October 8, 1 8y6] 



NA TURE 



557 



ilesciipiion is possible even as low elown as I'clromyzon, but when 

 •ve pass to the Ammocretes we find the arrangement of the 

 brancliial chamber has become so different that it is no longer 

 possil)le to describe it in terms of gill pouches. The nature of 

 the branchial chamber is seen in Kig. 3, which demonstrates 

 clearly that the IXth and Xth nerves supply a series of 

 separate gill-bearing structures or appendages, which hang 

 freely into a common respiratory chamber ; each one of these 

 appendages is moved by its own separate group of branchial 

 muscles, and possesses an external branchial bar of cartilage, 

 which, by its union with its fellows, contributes to form the 

 extra-branchial basket-work so characteristic of this primitive 

 respiratory chamber. The segmental branchial unit is clearly in 

 this case, as Ratlike originally pointed out, each one of these 

 suspended gills, or rather gill-bearing appendages ; it is 

 absolutely unnatural, as Xcstler (Anhi-o f. Naturgcsthich., 

 56, vol. i.) attempts to do, to take a portion of the space 

 between two consecutive gills and call that a gill-pouch. It is, 

 to my mind, one of the most extraordinary and confusing 

 conceptions of the current morphology to describe an animal in 

 terms of the spaces between organs, rather than in terms of the 

 organs by which those spaces are formed. We might as well 

 speak of a net as a number of h les tied together with string. 

 -Another most striking advantage is obtained by considering the 

 segmental unit to be represented by each of these separate 

 branchial appendages — viz. that we can continue the series in 

 the most natural manner (as seen in Fig. 3) in front of the limits 

 of the IXth and Xth nerves, and so find a series of appendages 

 in the oral chamber serially homologous with the branchial 

 appendages. The uppermost of the respiratory appendages is 

 the hyo-bianchial, supplied by the Vllth nerve, then, pa.s.sing 

 into the oral chamber, we find a series of non-branchial 

 appendages, viz. the velar and tentacular appendages, supplied 

 by branches of the Vth nerve. In fact, by simply considering 

 the tissue l>etwcen the .so-called gill-pouches as the segmental 

 unit, we no longer get lost in a maze of hypothetical gill pouches 

 in front of the branchial region, but find that the resemblances 

 between the oral and branchial regions, which have led to the 

 endless search for gill-slits and gill-pouches, really mean that 

 the oral chamber contains appendages just as the branchial 

 chamber, but that the former were not gill-bearing 



The study of Ammocoites, then, leads directly to the con- 

 clusion that the ancestor o( the vertebrate possessed an oral or 

 prosomatic chamber, which contained a series of non-ljranchial, 

 tactile and masticatory appendages, which were innervated from 

 the fused prosomatic ganglia or hind brain, and a branchial or 

 mesosomatic chamber, which contained a series of branchial 

 appendages which were innervated from the fused nieso.soniatic 

 ganglia or medulla oblongata. These two chambers did not 

 originally communicate with each other, for the embryological 

 evidence shows that they are separated at first by the septum of 

 the stomatodacum, and also that the oral chamber is formeil by 

 the forward growth of the lower lip. 



The phylogenetic test on the side of Limulusand its congeners 

 agrees in a remarkable manner with the conclusions derived 

 from the study of Ammocretes, for we see that the variation 

 which has occurred in the formation of Eurypterus from 

 Limulus is exactly of the kind necessary to form the oral and 

 branchial chambers of the Ammocwtes. Thus, we find with 

 respect to the me.sosoniatic appendages that the free, many- 

 jointed appendages of the crustacean become converted into the 

 plate-like appendages of Limulus, in which the separate joints 

 are still vi.sible, but insignificant in cr)mparison with the large 

 branchia'-bearing lamella ; then conies the in-sinking of these 

 appendages, as described by .Macleod (Arc/iiv de Bioiogif, vol. 

 v., 1884) to form the branchial lamella.', or so-called lung-books 

 of Thelyphonus, and the branchia.- of Kurypterus, in which all 

 semblance of jointed and free appendages disappears and the 

 V)ranchix- project into a series of chamb. rs or gill-pouches, each 

 pair of which in Thelyphonus open freely into communication. 

 In this way we see already the commencement of the formatii>ii 

 of a branchial chamber similar to that of .Ammoccetes. 



Si> alsr)with the innervation of these mesosomatic appendages, 

 originally a series of separate mesosomatic ganglia, each of 

 which innervates a separate a]ipendage ; then a process of 

 cephalisation takes place, in consequence of which, in the first 

 place, a single ganglion, the opercular ganglion, fuses with the 

 already fused prosomatic ganglia, as is seen in the stage of 

 Limulus ; then, as pointed out by Lankcster, in the different 

 groups of scorpions more and more of the mescsomalic ganglia 



NO. 1406, VOL. 54] 



fu.se together, and so we find the upward variation in this group 

 is di.stinctly in the direction of the formation of the medulla 

 oblongata roincidently with the formation of a branchial 

 chamber. 



In a |)reci.sely similar way, we find the variation which has 

 occurred in the prosomatic apjiendages leads directly to the 

 formation of the oral chamber and oral appendages of 

 .\mmocietes ; for the original chelate and locomotor append- 

 ages o[ Limulus become ci)nverted into the tactile non-chelate 

 appendages of Eurypterus (if. Figs. 4 and 5), and the small 

 chilaria (M) of Limulus, according to Lankester, fuse in the 

 middle line and grow forward to form the metastoma ot 

 Eurypterus, thus forming an oral chamber, into which the short 

 tactile appendages could be withdrawn, closely similar in its 

 formation to the oral chamber of Ammocceles. The prosomatic 

 ganglia supplying these oral appendages have already, in Limulus 

 (see Fig. 4), been fused together to form the infra-tesophageal 

 ganglia or hind brain. 



The phylogenetic te.st, then, both on the side of the 

 vertebrate and of the invertebrate, points direct to the 

 conclusion that the peculiarities of the trigeminal and vagus 

 groups of nerves are due to their origin from nerves supplying 

 prosomatic and mesosomatic appendages respectively. 



2. The anatomical test confirms and emphasises this con- 

 clusion in a most striking manner, for we find not only coin- 

 cidence of topographical arrangement, as already mentioned, but 

 also similarity of structure ; thus we see that the blood in the 

 gill lamella and velar appendages of .Ammocietes does not cir- 

 culate in distinct capillaries, but, as in the arthropod appendages, 

 in lacunar spaces, which by the subdivision of the surface of the 

 appendage to form gill lamelhe become narrow channels ; that 

 also certain of the branchial muscles and of the muscles of the 

 velar appendages are of the invertebrate type of so-called 

 tubular muscles. These invertebrate muscles are not found in 

 higher vertebrates, but only in Animoccetes, and moreover 

 disappear entirely at transformation. 



Origin of the Vertebrate Carlilaginoiis Skeleton. 



Perhaps, however, the most startling evidence in favour of 

 the homology between the branchial segments of Ammocretes 

 and the branchial appendages of Limulus is found in the fact 

 that a cartilaginous bar external to the branchia; exists in each 

 one of the branchial appendages of Limulus, to which some of 

 the branchial muscles are attached in precisely the same way 

 as in Ammoccetes. The branchial cartilages of Limulus (see 

 Fig. 4) spring from the entapophyses and form strong cartilag- 

 inous bars, which are extra-branchial in position, just as in 

 Ammoccetes ; in addition to each branchial bar, a cartilaginous 

 ligament passes from one entapophysis to another, so as to form 

 a longitudinal or entapophysial ligament, more or less cartilag- 

 inous, which extends on each side along the length of the 

 mesosonia. In precisely the same way the branchial bars ot 

 .Ammoccetes are joined together along each side of the 

 notochord by a ligamentous band of more or less continuous 

 cartilaginous tissue, forming a subchordal or parachordal carti- 

 laginous ligament. 



Further, we see that this cartilage of Limulus is of a very 

 striking structure, quite different from that of vertebrate 

 cartilage, and that it is formed in a fibro-massive tissue which, 

 like the matrix of the cartilage, gives a deep purple stain with 

 thionin, thus showing the presence of some form of chondro- 

 mucoid. This fibro-massive tissue is closely connected with the 

 chitinogenous cells of the entapophyses. 



Startling is it to find that the branchial cartilages of Ammo- 

 cretes possess identically the same structure as the cartilages ot 

 Limulus ; that the branchial cartilages are formed in a fibro- 

 ma.ssive tissue which, like the matrix of the cartilage, gives a 

 deep purple stain with thionin, and that this fibro-massive tissue, 

 to which Schneider (" Beilriige z. Anat. u. Entwicklungsgesch. 

 der Wirbelthiere," Berlin, 1879) gives the name of muco- 

 cartilage, or Vorknorpel, entirely di.sappears at transformation. 



Further, according to Shipley (Quart. Journ. of Micr. Sci., 

 1887), the cartilaginous skeleton of the Ammocretes when first 

 formed consists .simply of a series of straight branchial bars, 

 springing from a .series of cartilaginous pieces arranged bilaterally 

 along the notochord. 



The formation of the trabecule, of the auditory capsules, of 

 the crossbars to form the branchial basket-work, all occur 

 subsequently, .so that exactly those parts which alone exist in 

 Limulus are those parts which alone exist at an early stage in 



