428 



NA TURE 



[June lo, 19C9 



smell. The chief drawback to the ordinary commercial 

 method of preparing formaldehyde is, I am told, the 

 impossibility of preventing polymerisation. In the same 

 way, as Dr. Perman himself points out, hydrofluoric acid 

 at ordinary temperatures " consists mostly of molecules 

 H^F.." Hydrocyanic acid, again, shows a great tendency 

 lo polymerisation and to decomposition in the presence of 

 water. The possibility of ionisation in the presence of the 

 film of moisture on the surface of the olfactory membrane 

 and of the moist air in the nasal chambers must also be 

 taken into account. It is also possible that certain gases 

 produce an olfactive effect after the incorporation of water 

 in their molecules. 



In the second place, a distinction must be drawn 

 between indirect olfaction due to chemical action and 

 olfaction which can be accounted for only as the result 

 of the vibration of olfactory hairs. I, personally, should 

 hesitate to describe the effect upon my nervous system, 

 through my olfactory membrane, of pure ammonia, as 

 a sensation of smell. It seems to stand in an entirely 

 different category from the smelling of musk. To make 

 such a distinction recalls to mind the fact that olfaction 

 is the successor of chemical stimulation, chemiotaxis. The 

 sense of smell may be based upon the older and coarser 

 mode of action of olfactive bodies as well as upon the 

 more modern and refined. 



Either of the three substances which Dr. Perman has 

 cited as odorous is capable of producing a change in the 

 constitution of protoplasm such as cannot, we suppose, 

 be produced by the minimal amount of human effluvium 

 which enables a dog to track his master, or even by the 

 minimal quantity of drifting particles which are capable 

 of appealing to a man's far less sensitive nose. It can 

 be demonstrated experimentally that one part of mercaptan 

 in 50,000,000,000 of air gives a recognisable odour to the 

 mixture. Chemical action in such a case seems to be out 

 of the question. 



Although we cannot conceive the way in which so minute 

 a quantity of matter plays upon the instrument which 

 originates nerve-impulses, we picture the olfactory hairs as 

 answering to some change in the vibrations of the mole- 

 cules of air, or of the atoms within their molecules, due 

 to the influence of the olfactive particles. Such evidence 

 as is at present available, if we make allowance for the 

 sources of error to which I have alluded, points to the 

 conclusion that to produce this molecular or intra- 

 molecular change the added gas must be heavier than air. 

 That olfactivity is not proportional to density is sufficiently 

 evidenced by the aggressive scent of sulphuretted hydrogen 

 and of many other substances which are comparatively 

 light. _ In my letter of May 13 I suggested that the 

 inability of flies to distinguish between pure water and 

 water containing formaldehyde seems to point to the same 

 conclusion. Alex Hill. 



The Germ-layer Theory. 



The most important criticism in the review on May 13 

 of "The Origin of Vertebrates," by Dr. W. H. Gaskell, 

 is based on a dogmatic view as to the fundamental dis- 

 tinctness of the germ layers and their predetermination 

 for the formation of certain organs. It is evident that 

 your reviewer regards this as a settled fact. It is there- 

 fore only fair to point out that this is by no means the 

 opinion of all morphologists. Indeed, Morgan, Hertwig, 

 Braem, Driesch, Conklin, Jenkinson, and many others 

 grant little phylogenetic value to the germinal layers. 



The germ-layer theory requires the supposition that 

 there is a prelocalisation in the egg of the various sub- 

 stances necessary for the formation of the different organs, 

 and that these substances in its segmentation pass into 

 definite segments which form the germ layers. Now this 

 supposition is directly contradicted — or at least made 

 exceedingly improbable — by the results of the experimental 

 separation of the first two, four, eight, or sixteen cells 

 formed in the development of many animals. Further, 

 some of the facts of regeneration and budding show that 

 the ectoderm is on occasion quite capable of forming endo- 

 derm and mesoderm. The anomalies also which exist in 

 the formation of the layers in vertebrates are patent to 

 every student, while research on cell-lineages in the 



NO. 2067, VOL. 80] 



invertebrates has shown most diverse histories. So far a& 

 an independent observer can judge, the trend of modern 

 research is to show that embryology gives no sure evidence 

 of the homology of the germ layers. 



J. Stanley Gardiner. 

 Cambridge, May 22. 



Perhaps the reviewer should have made it plainer that 

 the difficulty he stated at the top of p. 303 is not admitted 

 by those morphologists who have ceased to believe that 

 the germ layers afford any criterion of homology. He 

 simply expressed his conviction, which he shares with 

 many, that it does count for something which layer a 

 structure develops from. He said that he was not pre- 

 pared to follow Dr. Gaskell in throwing the germ-layer 

 theory overboard, and that this made criticism difficult, a 

 discussion of the author's dismissal of the theory being 

 impossible in an article which appreciation of the book 

 discussed had already expanded far beyond the limits 

 prescribed. The Reviewer. 



Gaskell's "Origin of Vertebrates." 



In the review of my book on the " Origin of Verte- 

 brates," which appeared in Nature of May 13, the 

 reviewer, discussing my theory that the vertebrate central 

 nervous system represents the conjoint central nervous 

 system and alimentary canal of an arthropod, says " this 

 view lands us in difficulties which seem to us as insuper- 

 able as those of the reversal hypothesis seem to the 

 author." He then proceeds to say, " we want to know, 

 for instance, where the arthropod's mesenteron has gone? "■ 

 This is the " only one of the most obvious difficulties " 

 of which he makes mention. I wish he had mentioned 

 more, as I am most anxious to have all the difficulties of 

 my theory pointed out and fully discussed. 



He will find in my paper in the Quarterly Journal of 

 Microscopical Science, vol. xxxi., that I look upon the 

 peculiar tissue which fills up the space between the braini 

 and the cranial wall in Ammoccetes as the remains of the 

 corresponding tissue which surrounds the brain of such 

 animals as Limulus ; in other words, this tissue represents 

 the mass of generative glands and so-called liver-tissues in 

 these animals. This so-called liver, together with its duct 

 or ducts leading into the gut, constitutes the mesenteron, 

 and the most distinct remnant of such mesenteron ira 

 Ammocoetes is the tube, called by me the old liver-tube, 

 which leads from the fourth ventricle to terminate on the 

 surface of the brain at the conus post-commissuralis, as 

 is shown in a series of sections reproduced in that paper. 

 In my book I have discussed this vestige of the arthro- 

 pod's mesenteron on pp. 209, 210, 211, chapter v., but 

 have not re-published the series of sections given in my 

 former paper. In the summary of chapter v. I have not 

 mentioned this question of the vestiges of the arthropod's 

 liver, as it was not especially concerned with the subject- 

 matter of chapter v. ; possibly that is the reason why it 

 has failed to attract the notice of the reviewer. 



The reviewer says that " the tubular appearance of the 

 vertebrate central nervous system appears to some an 

 unimportant architectural consequence of the mode of 

 development from a medullary groove," and also in reply 

 to my argument " that the extraordinary resemblance 

 between the structure and arrangement of the central 

 nervous systems of vertebrates dnd arthropods is against 

 the view of their phyletic distinctness," he asserts that, 

 " given segmentation in two distinct types, we naturallv 

 expect similarity in the general plan of innervation." 

 But the whole point is that the tube is not a simple tube 

 such as would be formed by the coming together of 

 medullarv folds, but one, which invariablv possesses a 

 ventral diverticulum, the tube of the infundib'ulum, situated 

 in exactly the position of the arthropod oesophagus, on the 

 view of the phyletic relationship between the central 

 nervous systems of the arthropod and the vertebrate. 



The reviewer seems to think that I lay too much stress 

 on Ammocoetes and ignore Amphioxus and the tunicates, 

 and also that I am inclined to flit a little from type to 

 type, making use of arachnids. Peripatus, and annelids 

 when the Pal;Eostraca are insuflHcient. I thought I had 

 made it clear in my book that my object was to find out, 



