NO. 1116. PROCEEDINGS OF THE NATIONAL MUSEUM. 481 



I 



'. Xo marlvin.<;;s on breast or chest, 

 fc'. Darker, with dusky prevailing on upper parts. 

 c'. Bill larger, but wing averaging shorter (exposed lulmen 0.92-1.07, wing 



4.00-4.60). (Abingdon Island) 4. N.inrHonatiis (-p. ^Hi^). 



C-. Bill smaller and more slender, but wing averaging longer (exposed cul- 

 men 0.80-0.87, wing 4.20-4.68). (.James, Indefatigable, and .Tervis 



islands) 5. N. melanotts (p. 489). 



h-. I'aler, with grayish brown prevailing on upper parts. 

 &. Bill smaller (exposed culmen 0.72-0.80). (Albemarle Island.) 



6. N.jyarvuhis (p. 491). 



(•■'. Bill larger (exposed culmen 0.83-1.00). 



d . Larger, with shorter tarsus (exposed culmen 0.95-1.00, wing 4.30-4.45, 



tail 3 95-1.30, tarsus 1.2.5-1.35. (Tower Islaud) . . 7. X baiiri (p. 492). 



d'. Smaller, with longer tarsus (exposed culmen 0.83-0.91, wing 3.85-4.20, 



tail 3.50-3.90, tarsus 1.32-1.39). (Bindloe Island). 



8. N.bindloei (p. 492). 



I T am aware of tlic unsatisfactory nature of the above key, wliicli is 

 ' owiiiji' in large measure to the fact that in some cases the specimens 

 ui)on which it is based (Dr. Baur's Sjiecimens having- i)reviously been 

 returned to hiin) are in a state of plumage different from that of speci- 

 mens of the allied species, rendering exact comparison impossible. 



The species of this genus fall into two groups, which in a more exact 

 sense might be considered as S[)e(ies, the several allied forms being- 

 more i>ropeily subspecies or local races. Thus K. trij^isciatus, N. mac- 

 donaldi, and ISf, adamsi, with their double breastbands, represent one 

 type, Avhile N. Melanotis, N. parvuJns.^ N. jx't'sonattin, X. hauri, and X. 

 hiitdioci represent another. Were these forms of continental habitat, 

 where there was every probability or even certainty of intergradation, 

 I should consider them subspecies and give them trinomial instead of 

 binomial names (for example, Xcsoinimus trlfasciutus viacdonaldi, Xesom- 

 imus trifasciatus adamsi, Xesomimus melanotis personatus, Nesomimus 

 melanotis x>arvidvs, Xesomimus melanotis bauri, and Xesomhnns melanotis 

 Mndloei). But the case of insular forms is different, intergradation in 

 the same sense being impossible, and therefore, notwithstanding the 

 slight differences they present, I prefer to consider the several forms 

 as if they were distinct species. 



The genus Nesontimus occurs at present, or rather as recently as 

 Messrs. Baur and Adams' visit (1891), on all the larger islands .oi' the 

 Galapagos Archipelago, except Narborough, Duncan, and Charles. It 

 was formerly represented on the lasst-named by a peculiar species {N. tri- 

 faseiatus, closely related to X. macdoiialdi of Hood Island), which seems 

 to have, like Geospiza m((gnirostris of the same island, become extinct 

 there.i The genus was also found by Messrs, Baur and Adams to be 



'Messrs. Baur and Adams collected a Nesomimus on Gardner Islaud, which they 

 ideutitied as X macdonaldi, the Hood Island form; but as Gardner Islaud is much 

 closer to Charles than to Hood Island it is possible the species was .V. trifasciatus. 

 Unfortunately the specimens were lost, and the form must therefore remain in doubt 

 until more can be procured. 



Proc. N. M. vol. xix 31 



