PREFACE. Vll 



Difficulty has hitherto arisen from the fact that Variation is 

 not studied for its own sake. Each observer has some other object 

 in view, and we are fortunate if he is good enough to mention in 

 passing the variations he has happened to see in following his own 

 ends. From the nature of the case these observations must at 

 first be sporadic, and, as each standing alone seems to have little 

 value, in the end they are unheeded and lost. If there were any 

 centi^al collection of facts to which such observations might from 

 time to time be added, and thus brought into relation with cognate 

 observations, their value would at once appear and be preserved. 

 To make a nucleus for such a collection is the object of the present 

 work. 



The subject treated in this first instalment has been chosen for 

 the reasons given in the text. Reference to facts that could not 

 be included in this section of the evidence has as far as possible 

 been avoided, but occasionally such reference was necessary, 

 especially in the Introduction. 



It was my original purpose to have published the facts with- 

 out comment. This course would have been the most logical and 

 the safest, but with hesitation it was decided to add something of 

 the nature of analysis. I do this chiefly for two reasons. First, 

 in starting a method one is almost compelled to shew the way in 

 which it is to be applied. If it is hoped that others may interest 

 themselves in the facts, it is necessary to shew how and why their 

 interest is asked. In the old time the facts of Nature were 

 beautiful in themselves and needed not the rouge of speculation 

 to quicken their charm, but that was long ago, before Modern 

 Science was born. 



Besides this, to avoid the taint of theory in morphology is 

 impossible, however much it may be wished. The whole science 

 is riddled with theory. Not a specimen can be described without 

 the use of a terminology coloured by theory, implying the accept- 

 ance of some one or other theory of homologies. If only to avoid 

 misconception matters of theory must be spoken of. 



It seemed at first also that the meaning of the facts was so 

 clear that all would read it alike ; but from opportunities that 

 have occurred for the discussion of these matters I have found 

 that it is not so, and reluctantly I have therefore made such com- 

 ments as may serve to bring out the chief significance of the 



