( I'-^o ) 



40. Tanysiptera dea ellioti Sharpe. 



Differs from T. dea dea and T. dea riedeli by the absence of tlie pale apices to 

 the feathers of the mantle, which in this and all the following subspecies of T. dea 

 is uniform dark blue or black. 



T. dea ellioti differs from all other forms of T. dea in having the elongated 

 jiortion of the central rectrices much broader and much less developed spatulae. 

 The rest of the tail is also proportionately longer, which character it shares with 

 rosscliana. The perfectly adult male seems to have a pure white tail. This form 

 also differs from rosseliana by the much paler crown. The tail of the young is 

 blue in both ellioti and rosseliana. 



Inhabits the island of Koffiao, near Mysol. It was formerly erroneously said 

 to come from Mysol, but it appears to be a fact that no species of Tanysiptera 

 occurs on that island, which circumstance is also confirmed by the absence of any 

 Tanysiptera in the large collection Blr. Kiihn sent from Mysol. We have at present 

 no specimen of ellioti. 



41. Tanysiptera dea rosseliana Tristr. 



This beautiful form, when adult. Las the tail entirely white with the exception 

 of the attenuated portion of the central rectrices. It differs from all others in the 

 intense deep blue of the whole upper surface, whieh is palest and brightest on the 

 crown and ui)per wing-coverts. It occurs on the island of Rossel, in the Louisiades, 

 whence Mr. Albert Meek sent us seven specimens, namely, three fully adult males, 

 ovlb female in perfect plumage, two &A\\\t females in moult and one young bird 

 with an almost entirely blue tail, and rusty margins to some of the feathers of the 

 upper surface. 



42. Tanysiptera dea margarethae Heine. 



drown purplish blue with a much lighter, almost silvery-blue border. Ear- 

 coverts l)right deep blue. Upperside black with a purplish-blue wash. Central 

 rectrices blue with white spatulae. Inhabits the northern Moluccas, namely Batjan, 

 Halmahera and Morty. Si)ecimens from the latter island have very dark crowns, 

 but we are alile to match them with a specimen from Halmahera. 



We have the following 25 specimens : — 



1 c? Batjan (Powell coll.). 



4 cJ ad., 3 ? Batjan (Doherty coll.). 



1 ¥ Batjan (from Waterstradt's natives). 



3 c?, 1 ?, 1 sex?, juv. Batjan (Uoherty coll.). 



1 ad. Halmahera (from Bruijn's hunters). 



2 ad. Halmahera (from Bruijn's hunters, per Boucard). 

 1 juv. Halmahera (from Bruijn's hunters, per Boncard). 

 1 ad. without locality. 



3 ad., 3 juv. Morty (Dumas coll.). 



Since writing the details of the T. sabrina group, we found that six specimens 

 collected on Morty were taken at the same time as the five doris. This confirms 

 our contention that the white-backed sabrina group 'forms a distinct species from 

 the dark-backed T. dea group. 



