460 Birds of Celebes: Nectariniidae. 



formula C. frenatus = saleyerensis for a form which is equidistant from the typical 

 freiiatus and from saleyerensis, then C. frenattis <: saleyerensis with a more acute 

 angle for forms which incline towards saleyerensis in a less degree, and C. frenatus 

 {saleyerensis with a less acute angle for forms wliich incline towards saleyerensis in 

 a greater degree. To indicate the estunated degrees mathematically (<[ 45°) is too 

 fantastic a method to commend itself. 



More practical — but to be condemned perhaps as suggesting an accuracy of 

 knowledge which we do not j^ossess — is the use of numerals as j^ropounded else- 

 where (see also Haliastur Indus). Between the typical C. frenatits and the extreme 

 C. frenatus saleyerensis we believe we could arrange a series of 24 "subspecies" with 

 the small degree of difference seen between C. f plateni and C. f dissentiens taken 

 for the unit: then the typical form remains the typical C. frenatus, or C. frenatics2t 

 saleyerensiso; the Saleyer bird is C. frenatiiso saleyerensiso^; the North Celebes bird 

 C. frenatusso saleyerensis 4, standing much nearer to the typical than to the Saleyer 

 form;. the lowland bird of South Celebes (plateni) may be held to have, say, 7 times 

 as much affinity with the Saleyer race as with the typical form, or: C. frenatus 3 

 saleyerensis2j; the Bonthain bird standing a little nearer still to that of Saleyer 

 becomes: C. frenatus2 saleyerensiso2. 



These formulae are very easy to set up, but any one who will undertake to 

 prove that the numerals are erroneously apphed will have a difficult, and remarkable, 

 piece of work to do. 



It is preferable here to abstain from inventing a formula for the Bonthain 

 birds. These specimens are referred to under: 



'^ Cyrtostomus frenatus dissentiens (Hart.). 



t Cinnyris frenata dissentiens (1) Hart., Nov. Zool. 1896, 152, pt.; (2) id., ib. 1897, 155. 

 Distribution. Bonthain Mts. — Indrulaman (Everett 1); Loka (P. & F. Sarasin); "from 

 the lower hills up to 6000 ft." (Doherty 2). 



Further local variation. Adults fi-om Celebes may be at once distingivished from adults 

 fi'om New Guinea (Passim, Dore) by their browner olive-green upper surface. 

 Not so an adult male from Buru (Nr. 1855), which is like Celebes birds as to 

 its back, but is a trifle larger in size (wing 58 mm). An adult male from Bat j an 

 is a little bro^vner above than the New Guinea specimens, though hardly distin- 

 guishable. Adult males from Peling and Banggai are larger in bill and body than 

 the form of N. Celebes, but in colour they are often similar, that is, browner 

 olive-green than the typical, eastern, C. frenatus. An adult male and female fi'om 

 New Britain are the yellowest specimens as regards their upper surface we have 

 seen, and it appears as if an influence, which makes the birds become yellower or 

 browner respectively, is felt in increasing force the further south and east or west 

 the species ranges. In a similar manner Haliastur i?idus seems to have gradually 

 lost its dark stripes in extending its range from India to the Papuan Islands. Mr. 

 Biittikofer terms the characters j^ointed out by Prof. W. Blasius in C. frenatus 

 plateni inconstant, so that even the separation of the Celebes -form as a subspecies 

 appears unallowable; on the other hand, we find that individual variation in 

 C. frenatus as a whole seems to run within comparatively narrow hmits, though plateni, 

 etc. are undoubtedly connected with the typical form of New Guinea by every con- 

 ceivable intermediate form from the Moluccas. 



