376 MR. AV. p. PYCRAFT — CONTRIBUTIONS [Apr. 9, 



still perhaps the Artamidte should be split up into two sub- 

 families— -Artaminse, Yireolaniinfe. 



I have been much puzzled as to the systematic position of the 

 genus Sphecotheres. Generally regarded as one of the Oriolidfe, it 

 seems to me much more nearly allied to the Campophagidse, and 

 should indeed be included in this Family. The skvill bears a 

 general resemblance to that of Graucalus. It is certainly not an 

 Oriole ; at any rate if the skulls in the Museum Collection are 

 rightly labelled, and there seems to be no reason to doubt this. 

 The fact that Sphecotheres has not developed the peculiar spiny 

 rump-feathers so chaiacteristic of the Campophagidse may be 

 urged, by some, as an objection to the introduction of this genus 

 to the Family Campophagidse. This, however, does not seem a 

 very weighty objection, for the feathers in question vary in the 

 degree of their spininess very considerably, in some genera it is 

 hardly noticeable. 



The genus Tylas again — which includes five species — has 

 puzzled me greatly. Included by Dr. Sharpe with the Pycno- 

 notidse, it seems to me to belong rather to the Prionopidte, though 

 in some I'espects it recalls the Muscicapidpe. It has certainly no 

 aifinities with the Pycnonotidfe. 



The resemblances which the birds of the genus Tylas 

 undoubtedly present, in many osteological characters, to the 

 Laniine group are significant ; and this because, curiously enough, 

 they have been, and still are, regarded as " Babblers " (Pycno- 

 notida?) which have, for purposes of mimicry, assumed the guise 

 of Shrikes ! 



Briefly, the genus Tylas includes five species, all of which are 

 peculiar to Madagascar. All bear a strong likeness one to the 

 other ; and resemble scarcely less closely a member of the genus 

 Xenopirostris — X. polleni. This bird is considered by some to be 

 a Shrike, while by others it is, and probably rightly, included 

 among the Artamidse, or the closely allied Vangidfe — the two 

 Families must I think be united. There are three species in the 

 genus Xenopirostris^ two of which are dull-coloured and bear no 

 likeness to X. polleni. Thus all the five species of Tylas seem to 

 have more or less completely assumed the livery of X. polleni. 

 On the assumption that the genus Tylas belonged to the 

 Pycnonotidfe, this resemblance to an agressive Shrike was truly 

 remarkable ; but the likeness loses little of its interest even now 

 that it is shown that this genus really belongs to Pi-ionopidpe, a 

 Family which must be included among the Shrikes. 



Finally, I propose to arrange the Diacromyodean Families 

 herein described in four groups, as follow : — 



Suborder Diacromyodi. 



Group I. Hirundines. 



Fam. 1. Hirundinidfe. 



