embryo 

 appear to me to be decidedly more arachnidan than crustacean. The 

 embryo, when the appendages are first formed, has a decidedly arach- 

 nidan facies. It will be remembered that when the limbs are first formed 

 they are all post-oral. They resemble in this respect the limbs of the 

 Arachnida, and it seems to be probable that the anterior pair is equivalent 

 to the chelicerae of Arachnida, which, as shewn in a previous section, are 

 really post-oral appendages in no way homologous with antennae 1 . 



The six thoracic appendages may thus be compared with the six 

 Arachnidan appendages ; which they resemble in their relation to the 

 mouth, their basal cutting blades, etc. 



The existence of abdominal appendages behind the six ccphalothoracic 

 does not militate against the Arachnidan affinities of Limulus, because in 

 the Arachnida rudimentary abdominal appendages are always present in 

 the embryo. The character of the abdominal appendages is probably 



1 Dohrn believes that he has succeeded in shewing that the first pair of appendages 

 of Limulus is innervated in the embryo from the supra-cesophageal ganglia. Hb 

 observations do not appear to me conclusive, and, arguing from what we know of the 

 development of the Arachnida, the innervation of these appendages in the adult can be 

 of no morphological importance. 



