November 24, 1892] 



NA TURE 



79 



irrelevant consideration ; and I have never been able to under- 

 stand the stress laid upon it by acute thinkers. It i> because 

 the triangle is as far as I can perceive isosceles, that I intuit it 

 to be as far as I can perceive equal-angled. 



It has, I believe, been already explicitly recognized by cer- 

 tain logicians that a "symbolically" proved conclusion need 

 not give any actual information about " real things." Indeed 

 some go further ; but I do not know that any have gone so far 

 as to say that it would not give any information about ideas — 

 although perhaps this may be the logical conclusion. 



Cambridge, November lo. E. E. Constance Jones. 



Ice Crystals. 



Your correspondent, C. M. Irvine (vol. xlvii. p. 31) will find 

 letters on this subject in Nature, vol. xxxi. pp. 5, 81, 193, 

 264, 480, and in vol. xxxiii. pp, 461, 486. 



Prof. (?) McGee's letter at p. 480, of vol. xxxi. , gives a list of 

 communications on the same subject in earlier volumes. 



B. WooDD Smith. 



The Late Prof. Tennant on Magic Mirrors. 



Several scientific friends tell me that the late Prof. Tennant, 

 the well-known mineralogist, published some twenty or twenty- 

 five years ago a small pamphlet on Magic Mirrors. Failing to 

 find a copy even in the library of King's College, I invite the 

 readers of Nature to assist me to discover one. 



Silvanus p. Thompson. 

 City and Guilds Technical College, Finsbury, 

 November 15. 



On a Supposed Law of Metazoan Development. 



Under the title of " The Relations of Larvce to Adult 

 Forms," I recently read a paper before Section D at the 

 Edinburgh meeting of the British Association. The subject 

 dealt with was of so extensive a nature, and the time available 

 was so limited, that I fear much that was said must have ap- 

 peared vague and ill-founded, if not entirely incomprehensible. 

 The material of the essay had, indeed, been prepared with the 

 intention of devoting at least an hour to its delivery : as it 

 happened, I found myself under the necessity of cutting out whole 

 passages of my notes whilst speaking. 



The few lines of the report in Nature (vol. xlvi. p. 404), 

 ■convey a very inadequate idea of what I aimed at proving in the 

 paper, and hence I am tempted to offer a fuller account to the 

 readers of this journal. 



The subject of the es^ay furnishes a problem which must 

 interest every embryologist, even though he should reject the 

 conclusions to which observation and reflection have led me. 



In working out the complete paper so many novel and con- 

 firmatory points have been met with, so much of imp >rtance in 

 the writings of the older embryologists, and more especially in 

 the memoirs of Johannes Miiller on the Echinoderm larvae, has 

 been unearthed, that an extension of the original plan of the 

 work has been rendered necessary. 



My conclusions, moreover, are so much in conflict with pre- 

 vailing doctri les that any haste in producing the full argument 

 would be unpardonable, although a preliminary sketch by way 

 of clearing the ground may be justifiable. On a subsequent 

 occasion an attempt would be made to show how the researches 

 of recent years had, with a few notable exceptions (such as the 

 work of R. S. Bergh, J. Kennel, and N. Kleinenberg), tended 

 away from rather than in the direction) of a recognition of 

 the fundamental fact of an alternation of generations as under- 

 lying Metazoan development, in that they had been concerned, 

 for example, with unnecessary attempts at homologizing the 

 " mesoderm" and its mode of formation throughout the animal 

 kingdom. 



If the fac's in support of my case should not be as complete 

 a-; the published researches of the last thirty years on the 

 ontogeny of very many animals might lead one to anticipate, 

 the circumstance would have an obvious explanation. 



With the death of Johannes Miiller — \ man whose brilliance 

 as an enbryolo^ist wa? only surpiss;d by his greatness as an 

 anitomist — there closed one chapter, and that one of the finest, 

 in the history of comparative embryology. What influence the 

 publication of "The Origin of Species" had upon the subse- 

 quent progress of che science is too well known to need further 

 expatiation here. The pernicious search after pedigrees, 



NO. 1204, VOL. 47] 



initiated by Haeckel, led lo an era of activity during which 

 every fact with no apparent bearings on phyiogeny was ignored. 

 As a consequence the work of Miiller on the Echinoderm larvae 

 and the essay of Steenstrup on "Alternation of Generations" 

 became more or less mere curiosities in the history of the science. 

 With little exception embryolo^ical speculation of the past thirty 

 years has been naught else than a pursuit of will-o'-the-wisps. 



rtbekoves us now to revert to the pxth along which yohanncs 

 Miiller laboured. 



My own embryological conclusions, like those of contempor- 

 aries, hive not hitherto been influenced by the embryological 

 works of Miiller ; for it was not until after my paper had been 

 read that a first study of the Echinoderm memoirs convinced me 

 how nearly he had anticipated what follows. 



Before passing to the subject, one further remark may be 

 permissible. Owing to lack of time when reading the paper, 

 no opportunity offered itself for pointing out the analogy which 

 obtains between the suggested mode of Metazoan development 

 and the accepted fact of an alternation of generations in the 

 life-histories of all plants above the lowest Thallophytes. 

 Furthermore nothing was said about the mode of formation of 

 the " mesoderm " in certain cases as one or more outgrowths of 

 the endoderm ; alth>ugh the writer was fully alive to the ex- 

 planation which from his standpoint could be offered. This 

 and other questions of alike character would receive consideration 

 in the complete paper, in which it would be demonstrated that 

 such things and processes need be neither " palingenetic " nor 

 "coenogenetic," but that the analogy of the formation of 

 imaginal discs in Insecta, or in the Piiidiutn of the Nemertine, 

 ought to suflice to account for them. As an instance, the for- 

 mation of the mesoblastic somites in Amphioxus as evaginations 

 of the endoderm may be only a mode in which certain parts of 

 the adult are in that particular case laid down upon the larva. 



And now, after this digression, to return to the question 

 under consideration. Two modes of development have long 

 been distinguished, viz., larval witk metamorphosis ?lx\A fatal and 

 direct. Cases are known in which there subsists no homology 

 between the larva and the adult, and even such in which the 

 larva {Bipinnaria asterigera) is said to exist apart for a time after 

 it has given rise to the Echinoderm. In many such, moreover, 

 the sole larval organ carried over to the adult is the alimentary 

 tract, all other organs of the larva, such as nervous system, 

 sense organs, locomotor and excretory organs, mouth and anus, 

 &c., being replaced by new formations in the adult. The new 

 organs are thus not homologous with tho-e of the larva ; indeed, 

 neither as a whole nor in its parts is the larva the homologue of 

 the adult form ; but the latter arises upon the former by a mode 

 of asexual generation. 



The birth of the Nemertine on the Pili/iium, and that of the 

 Echinoderm upon the Pluteus, or upon the Bipinnaria asterigera, 

 may be cited as examples, and the question may now be asked. 

 What becomes of the larva when (a) food-yolk is more or 

 less abundantly acquired, and (fl) when uterine development is 

 initiated? Does the larva really disappear ? Anticipating the 

 sequel, it is asserted that the larv.i never vanishes from the 

 development, but is always present in more or less disguised 

 form. In all cases the adult or ima^jo would appear to arise 

 upon it just as is so obviously the case in the examples previously 

 cited. 



In the complete paper the modifications of the process 

 throughout the Metazoa would be considered ; in this place 

 generalities alone can be dealt with. If the larva be laded 

 with food-yolk it becomes transformed into a more or less 

 obvious bUs'oderm, upon which the imago or mature form takes 

 its origin. Certain of the larval or,ians — such as those of locomo- 

 tion — may then disappear, but others, such as the larval excre- 

 tory and nervous mechanisms i^e.g., Hirudinea, according to 

 Bergh's researches, Ichthyopsida from my own work) would per- 

 sist. Considerations of space do not permit me to enter 

 fully into details regarding Molluscan development. The 

 published work on this group furnishes one with useful 

 material in support of my case ; and the group is an in- 

 teresting one in connection with this question of the relation 

 of the larva to a blastoderm. In the Mollusca one can readily 

 find all gradations from cases in which the adult is gradually 

 substituted for a pelagic larva {Patella), through those in which 

 the larva is somewhat burdened with foodyolk {Buccinum), to 

 others, finally, in which there is a large yolk-sac and a blasto- 

 derm, on which the adult form a.T\ses (Ceplialopoda). Incident- 

 ally I may remark that it was the study of some Buccinum 



