1906.] OF THE TRACflEOPHOXE PASSERES. 157 



truth if we reduce it to the status o£ a sub-family, and create, to 

 accompany it, the sub-families Ptei'optochinje and Hylactin^e — 

 the former of these being regarded by Sharpe (' Hand-list ') as a 

 Family, and the latter as a genus only thereof. 



The essential feature of the Gonopophagidee is the 4-notched 

 sternal plate, and the most primitive member of the family is 

 Gonopophaga. In the peculiar character of the sternum (p. 147) 

 this family is unique among the Passeres. 



The Furnariine forms need careful revision. The composition 

 of this family roughly corresponds to the Dendrocolaptidee of 

 Dr. Sharpe minus the Dendrocolaptinpe, which, it seems to me, 

 should be regarded as a separate Family. 



As touching the Xenicidge, I have recently elsewhere (8) con- 

 tended that this Family is more or less nearly related to the 

 Synallaxidee, and this largely, but not entirely, on account of 

 osteological characters. Moi^e primitive in some respects than 

 this Family, they differ chiefly in the form of the syrinx, which is, 

 tracheo-bronchial, and therefore the Xenicidse would appeal- to be 

 at the bottom of the tracheophone stem, the members of which 

 split up into holorhinal and schizorhinal types. 



The scutellation of the podotheca, largely used in Sclater and 

 Salvin's classification of the Group (10), cannot be relied on as a 

 guide to the closer bonds of affinity. Thus, in the Gonopophagidje, 

 Conopophaga is exaspidean, like the Tyi'annidse ancl Pipridse ; 

 Pteroptochiis is taxaspidean,like the Formicariidse and Philepittidfe, 

 while the Dendrocolaptinfe and Synallaxinte are endaspidean. 



In the matter of pterylosis all the Tracheophonefe have a long 

 10th remex and a vestige of the 11th, and all have a more or less 

 saddle-shaped expansion to the pt. spinalis, the tract behind this 

 being feebly developed. 



The curious form- of the nostrils of Xenicus and the I'emarkable 

 structure of the external ear I have already described at length 

 (8). But little attention has ever been paid to this apertvire, 

 and it is probable that a careful study of the form of the external 

 ear will be rewarded by interesting i-esults. 



The external nares in Scytalopus and Conopophaga are covered 

 by a leaf -shaped operculum. 



The delioideus longicseiudbrevis — musclesof the shoulder-girdle — - 

 ai'e, as Dr. Mitchell has showai, of considerable value as factors in 

 classification. In the paper on Acanthidositta, to which I have 

 several times referred here, I have shown that these muscles, in 

 this genus, have presei-ved their primitive character to a very 

 unusual degree : the longus portion being two-headed, the second 

 head being attached to the os htLinero-scapidcm'e and forming with 

 the claviculo-scapular head a large and powerful muscle inserted 

 into the ectepicondylar process of the humerus by a short tendon. 

 In such Tracheophonepe as I have been enabled so far to examine 

 in this respect, I find the more normal, specialised, condition to 

 obtain. Thus in Scytalopus and Conopophaga and in Formicivora 

 the longus portion has lost the second head, though in Conopo- 



