MK.M()I1{S Ol' THK NATIONAL ACADK.MV Ol' StIENlM'iy. 319 



by IIk' wnttT, imist li;ivt' siiimltiint'ously Ih'coiiic sokU'iod liy tlicir od^'cs to the walls oC tin- ;rill- 

 ohiiiiibi-r. 



For our earliest iiit'orinatioii reiianliiij;- tke mode of development ol'tlie Imi^s of spiders we are 

 iudebted to S«liiiiike\vitsoli (I8S5). He states: 



Mnllbiir liiis sliowii ■ that tlie trachoir iiu<l liinKK of Arauoina are formed by the invagination of tlio octnilorni. I 

 am unal>l<> to conlirin this oliHorvation. In tlio ombryo of Lj/roiaiiaccala, licfon- hutching, 'the traclioii' arc ri-prcMcntod 

 l>y a [>rin<-i]ial trunk from whii'li ariso four secondary branches. Their wall is formed of the epithclia] cctoilcrmic 

 hiyer lined by a honiogcneons internal coat and covered by an external coat containing nuclei of mesodermal origin 

 (V, chap. III). In this same stage the strncturo of the lungs is very interesting (Tl. XXII, I'ig. 1, pm). A. Milne- 

 Edwards an<l J. MacLeod have endeavored to prove the homology of the lungs of Arachnida with the gills of Limuli. 

 The lungs of the embryo of Li/cosn Mccala I'onsist of true trachea- arranged in bundles; the principal trunk (I'l. XXII, 

 Kig. 1), which opens outward by the stigmatic orilico ("O. divides into live secondary branches. The e|)ithelial 

 layer of these last is represented by llattened colls. The secondary branches, flattened from above downward, cnti-r 

 the lavity surrounding the lungs, and in the spaces between these are to be seen the blood coriuiscles. Thus the em- 

 bryonic lung consists of a trachea arranged enfaiseeaii; the only difference to bo noted is the absence of the external 

 tunic, which I have notobserved, either on the lung of the embryo or on that of the adult (p. .561). 



Ill the following year, in his "Les Araebnides et leiirs Attinites" (ISSfi), he reiterates this view. 



In the same year, Loey, in his "Observatious on the development of A^elena H/fyJa," jjives a still 

 more eomi)Iete aeeouut of the mode of orij^in of the lungs of spiders, with excellent figures. He 

 states that the lungs arise as a pair of extensive iuvaginatious at about the same time as the proc- 

 toda-um. '' In sagittal sections of early stages the lungs appear as oblong plates of cells, the large 

 oval nuclei of which are arranged in parallel rows." The flattened sacs are hollow and divided at 

 intervals by 2eell eolnmns, the intersijacea sometimes containing blood corpu.seles. The details are 

 mostly histological and the author does not enter into any morphological comparison with the book- 

 lungs of Limnlns, but nothing is .said contrary to the view that the lungs are a bundle of trache:e. 



In his paper "On the origin of vertebrates from Arachnids " (1890), Tatten refers to the chitin- 

 lined tubes of the scorpion, " which serve for the support of muscles." These, as we understand 

 them, are the "muscular stigmata" of Lankester aud occur in insects as well as in Arachnida and 

 Podostomata, as we have ob.served them in a specimen of Asaphus as well as in Limiilns. As 

 I'atten remarks: "The chitinized tubes are comi)arable with the three or four pairs of tracheal 

 invaginations which in insects give rise to the tentorium." He then adds : 



I regard the lung-books of Scorpio anil the chitin-liued tubes described above as belonging to the same category, 

 for aft«r careful study I have foBud nothing to in<Ucate that they arise as modifications of rudimentary abdominal 

 .ippendages (p. 355). 



In 188.J Lankester discards his first hyjiothesis, which he regards as "overstrained," aud 

 replat'cs it by a second "perfectly simple" one. 



In his notes on the embryology of Limulus (18S5) Kingsley describes the early stages of the 

 ab<lominal limbs and discusses at some length their homology with the "pulmonary books" of 

 Arachnids, and he admits its general validity, though not ready "to follow all of Prof. Lankester's 

 intermediate steps, nor those of MacLeod." His hypothesis and figures are ingenious and more 

 simple than tho.se of Lankester. As he remarks: 



As I have mentioned above, tlie process of formation of the gill-leaves is largely by a process of outgrowth, 

 but there is also a slight ingrowth, espi'cially noticeable at the distal portion of the appendage. This, however, disl 

 appears with growth, but is very noticeable in all my sections. To transform the gill of Limulus into the liii'ig of 

 Scorpio it is only necessary that, together with the sinking of the whole organ, as described above, the inpushings 

 of the integument to form the lamellae should be e.^aggerated and the outgrowth CT)rrespondingly decreased. On 

 PI. xxxvn, Figs. 18 to 20, I have diagramuiatically illustrated the steps in the process, the gill-le.-ives being few in 

 number to secure clearness. In 18 wo have the typical condition fouml in Limulus, one appendage being shown half 

 in section and half in perspective. In 19 we have an intermediate cimdition, when, as suggested above, the auintal 

 was leaving the water and seeking a terrestrial life. Here the gill-bearing appendage (ga.) is partially sunk in the 

 surrounding tissues to secure protection. The same causes would also tend to produce a similar change in the gill 

 leaves (.111.) and they would also tend to be lornied rather as ingrowths than as protruding proces.sea. This change 

 in structure would be the more readily eflectod on account of a change of the medium of respiration. A gill needs 

 either to project into the water or to have some means of instantly changing the fluid which bathes it. An org.in 

 for aerial respiration, on the other hand, is not so restricted in its misition, since the air is more fluiil and more 

 elastic and hence more readily changed. Another ailvantage to the animal resulting from the change is that the 

 oxj'gen is thus brought nearer to the tissues rei|uiring it. 



'I can find no reference in Halfour's memoir to this subject. 



