GENERAL CONSIDERATIONS. 359 



Limulus by complete invagination. The limb was not onlj^, according to this 

 view, invaginatecl as a whole, but each single branchial lamella was separately 

 invaginated, the limb, as it were, growing into the body instead of growing out, 

 so that the interstices between the lamellae then became the lamellae of the 

 book-lung. This view somewhat resembles that of Kikg.sley. It appears to us 

 that MacLeod's view is the simplest explanation, and the one best agreeing with 

 the facts. MacLeod (No. 21) starts from the assumption that the lamellae of 

 tlie book-like gills are homologous M-itli those of the book-like lungs. The gill- 

 liearing limbs of Limulus are usuall}^ closely jiressed against the ventral side of 

 the abdomen. The brancliial lamellae develop only on those ujiper surfaces 

 which are pressed against the body. The ventral side of the limbs, in Limulus, 

 already shows a depression corresponding to the branchial lamellae. If we 

 imagine the respiratory limbs shifted further apart than they are in Limulus, 

 and the edges of the depression just alluded to fused with the edges of the leaf- 

 like limb, a closed space, the lung-sac, will thus be formed. Tlie free posterior 

 edge of the limb would then become the anterior margin of the stigma belonging 

 to this sac. By this assumption, MacLeod is able to explain certain features in 

 the structure of the Arachnid lung : eg., the facts that some of the lung-lamellae 

 are free not only at their posterior edges, bi;t at their lateral edges also, and that 

 tlie corresponding lung-sacs of the two sides are connected, etc.* 



The agreement in the internal anatomy of the two forms is no less 

 remarkable than that in the outer segmentation of the body and in 

 the structure and functions of the limbs. The presence of an 

 cndosternum has already been mentioned. We shall here merely 

 recall to mind the large, l)ranched liver, opening through several 

 efferent ducts into the intestine, the retiform rudiment of the genital 

 glands, the presence of a circum-oesophageal, arterial vascular ring- 

 accompanying the oesophageal commissure (and in Limulus developing 

 into an actual vascular sheath), and, finally, the presence of a gland 

 (brick-red gland of Linnilus, coxal gland) on the coxae of the fifth 

 pair of limbs (third ainbulatory limbs). 



The agreement which we have thus pointed out in the structure 

 and development of Limulus and of the Arachnida is so remarkable 

 that we can hardly avoid the conclusion that the two forms are 

 genetically related. We therefore accept the view that the 

 Arachnida have developed from the Palaeostraca through adaptation 

 to terrestrial life. 



It may be further mentioned here that adaptation to life in fre.sh water, and 

 ])Ossibly on land (?) had, perhaps, taken ]ilace even in the Gigantostraca them- 

 selves. According to Zittel (Xo. 7), they are found in the coal formations 

 associated with land plants, scorpions, insects, fish, and fresh water amphibia. 



* [Bkrnaed (App. to Lit. on Trilobita IV.), on the other hand, has endea- 

 voured to show that the lung-books of Scorpio developed in situ, as adaptations 

 to the circulatory system. He further claims that the Arachnids originally had 

 more stigmata than there were respiratory limbs in Limulus. — Ed.] 



