VOl 'lfl^ XI1 ] VON Ihering, The Dendrocolaplidcc. 149 



we should be able to find other characters to support our action 

 and these, I believe, exist. 



The true Furnariinse have the tail truncated while in the genera 

 Lochmias, Upucerthia and other Cinclodinse the exterior rectrices 

 are successively shortened. If we consider that this latter condi- 

 tion prevails in general throughout the Dendrocolaptidse we must 

 realize that the tail structure in the true Furnariinse is quite a 

 remarkable peculiarity. 



The Furnariinse have probably originated through localization 

 in the vast prairies of the La Plata states and the adjacent parts 

 of Brazil and Bolivia, while the origin of the Cinclodinse has been 

 in Patagonia and the Andes. 



It is not easy to trace the lines of dispersal which have brought 

 about the present distribution of the South American Dendro- 

 colaptidse but some light is thrown upon the matter by the study of 

 ornithological literature. Of special interest in this connection is 

 the history of Furnarvus, the Ovenbird, one of the characteristic 

 species of the central Brazilian and Argentine fauna which seems 

 to be still extending its range. When Natterer in the years 1818- 

 1823 explored the state of Sao Paulo, he did not meet with it 

 although at the present time it is common in the valley of the Para- 

 hyba river and appeared some fifteen years ago at Campinas 

 where it nests. 



We may also infer that the genus Cinclodes in eastern Brazil is a 

 relatively recent immigrant, as also the few species of Pteroptochidse, 

 a family of Patagonian-Andean origin. 



Of several genera of the Dendrocolaptidse the skull is unknown 

 to me, such as Margarornis and Glyphorhynchus, so that I cannot 

 form an opinion upon their relationships from cranial characters. 

 It is not, however, my intention to propose here a new system of 

 classification for the family, my aim being rather to furnish new 

 facts based upon biological and anatomical observations which 

 may eventually be of value in the construction of such a system. 



As in the Furnariinse two lines of development have\been demon- 

 strated we can presume that the Dendrocolaptinse sprang from 

 two different groups of the Philydorinse. Probably the case is 

 more or less the same with respect to the somewhat aberrant 

 genera Sclerurus, Glyphorhynchus and Margarornis. 



