BY R. J. TILLYARD. 277 



the d liferent Archetj^pes; also P, Q, B, S, not having the 



same values for different Orders, we may omit these letters, and 

 consider that the whole series of characters is comprised in a 



single (longer) series of variables a, b, c, d, e, J] etc. 



Taking a group of Orders, we may now represent the most 

 archaic value of the character a hy A, while «', a\ etc., represent 

 various specialised conditions of the same. Then we shall be 

 able to write the Archetypes of different Orders in this way 



Archetype of Order 1: - ^4, h", C, d', e, F, g" . . . (say). 



Archetype of Order 2: -a, h\ C, d", EJ'\ G . . . (say). 



Archetype of Order 3:—^, B, c', Z>, e",f\ g' . . . (say); and 

 so on. 



The condition that any one Order may be ancestral to another 

 can now be determined by comparing the Archetype of the 

 supposedly higher Order with any supposedly ancestral group 

 lying within the bounds of the Lower Order. But, as the 

 characters of the Archetype of the supposedly ancestral Order 

 are even more archaic than those of the supposedly ancestral 

 group lying within that Order, the determination may be made, 

 once for all, by comparing the Archetypes of the two Orders. 

 The following rule may be laid down : — 



The condition that one Order may be truly considered ances- 

 tral to another, is that there must not exist a single character in 

 the Archetype of the former, ivhich is more highly specialised than 

 the corresponding character in that of the latter. 



For example, consider the case of Orders 1 and 3 above. The 

 character A is common to both Archetypes. For the character 

 6, the Archetype of Order 3 is the older {B against h"). But for 

 the character c, the Archetype of Order 1 is the older (C against 

 c). Hence, clearly, neither of these two Orders can be con- 

 sidered as ancestral to the other; and we must go back, for their 

 common ancestor, to an Order (probably no longer existing) 

 whose Archetype shows the characters A, Jj, C, . . . . 



The Recognition of Embryonic ^Structures. 

 It would seem necessary here to call attention to the fact, 

 which seems to have been quite ignored by most of those ento- 

 20 



