358 Letters, Announcements, ^c. 



there are two kinds of characters : — Istj those which are dia- 

 gnostic of the group ; and, 2nd, those which are common to all its 

 members, and are, so far, characteristic, though they may not 

 be diagnostic. Thus in defining the class Mammalia, one does 

 not omit to state that the blood is hot, though the warmth of 

 the blood is not a diagnostic character of that group ; and in 

 attempting to define the Ratitce and the Cai'inatce, characters 

 which are common to all the members of those groups, though 

 they may not be absolutely diagnostic of them, should surely not 

 be omitted. 



Further, it must be recollected that the diagnosis of a group 

 may rest not merely on a particular character confined to the 

 group, but on a peculiar combination of characters. 



And it may happen that a well-defined group shall not have 

 a single structural feature peculiar to itself, its peculiarity lying 

 entirely in the mode of combination of those features ; so that 

 if each one of the seven characters of the Raiitce which I have 

 enumerated were discoverable in some other animal, but in a 

 different state of combination (if I may express myself chemi- 

 cally), I do not think the goodness of the definition would be 

 interfered with. 



I quite agree with you, that " a really natural arrangement 

 can only be made out by taking an aggregate of characters'^; 

 and, practically, I have endeavoured to express this belief by 

 enumerating seven characters for the Ratita and three for the 

 Carinata. 



On the other hand, whatever one's notions may be about what 

 is philosophical and what otherwise, it is a matter of fact and 

 every-day experience in zoology, that the modifications of a soli- 

 tary organ will sometimes afford indications of affinity of great 

 value throughout a whole class, or even subkingdom. 



What to an a priori speculator could seem more unphiloso- 

 pliical and one-sided than to attempt to arrange the Vertebrata 

 cording to their occipital condyles, or according to the way in 

 which the lower jaw is connected with the skull ? And yet by 

 either of these characters one would be able to assign 999 verte- 

 brate animals out of 1000 to their proper divisions. 



Or, again, what can be (theoretically) more open to criticism 



