INTRODUCTION. 



The sj-stematic arrangement followed in the present volume is, in 

 part only, my own. I had to accommodate my views to those held 

 and expressed by the authors of the previous volumes, notably in 

 vol. iv. p. 6, where the limits of the last two families of the Cichlo- 

 morphse, Paridte and Laniidae, were defined, and in vol. v., where the 

 limits of the family Turdidffi^'were so much contracted that, in order 

 to render a Catalogue of Cichlomorphous genera complete, I had to 

 admit into the families which fell to mj' share genera which, in my 

 opinion, ought to have been placed elsewhere. 



It is easy to form a satisfactory diagnosis of so-called typical 

 families if we confine ourselves to such forms as we consider taxo- 

 nomic centres ; but if we have to include the more peripheral forms, 

 which are frequently highly specialized, as well as the more 

 generalized forms, where the various groups inosculate with and 

 overlap each other, the task of defining the families becomes ex- 

 tremely difficult, as was already experienced by the author of the 

 fourth volume (p. 6). 



The combination of the Parince with the Regulince seems to me 

 artificial, and the latter probably had better have been treated in 

 vol. vi. Moreover, Regulus and Leptopcecile being apparently closely 

 allied to each other, I found it impossible to give a diagnosis of the 

 family Paridoe. 



A discussion of the probable relationship between the Paridce and 

 Sittince had become unnecessary through the establishment of the 

 group Certhiomorpha? (Cat. B. vol. iii. p. 3). 



Regarding the Laniidae I beg to refer to pp. 88 and 89. 



