G32 REPORT— 1902. 



three determined advances upon them, and of these the third and most recent 

 may be first discussed. It has for its object the attempt to prove that they are 

 intimately associated with the cephalaspidian and other shield-bearing fishes of 

 the Devonian and Silurian epochs, and that through them they are ancestral 

 to the Vertebrata. The latest phase of this idea is based on the supposed 

 existence in a Cepkalasjns of a series of twenty-five to thirty lateral appendages 

 of arthropod type.'*** When, however, it is found that the would-be limbs are 

 but the edges of body-scutes misinterpreted, suspicion is aroused ; and when, 

 working back from this, an earlier attempt reveals the fact that the author, 

 compelled to find trabeculse, in order to force a presupposed comparison between 

 the architecture of the Cephalaspidian head-shield and the Limulus' prosomal hood, 

 resorts to a comparison between the structure of the former in general and that of 

 the cornu of the latter, with details which on the piscine side are not to date, the 

 argument must be condemned.'"'^ It violates the first principles of comparative 

 morphology, and is revolting to common sense ; and as to the fishes concerned, 

 we know that they have nothing whatever to do with the Limuloids, for we 

 have already seen that, with their allies the Pteraspidiae, they are a lateral branch 

 of the ancestral piscine 8tem.'-"'° 



The second advance upon the king crabs has very much in common with the 

 first. It has engrossed the attention of an eminent physiologist for the last six or 

 seven years, and by him it was in detail set before Section I at our meeting of 1S96. 

 Suffice it to say that it specially aims at establishing a structural community 

 between the king crabs and certain vertebrates, favourable to the conviction that 

 the Vertebrata have had an arthropod ancestry.' '' When we critically survey 

 the appalling accumulation of words begotten of this task, it is sufficient to con- 

 sider its opening and closing phases. At the outset, under the conclusion that 

 the vertebrate nervous axis is the metamorphosed alimentary canal of the arthropod 

 ancestor, the necessity for finding a digestive gland is mainly met by homologising 

 the so-calltd liver of the arthropod with the cellular arachnoid of the larval 

 lamprey, in violation of the first principles of comparative histology ! '"'- At the 

 close we find ingenious attempts to homologise nerve tracts and commissures 

 related to the organs of sense, such as are invariably present wherever such 

 organs occur.'^^ Sufficient this to show that the comparison, in respect to its leading 

 features, is in the opening case strained to an unnatural degree, in the closing case 

 no comparison at all. Finding, as we do, that the rest of the work is on a par 

 with this, we are compelled to reject the main conclusion as unnatural and 

 unsound ; and when we seek the explanation of this remarkable course of action, 

 we are forced to believe that it lies in the failure to understand the nature of the 

 morphological method. For the proper pursuit of comparative morphology, it 

 is not sufficient that any two organisms chosen here and there should be com- 

 pared, with total disregard of even elementary principles. Comparison should 

 be first close and with nearly related forms, passing later into larger groups, with 

 the progressive elimination of those characters which are found to be least 

 constant. And necessary is it, above all things, that in instituting comparison it 

 should be first ascertained what it is that constitutes a crustacean a crustacean, a 

 marsipobranch a cyclostome, and so on for the rest. We have tried to accept this 

 theory, fascinated both by the arguments employed and by the idea itself, which 

 for ingenuity it would be difficult to beat, but we cannot ; and we dismiss it as 

 misleading, as a fallacy, begotten of a misconception of the nature of the morpho- 

 logical method of research. '^^ It is of the order of events which led Owen to 

 compare a cephalopod and a vertebrate,'-'' led Lacaze-Duthiers to regard the Tuni- 

 cates and Lamellibranchs as allied ; '■•"^ and with these and other heresies it must 

 be denounced. 



Passing to the third advance, extending over the last twenty years, it may be 

 said to consist in the revival of a theory of 1829, which boldly asserts that 

 Limulus is an Arachnid. In the development of the defence there have been 

 two weak points but lately strengthened, viz., the insufficient consideration of 

 the palseontological side of the question and of the presence of tracheae among the 

 Arachnida.'^' Under the former there was, until recently, assumed the absence of 



