for Systematic Biology. 269 



in the nature of the case, necessarily mere conjecture, I gradually 

 realised exactly what it was possible to do with the specimens so 

 that the work which had been expended on them could be recorded, 

 and, at the same time, a start made towards a natural classification. 

 I found it, however, almost useless to suggest this practical 

 solution to my fellow workers. One here and there, who has 

 found exactly the same kind of difficulty as I have with the 

 Corals, expressed approval, but the great mass of modern systema- 

 tists say they do not want it. For their worst difficulties, they 

 say, a little patching of the old system will amply suffice and 

 so on. I perceived further that this attitude is not solely due to 

 the fact that, with regard to the more stable forms of life, it is 

 still possible, by means of the old formula of naming, to do accu- 

 rate scientific work, preparatory towards a natural classification, 

 but also largely to the fact that the prevailing ideas as to the 

 aims of classification are not sufficiently defined. Discussions 

 have been all about nomenclature and not about classification, 

 that is, not about the principles that should inspire nomen- 

 clature. 



It became quite clear, then, that before any better method of 

 naming is demanded, a need for it must be established. The 

 actual formulae of classification are purely secondary; they are 

 only the symbols which we agree to use to express an order of 

 ideas. Before any reform in our methods of designating specimens 

 can be accepted, it is essential that the ideal aims of classification 

 should be very clearly understood. When that is once achieved, 

 it will be psychologically impossible to revert to any system which 

 totally fails to express them. I propose, then, after describing 

 the aims of classification as the theory of Evolution has silently 

 advanced them, to show that the present method of classifi- 

 cation, which has survived from pre- Darwinian days, fails to 

 express these aims, and fails even in those cases in which, owing 

 to the stability of the group, the work done is as exact as it could 

 be, while, in cases where the forms are very inconstant, the old 

 formula, as we are at present forced to use it, positively hinders the 

 attainment of any good object whatever. That being established, 

 I shall sketch the line of reform of method which the exigencies 

 of my own work suggested to me, and which, the longer I work 

 with it becomes more practical and promising. I am aware that 

 to attempt to alter our time-honoured methods of classification 

 may be a daring thing to do, but the strength of my position lies 

 in this, that I am simply appealing to the first principles of 

 scientific work, and maintaining that, in classification as in all 

 other departments, we must have a system of work which allows 

 us to begin by collating the facts and nothing but the facts; 

 and further that the time has come when what all evolutionists 



