While collecting in the densely forested lands along the Caribbean shores of Panama, I was so fortunate as to secure a specimen of the peculiar little tropical hawk, Harpagus bidetatus fasciatus. (The following week Mr. Frank Richardson obtained a second specimen in the same area.) The general appearance of the bird when I picked it up was that of an Accipiter with an abbreviated tail, but I was completely mystified on close scrutiny to find two corneous denticles on each side of the bill. These "teeth" are even better developed than the single tooth of typical falcons. I had never before met such a bird, and the only books available in the field contained no record of it. He had to appear in my field record under the pseudonym of "Double-toothed Hawk." On my return to California the generic and specific identity of the hawk was deter- mined from Carriker's paper on the Birds of Costa Rica (Ann. Carnegie Mus., vol. 6, no. 7, 1910, pp. 314-915) and the Catalogue of Birds of the British Museum (R. Bowdler Sharpe, vol. 1, 1874). Both these publications place the genus Harpagus in close proximity to the falcons and to the kites. Sushkin (Nouv. Mem. Soc. Imp. Sci. Nat., Moscou, vol. 16, pt. 4, 1905) states that such an allocation of the genus is based upon the denticulation of the bill which he considers not fundamental: His opinion is that Harpagus is not to be assigned to the Falconidae. Peters, in his Birds of the World (Harvard Univ. Press, vol. 1, 1931), places the genus in the family Accipitridae and in the subfamily Milvinae immediately adjacent to the Accipitrinae. My own immediate reaction, having only the fresh bird in hand, was that the kinship was with the accipitrids and not the falconids in spite of the denticulate bill. It was with much interest, then, that I looked forward to a study of thd bird's skeleton. Tarsus, tibia coracoid and humerus were compared in detail, with the result that my first impression was greatly strengthened and I came to agree pretty closely with Peters' arrangement. He recognizes nine subfamilies of the Accipitridae which he lists seriatim with Accipitrinae as no. 4. The genus Harpagus is assigned to subfamily no. 3, the Milvinae, including the old world type, Milvus milvus milvus. The genus Falco is placed at the extreme end of the series of four subfamilies in the Falconidae. Assuming that Peters' arrangement is based upon his impression of the degree of kinship, Har- pagus would appear almost.as remote from Fatco as it could well be. Even the aberrant Pandion lies between them! Peters' arrangement,. exclusive of the cathartids and Sagittarius is as follows: Superfamily Falconidae 7. Circinae Family Accipitridae 8. Circaetinae Subfamilies 1. Elaninae 9. Pandioninae 2. Perninae Family Falconidae 3. Mjlvinae Subfamilies 1. Herpetotherinae 4. Accipitrinae 2. Polyborinae 5. Buteoninae 3. Polihieracinae 6. Aegypiinae 4. Falconinae This arrangement seems a smoothly flowing progression except that one might say it is necessary to "shift gears" to include Pandioninae in the itinerary. A more detailed comparison of the several skeletal elements is as follows. Harpagus bidentatus, Elanus leucurus, Accipiter cooperil, and Falco columbarius are taken as typical forms. Tarsometatarsus.--The proximal end of the tarsus, viewed from the anterior face, shows Harpagus to be close to Accipiter, less like Elanus, and far from Falco. The pos- terior and the articular views of the bone show exactly the same relationship. When the distal end is studied, there appears a more exaggerated inner trochlea than any of the other species. This throws the center of the foot farther toward the inner side than any of the others, though the degree is approached by Elanus. The outer trochlea is weak with little backward projection. This gives much less "transverse arching" to the foot. Accipiter approaches this character. The outer toe is less raised than in Falco or Elcmus, but is much like that in Accipiter. Tibiotarsus.--The tibia is strikingly like that of Accipiter in all respects except that the outer condyle is deeply cut away on the proximo-anterior sector of its arc. The outer condyle is thus reduced to a much smaller mass than the inner. None of the other genera has so marked a discrepancy between the condyles, though it is indicated in Falco. The tendinal groove in Falco and Elanus is broad and deep; and, of course, in Falco, it is bifurcate. In Accipiter and Harpagus this groove is much constricted. The width of the tibial shaft just proximal to the condyles is again suggestive of Accipiter in that the shaft is but little narrower than the condyles. In Falco and Elanus a marked widening through the condyles gives the impression of greater strength in the ankle joint. The posterior aspect of the condylar region is quite unlike Falco in its lack of a definite groove for the tendon of Achilles. It resembles Accipiter more than it does Elanus in that the internal condyle is not marked off by an obliquely transverse ridge.  Coracoid.--The coracoid of Falco differs very widely from that of Elaus or of Accipiter. Harpagus falls between these two latter genera in the characters of this bone. Falco has a long, slender procoracoid which is strongly flexed over the triosseal canal. This process is but moderately developed in the other three genera here compared. A coracoidal fenestra is complete in Elanus, is strongly notched in Falco, is faintly indicated in Harpagus, and entirely wanting in A1/2cipiter. The sterno-coracoidal imprint reaches far up the coracoidal shaft in Falco, is less extensive in Elanus, still less in Harpagus, and least in Accipiter. The sterno-coracoidal process follows the same sequence in respect to its breadth. The glenoid facet is broad- est in Accipiter and Harpagus, is narrower in Falco, and narrowest in Elanus. Humerus.--The affinities as displayed by the humerus all lie with Accipiter. Elanus has a long and almost straight humerus with subdued topography such as one finds in the sailing fliers. Falco stands at the opposite extreme with a short, sigmoidal humerus, widely expanded at the ends and very rugged in its contours. Such characters are found in birds of great speed and endurance on the wing. Accipiter and Harpagus, very much alike, occupy the intermediate ground that is correlated with the habit of a still-hunter who catches his prey by a quick dash from some place of concealment. Our much too brief observations in the field strengthen this latter deduction. Both specimens were taken in dense jungle from the jacaranda trees. Any bird, to be visible from the ground, must have come well down into the crown of the trees below the denser leaf zone, a typical accipitrine behavior. Insects constituted the only fbod identified in the stomach. University o! California at Los Angeles, April x7, 957.