474 



NATURE 



[September ii, 1890 



have existed or may still exist in which the nerve tube fulfilled a 

 non-nervous function, and possibly acted as part of the ali- 

 mentary canal, a suggestion that has recently been revived in a 

 somewhat extravagant form. 



A passage of food particles into the alimentary cavity through 

 the neural tube has not yet ;been seen, and probably does not 

 occur, as the larva still possesses suflficient food yolk to carry it 

 on in its development. It is therefore permissible to hold that 

 the neurenteric canal may be a mere embryological device, and 

 devoid of any deep morphological significance. 



The question of variation in development is one of very great 

 mportance, and has perhaps not yet received the attention it 

 deserves. We are in some danger of assuming tacitly that the 

 mode of development of allied animals will necessarily agree in 

 all important respects or even in details, and that if the develop- 

 ment of one member of a group be known, that of the others 

 may be assumed to be similar. The more recent progress of 

 embryology is showing us that such inferences are not safe, and 

 that in allied genera or species, or even in different individuals 

 of the same species, variations of development may occur 

 affecting important organs and at almost any stage in their 

 formation. 



Great individual variations in the earliest processes of develop- 

 ment, i.e. the segmentation of the egg, have been described by 

 different writers. 



In Renilla, Wilson found an extraordinary range of variation 

 in the segmentation of eggs from which apparently identical em- 

 bryos were produced. In some cases the egg divided into two 

 in the normal manner ; in other cases it divided at once into 

 eight, sixteen, or thirty-two segments, which in different speci- 

 mens were approximately equal or markedly unequal in size. 

 Sometimes a preliminary change of form occurred without any 

 further result, the egg returning to its spherical shape, and 

 pausing for a time before recommencing the attempt to segment. 

 Segmentation sometimes commenced at one pole, as in telo- 

 lecithal eggs, with the formation of four or five small segments, 

 the rest of the egg breaking up later, either simultaneously or 

 progressively, into segments about equal in size to those first 

 formed ; while lastly, in some instances segmentation was very 

 irregular, following no apparent law. 



It is noteworthy that the variability in the case of Renilla is 

 apparently confined to the earliest stages, for whatever the mode 

 of segmentation, the embryos in their later stages were indis- 

 tinguishable from one another. 



Similar modifications in the segmentation of the egg have 

 been described in the oyster by Brooks, in Anodon and other 

 MoUusca, in Hydra, and in Lumbricus, in which last Wilson 

 has recently shown that marked differences occur in the eggs 

 even of the same individual animal. In the different species of 

 Peripatus there appear also to be considerable variations in the 

 details of segmentation. 



In the early embryonic stages after the completion of seg- 

 mentation very considerable variation may occur in allied species 

 or genera. Among Ccelenterates, for instance, the mode of 

 formation of the hypoblast presents most perplexing modifica- 

 tions : it may arise as a true gastrula invagination ; as cells 

 budded off from one pole of the blastula into its cavity ; as cells 

 budded off from various parts of the wall of the blastula ; by 

 delamination or actual division of each cell of the blastula wall ; 

 or it may be present from the start as a solid mass of cells in- 

 closed by the epiblast cells. It is in connection with these 

 variations that controversy has arisen as to the primitive mode 

 of development of the gastrula, a point to which I shall return 

 later on. 



Among the higher Metazoa or Ccelomata the extraordinary 

 modifications in the position and in every conceivable detail of 

 formation of the mesoblast in different and often in closely allied 

 forms have given rise to ardent discussion, and have led to the 

 proposal of theory after theory, each rejected in turn as only 

 affording a partial explanation, and now culminating in Kleinen- 

 berg's protest against the use of the term mesoblast at all, at any 

 rate in a sense implying any possibility of comparison with the 

 primary layers, epiblast and hypoblast, of Coelenterata. 



This is not the place to attempt to decide so difficult and 

 technical a point, even were I capable of so doing, but we may 

 well take warning from this extraordinary diversity of develop- 

 ment, the full extent of which I believe we as yet realize most 

 imperfectly, that in our attempts to reconstruct ancestral history 

 from ontogenetic development we have taken in hand no light 



NO. 1089, VOL. 42] 



task. To reconstruct Latin from modern European languages 

 would in comparison be but child's play. 



Of the readiness with which special developmental characters 

 are acquired by allied animals the brothers Sarasin ^ have given 

 us evidence in the extraordinary modifications presented by the 

 embryonic and larval respiratory organs of Amphibians. 



Confining ourselves to those forms which do not lay their eggs 

 in water, and in which consequently development takes place 

 within the e?g, we find that Ichthyophis and Salamandra have 

 three pairs of specially modified external gills. Nototrema has 

 two pairs ; Alytes and Typhlonectes have only a single nair, 

 which in the latter genus take the form of enormous leaf-like 

 outgrowths from the sides of the neck. In Ilylodes and Pipa 

 there are no gills, the tail acting as the larval respiratory organ ; 

 and in Rana opisthodon, according to Boulenger, larval respira- 

 tion is effected by nine pairs of folds of the skin of the ventral 

 surface of the body. 



Most of these extraordinarily diversified organs are clearly 

 secondarily acquired structures ; it is possible that they all are, 

 and that external gills, as was suggested by Balfour for Elasmo- 

 branchs. are to be regarded as embryonal respiratory organs 

 acquired by the larvae, and of no ancestral value. The point, 

 however, cannot be considered settled, for on this view the 

 external gills of Elasmobranchs and Amphibians would be in- 

 dependently acquired and not homologous structures, a view 

 contradicted by the close agreement in their relations in the two 

 groups, as well as by the absence of any real break between 

 external and internal gills in Amphibians. 



It is well known that the frog and the newt differ greatly in 

 important points of their development. The two-layered con- 

 dition of the epiblast in the frog is a marked point of difference, 

 which involves further changes in the mode of formation of the 

 nervous system and sense organs. The kidneys and their ducts 

 differ considerably in their development in the two forms, as do 

 also the blood-vessels. 



Concerning the early development of the blood-vessels, there 

 are considerable differences even between allied species of frogs. 

 In Rana esculenta Maurer finds that there is at first in each 

 branchial arch a single vessel or aortic arch, running directly 

 from the heart to the aorta : from the cardiac end of this aortic 

 arch a vessel grows out into the gill as the afferent branchial 

 vessel, the original aortic arch losing its connection with the 

 heart, and becoming the efferent branchial vessel. Afferent 

 and efferent branchial vessels become connected by capillaries 

 in the gill, and the course of the circulation, so long as gill- 

 breathing is maintained, is from the heart through the truncus 

 arteriosus to the afferent branchial vessel, then through the 

 gill capillaries to the efferent branchial vessel, and then on to 

 the aorta. When the pulmonary circulation is thoroughly 

 established, the branchial circulation is cut off by the efferent 

 vessel reacquiring its connection with the heart, when the blood 

 naturally takes the direct passage along it to the aorta, and so 

 escapes the gill capillaries. 



In Rana temporaria the mode of development is very different : 

 the afferent and efferent vessels arise in each arch independently 

 and almost simultaneously : the afferent vessel soon acquires con- 

 nection with the heart ; but, unlike R. esculenta, the efferent 

 vessel has no connection with the heart until the gills are about 

 to atrophy. 



In other words, the continuous aortic arch, from heart to 

 aorta, is present in R. esculenta prior to the development of the 

 gills : it becomes interrupted while the gills are in functional use, 

 but is re-established when the.=e begin to atrophy. In R. tem- 

 poraria, on the other hand, there is no continuous aortic arch 

 until the gills begin to atrophy. 



The difference is an important one, for it is a matter of con- 

 siderable morphological interest to determine whether the con- 

 tinuous aortic arch is primitive for vertebrates : i.e. whether it 

 existed prior to the development of gills. This point could be 

 practically settled if we could decide which of the two frogs', 

 R. esculenta and R. temporaria, has most correctly preserved its 

 ancestral history in this respect. 



About this there can be little doubt. The development of the 

 vessels in the newts, a less modified group than the frogs, agrees 

 with that q{ R. esculenta, and interesting confirmation is afforded 

 by a single aberrant specimen of R. temporaria, in which Mr. 

 Bles and myself found the vessels developing after the type of ^. 



' " Ergebnisse naturwissenschaftlicher Forschungen auf Ceylon," vol. ii- 

 chap. i. pp. 24-38. 



