28 PROCEEDINGS OF TUE 



eo-calletl fiiiuilii's aiv also |)ol}i)li}'li-ti(; ; and it would he more 

 convenient iF lie could devise a nonienclature which would not 

 011I3' indicate the lineage hut also the stage of evolution therein. 

 He deals in fact with slowly changing grades w hicli are approxi- 

 mately the same in each parallel lineage of a group as it is traced 

 tlirougli successive periods of time. Jlis ideal delinilion of any 

 category inclutles not merely the usual diagnostic characters, hut 

 also a statement of tendencies in evolution. He has already 

 api)lied this method of classification to certain ammonitts anil 

 bracliiopods with much success; and, where each individual retains 

 the whole of its siveleton from early youth to old age, the method 

 is often easily applicable. 



As concrete instances of these lineages, which have now been 

 well studied, may be mentioned those of the ammonites, grapto- 

 lites, Cretaceous ])olyzoans, and Palaeozoic corals. Equally 

 important are the parallel lineages which have been recognised 

 among several groups of Tertiary mammals. On the whole, 

 perhaps, the case of the graptolites is the simplest illustration lo 

 begin wiih and may be most concisely stated*, 



in the earliest known graptolites tlie polyps are arranged on an 

 irregularly branching skeleton, often forming a hard network. 

 The thecte, or cells which lodged the polyps, are varied in shape 

 and eviilently show that these primitive skeletons belong to 

 Several genera or species. {Somewhat later, graptolites with many 

 synnnetrically arranged branches make their appearance ; and it is 

 10 be noted that the symmetry of tlie arrangement would be 

 advantageous in tending to ensure an equal supply of food to 

 each branch. Xext, there appear succeissively eight- branched, 

 four-branched, and two-branclaed forms. Eventually, just before 

 the race becomes extinct, nearly all the graptolites are simple 

 rods. In each of these grades the variety in the shapes of the 

 thecae is at least as great as in the primitive ancestors. Presum- 

 ably therefore each of the original types has independently 

 followed the same trend of evolution, successively producing 

 colonies of the same simplified shape. The Tetrngraptids, Diplo- 

 graptids, and Monograptids, therelore, are not respectively true 

 genera as was originally supposed when they were named ; they 

 are merely cjrresponding stages in the evolution of several 

 parallel groups, which were all striving towards a more effective 

 and uniform distribution of food to the colony. 



The case of the Pakeozoic corals has lately been illustratetl Ijv 

 an instructive diagram by Dr. A\\ D. Lang f. In each lineage, as 

 distinguished by the characters of its septa, tabulae, etc., the 

 initial simple coral lirst begins to form a loose cluster. iSomewhat 



* H. A. Nicholson &, J. E. Mnrr, " Notes on tiie PJiylogeny of the Qrapto- 

 htes," Geol. .Mag. [4] vol. ii. (ISll.")), pp. .")-20-r)3'.t. Also Gertrude L. Eiie^ 

 "The Gra))l(iiite Fuuiias of tlie Britisli I^iu.*." Proc. Geol. Assoc, vol. xxxiii. 

 (r.)22), pp. HJS-lioU. 



t W. J). Lang, " Trends in Biili.sli Carboniferous Coral?," I'roc. Geol. Assoc, 

 vol. xxxiv. (1023). p. 122. 



