O. — GEOLOGY. 65 



inferior. Through, the more precise observations of Von Baer, Louis 

 Agasaiz, and others, the idea grew until it was crystaUised by the 

 poetic imagination of Haeckel in his fundamental law of the reproduction 

 of life — namely, that every creature tends in the course of its individual 

 development to pass through stages similar to those passed tlu'ough 

 in the history of its race. This principle is of value if applied with 

 the necessary safeguards. If it was ever brought into disrepute, it 

 was owing to tlie reckless enthusiasm of some embryologists, who 

 unwarrantably extended the statement to all shapes and structures 

 observed in the developing animal, such as those evoked by special 

 conditions of larval existence, sometimes forgetting that every con- 

 ceivable ancestor must at least have been capable of earning its own 

 livelihood. Or, again, they compared the early stages of an individual 

 with the adult structure of its contemporaries instead of with that of 

 its predecessors in time. Often, too, the searcher into the embryology 

 of creatures now living was forced to study some form that really was 

 highly specialised, such as the unstalked Crinoid Antedon, and he 

 made matters worse by comparing its larvae with forms far too remote 

 in time. Allman, for instance, thought he saw in the developing 

 Antedon a Cystid stage, and so the Cystids were regarded as the ancestors 

 of the Orinoids ; but we now find that stage more closely paralleled 

 in some Crinoids of Carboniferous and Permian age, and we realise 

 that the Cystid structure is quite different. 



Such errors were due to the ignoring of time relations or to lack 

 of acquaintance with extinct forms, and were beautifully illustrated 

 in those phylogenetic trees which, in the 'eighties, every dissector of 

 a new or striking animal thought it his duty to plant at the end of 

 his paper. The trees have withered, because they were not rooted in 

 the past. 



A similar mistake was made by the palaeontologist who, happening 

 on a new fossil, blazoned it forth as a link between groups previously 

 unconnected — and in too many cases unconnected still. This action, 

 natural and even justifiable under the old purely descriptive system, 

 became fallacious when descent was taken as the basis. In those days 

 one heard much of generalised types, especially among the older fossils ; 

 animals were supposed to combine the features of two or three classes. 

 This mode of thought is not quite extinct, for in the last American 

 edition of Zittel's ' Palaeontology ' Stephanocrinus is still spoken of as 

 a Crinoid related to the Blastoids, if not also to the Cystids. Let it 

 be clear that these so-called ' generalised ' or ' annectant ' types are 

 not regarded by their expositors as ancestral. Of course, a genus 

 existing at a certain period may give rise to two different genera of a 

 succeeding period, as possibly the Devonian Coelocrinits evolved into 

 Agaricocrhui!;, with concave base, and into Dorycrinus, with convex 

 base, both Carboniferous genera. But, to exemplify the kind of state- 

 ment here criticised, perhaps T may quote from another distinguished 

 writer of the present century : ' The new genus is a truly annectant 

 form uniting the Melocrinidae and the Platycrinidae. ' Now the genus 

 in question appeared, so far as we know, rather late in the Lower 

 Carboniferous, whereas both Platycrinidae and Melocrinidae were already 



1920 g 



