( 141 ) 



more outlying islands, Chatham, (Jharles, and Hood, and again Bindloe, Abingdon, 

 and Tower, as well as Wenman and Culpepper, have generally more ditFerentiated 

 forms. Abingdon and Bindloe have aj)parently almost always the same subspecies, if 

 the species is found on both. The ornis of Hood and Tower is very poor in the number 

 of species, but very different. Wenman and Culpepper have also very few species 

 of land l)irds, but they are nearly all different from those of the other islands. The 

 Xesomimiis of Wenman, however, we cannot distinguish from that of the central 

 islands. Wenman and Cnlpepper have often, but by no means always, the same forms. 



Botli Wenman and Cnlpepper are as yet apparently insufficiently explored. 

 Narborough and Albemarle have the same forms, as might be expected from their 

 close proximity, but the Xesomimns seems to lie quite different, although closely 

 allied. These facts seem not to have been explained before ; in fact, only a large 

 material could help to show them. 



The evident affinities with tlie North and Central American ornis (cf Phoeni- 

 copterus ruber, Haetnatnpns galapiigensis (very closely allied to //. frit~(tri from 

 Lower California), Ardea herocliaa, and perhaps Myiarchus and Dendroicu) are easily 

 explained by the tendency of northern forms to migrate in a southerly direction, 

 while tropical forms do not actually migrate. Analogous facts are observed in the 

 ( 'anary and Cape Verde Islands, where a great proportion of European forms is 

 found. The southern hemisphere has evidently sent hardly any colonists to the 

 Galapagos Islands. The only striking example of these is Spkeniscus menJicnlns. 

 This is by far the most northern home of any member of the order Impennes, 

 which is entirely restricted to the southern hemisphere. No species is known to 

 cross the equator, only three reach as far north as Pern, Rio Grande do Sul, and 

 South Africa, and Spheniscus mendiculus alone lives on the equator. Of no other 

 resident bird can we confidently say that it can only have been derived from the 

 southern hemisphere. Even migrants from the south are not known to occur, 

 with the exception of the alleged occnrrence of a single specimen of Querquedula 

 cersicolor (see p. 203). 



It has been thought to be possible that the larger islands might have different 

 representative forms in various parts. Especially of the large island of Albemarle 

 the probability has been suggested, that tlie birds of North- and South-Albemarle 

 might differ, as the tortoises from these parts are recognised as two different species. 

 Oar collections, however, do not support this idea, but they seem to prove that only 

 one representative form is found on every island. 



Salvin has also raised the question, whether the elevated interior portions 

 of the islands, where " clouds usually hang over the higher mountains, where the 

 moisture is far greater than on the sea-shore, and consequently the vegetation is far 

 more luxuriant," were inhabited by different birds. Mr. Harris and his companions 

 did not find tiiis to be the case, but they found the same subspecies in suitable places 

 in the various parts of the islands. 



During our work we have had most assistance from Mr. Kidgway's admirable 

 work on the Birds of the Galapagos Archipelago. Principally we agree with 

 Mr. Ridgway in going as fiir as possible in distinguishing, and consequently naming, 

 as many forms as jwssible ; and we fully bear witness " that the real promoter of 

 chaos and enemy of order is the ' lumper,' and not his nuich maligned co-worker, 

 the ' hair-splitter.' " 



Our material, consisting of not less than 3073 skins from the recent expedition 

 under Mr. Harris, and the Baur collection of about llno skins, is by far the largest 



