December 15, 1910J 



NATURE 



205 



notion of the general form of certain Silurian " fish-like " 

 forms. Assuming, however, that it is the case that to 

 the onlooker there is a striking resemblance between the 

 earliest fishes and the Arthropoda, it is necessary to point 

 out that striking resemblance in superficial characters 

 provides a type of pitfall which the morphologist has at 

 an early stage in his education to school himself to avoid. 

 He comes across cases of amazing resemblance, e.g. in 

 pairs of " mimetic " butterflies, between a marsupial and 

 a placental mammal, between the organ of vision of one 

 of the higher insects and that of one of the higher 

 Crustacea, between the skeleton of a flagellate and that 

 of a radiolarian, and he learns to recognise that super- 

 ficial resemblance may, and frequently does, provide a 

 cloak for fundamental unlikeness. It is, in fact, one of 

 the main parts of his business as a morphologist to find 

 out whether in each particular case the striking resem- 

 blance so apparent to the onlooker is an expression of 

 resemblance in fundamental points of structure or whether, 

 on the other hand, it is merely superficial. 



I think I have now said enough to make apparent how 

 greatly some of us who devote ourselves to the problem 

 of vertebrate phjiogen\- differ from Dr. Gaskell and those 

 with him in what we regard as the necessary principles 

 in accord with which morphological work must be done. 

 As regards Dr. Gaskell's main thesis, that vertebrates are 

 descended from arthropods, w-e take the position that what 

 is known of the morpholep- of vertebrates in general, and 

 of arthropods in general, does not justify the regarding 

 of that view as a reasonable working hypothesis. Were 

 Dr. Gaskell to increase by several fold his mass of de- 

 tailed anatomical resemblances between an undoubted 

 arthropod and an undoubted vertebrate, we should feel 

 ourselves confronted, not by a demonstration of near 

 genetic affinity, but rather by a fascinating puzzle in the 

 way of convergent evolution. It would take up too much 

 space, and perhaps serve little purpose, to indicate the 

 general considerations, the cumulative effect of which is to 

 force zoologists into the position I have indicated. I may, 

 however, just indicate one feature, the character of the 

 skeleton, which, as it happens, is a highly characteristic 

 feature alike in the arthropods and in the vertebrates. 

 In the .-Xrthropoda we find one of the finest evolutionary- 

 inventions existing in the animal kingdom — a supporting 

 skeleton formed out of waste products of metabolism, and 

 so spread over the surface of their body as to form an 

 armour effectively protecting the delicate living tissues of 

 the body from the most varied kinds of dangers. That a 

 group of free-living arthropods should have given up this 

 magnificent protective device is not, of course, incredible, 

 but before being accepted as probable it would have to 

 be supported by an overwhelming mass of evidence. 



What evidence do we, in fact, find as to the nature of 



the skeleton of the primitive vertebrate? We find in the 



vertebrate that the skeleton in the earliest stages of its 



development consists of a cellular rod cut off from the 



dorsal wall of the gut and running longitudinally along 



the median plane of the body. This rod, the notochord, 



which still persists as the main axial skeleton in the 



adults of some of the lowest vertebrates, occurs during 



embryonic development, not merely in a few vertebrate 



forms, but in every lower vertebrate the embr\-ology of 



which has so far been investigated. There is not a single 



exception. Could evidence be more overwhelming that the 



ancestral vertebrate was a creature with a skeleton in the 



form, not of a hypertrophied cuticle, but of a cellular 



notochordal rod formed from the wall of the gut? Dr. 



Starling would object that " no palaeontological evidence 



seems to be brought forward in favour of this hypothesis." 



1 The answer is a perfectlj- simple one. Palaeontology can 



from the nature of the case offer us hardly any evidence 



I whatever in regard to structures composed of soft, perish- 



j abl« organic material. By far the greater part of the 



I tissues and organs with which the morphologist deals are 



i composed of such soft, perishable materials, and all these 



j structures, upon the cumulative evidence of which (taken 



I in conjunction with the much smaller amount of evidence 



I obtained from the skeleton) he bases his conclusions, are 



I almost entirely absent from the geological record. I 



I might go farther and point out that a highly developed, 



highly rigid skeleton is in itself a product of long-con- 



Tued evolutionary change. Each animal possessing such 



NO. 2146, VOL. 85] 



a skeleton is the descendent of a long line of soft proto- 

 plasmic forms in which the skeleton had not yet become 

 evolved, and these ancestral forms have, so far as 

 palaeontology is concerned, vanished for ever from our 

 ken. It is upon the study of the comparative anatomy 

 and embryology of existing forms alone that we have to 

 depend when we endeavour to form a picture of what 

 these ancestral forms were like. 



The foregoing paragraphs are not meant as a criticism 

 of Dr. Gaskell's hypothesis. They are merely meant to 

 direct attention to an extraordinary want of agreement 

 as to methods or principles of morphological research. 

 It is clear that work in any department of science must 

 be done according to some definite set of principles if it 

 is to be of any appreciable value. In morphology, as in 

 other sciences, there are certain generally accepted prin- 

 ciples. It seems to me not unreasonable to ask that 

 workers who take up morphological research should either 

 accept these general principles and be guided by them, or, 

 if they find themselves driven to formulate a new and 

 better set of principles, that they should at least state 

 these clearly and give their fellow-workers the opportunity- 

 of judging in what respects they are better and more 

 trustworthy' than those in ordinary use. Unless this is 

 done there is apt to be caused an irritating waste of time 

 and energy. There is, further, the danger that important 

 work may be rejected without adequate examination, not 

 because of its inferior quality, but simply because of the 

 difficulty in the way of discovering common factors 

 between it and work on the more ordinary and orthodox 

 lines. J. Graham Kerr. 



The University, Glasgow. 



Mendelian Expectations. 



Mr. Lewis Bonhote has confirmed Mr. R. Staples 

 Browne's statement that the web-foot in pigeons is a 

 simple Mendelian recessive, but he finds that when webbed 

 birds from two different strains are crossed an irregular 

 result is obtained, viz. four normal and one webbed. 

 Moreover, mating the first crosses yielded results in almost 

 every case contrary to Mendelian expectations, normals 

 throwing webs and webs throwing normals (Nature, 

 December i, p. i6o). Similar results have been obtained 

 with West Highland terriers, in which white is apparently 

 recessive to yellow. The offspring of pure-bred white 

 terriers belonging to the same strain are white, but the 

 offspring of pure-bred white terriers from different strains 

 are sometimes yellow. Further, a hybrid (yellow) which 

 produced more than 50 per cent, of white pups to a white 

 dog of her own (Inverness) strain produced only yellow 

 pups to a white dog of a different (Poltalloch) strain. 



The explanation of these " irregular " results seems to 

 be that the normal toes, presumably latent in webbed 

 pigeons, and the yellow coat, presumably latent in white 

 terriers, are restored when two strains having a somewhat 

 different history' are interbred, i.e. mingling the blood of 

 two strains induces reversion. Because *' points " are lost 

 when two strains are crossed, many breeders are extremely 

 reluctant to introduce new blood even when their stock 

 is obviously deteriorating from in-and-in breeding. 

 Recently a very successful breeder assured the writer that 

 nothing in the world would induce him to use the blood 

 of another strain to improve his white Highland terriers, 

 and it is notorious that breeders of sheep and cattle have 

 once and again allowed their flocks and herds to lapse 

 owing to their reluctance to infuse fresh blood from other 

 strains. Further, Von Oettingen has pointed out that, in 

 the case of the English racehorse, the more remotely 

 related the parents the less chance there is of the offspring 

 winning races. 



If crossing two strains is liable to lead to reversion, we 

 can understand why in some hands breeding is such a 

 lotter}', why, e.g., the offspring of two record racers or 

 trotters are swnetimes complete failures. The English 

 thoroughbred breed is made up of several distinct tvpes, 

 each of which is now and again represented by a Derby 

 winner. When two fleet but not too closely related 

 members of the same t^pe are mated, the result may 

 prove highly satisfactorj, but when the sire belongs to 

 one tj'pe and the dam to another, and when, in addition. 



