BIRDS OF NORTH AND MIDDLE AMERICA. H39 



(3) The Elaininse restricted by elimination of the genera Rhynclio- 

 cyclus, Legatus, Myiozetetes, Conopias, Pitangus, Sirystes, and 

 Myiodynastes, which, except the first (referred to Euscarthininie), 

 should constitute a "subfamily" Pitanginir, morphologically inter- 

 mediate between the Elainin;i> and Tyrannin;r, but which perhaps 

 should be united with the latter. 



These modifications in the arrangement of the Tyrannidse proposed 

 by Doctor Von Ihering are in my opinion entirely justifiable, but I 

 am convinced that they are but a step toward further changes which 

 are necessary before a correct exposition of the phylogeny of the 

 group can be made; and, while not prepared to forecast these with 

 any degree of certainty, will express my belief that the genera Agri- 

 ornis and Muscisaxicola are each quite sui generis and are not as 

 closely related to Tsenioptera as the latter is to typical Tyrannidse; 

 and that several genera do not belong to the Tyrannidas at all, but 

 must be transferred to other groups, none of them possessing the 

 exaspidean tarsal envelope, the only external character, except that of 

 less pronounced syndactylism (very variable within the group), which 

 can be said to be really diagnostic of the family. These are (1) 

 Lawrencia, which, having acutiplantar tarsi, is a member of the 

 Oscines;" (2) Stigmatura,^ (3) Hapalocercus,'^ and (4) Hahrura,^ 



aLawrencia has ten obvious primaries, the tenth being about half as long as the 

 ninth; the basal phalanx of the middle toe is completely united to the outer toe and 

 almost wholy adherent to the inner toe, thus agreeing, as in the wing-structure, with 

 the Vireonidse, where thegenus proljal)ly l>elongs, though differing from other vireonine 

 genera in its depressed and broadly triangular bill. 



bSligmatura resembles in general form the formicarian genus Formicivora and 

 possibly belongs to the same family. 



c Hapalocercus is also possibly formicariine in its ailinities. One species has l>oen 

 referred to it which has the typical tyrannine exaspidean tarsi and therefore certainly 

 is not congeneric. This is Aleclrurvs flaviventris D'Orbigny and Lafresnaye, type of 

 the genus Myiosyvipotes Reichonbach (Av. Syst. Nat., 1850, pi. G5), and therefore t(j 

 be known as Myiosyvipotes flaviventris. I have not seen //. fulviceps nor H. acuti- 

 pennis and therefore can not say whether they are congeneric with //. melacoryphus 

 or not; but the former doubtless is, since its very near relative, //. pauliis Bangs, is a 

 true Hapalocercus. 



<*The tarsal envelope of Habrura, while less typically taxaspidean than tliat of 

 Hapalocercus, nevertheless conforms to the same type and is by no means exaspidean. 

 The inner side of the planta tarsi consists of a single continuous series of well-defined 

 quadrate scutella; but on the outer side the acrotarsium extends quite to the posterior 

 edge of the tarsus except for the upper third, where three or four rather large and very 

 distinct longitudinal scutella occupy approximately the posterior half. This last 

 character is seen in many typical Tyrannid;v, but in none of the latter is then^ even 

 any indication of the well-defined and continuous series of qua,drate scutella along 

 the posterior half of the inner side of the tarsus. The nostrils in Habrura are quite 

 different from those of Hapalocercus, roundish and nonoperculate, while in the latter 

 they are more longitudinal, relatively larger, and overhung by a rather broad membrane- 

 ous operculum. The proper ])osition for these two genera is a question which I am 

 not able to decide, but woidd suggest that Hapalocercus may be a member of the 

 Formicariida?, while Habrura might not be out of place in the Cotingidaj. 



