66 ON THE GENUS LATHAMUS. 



feathers wide. In the more anterior parts of each arm the most internal 

 feather of each row is often placed in front of and at an angle with the 

 other feathers composing it, and so comes to stand between two rows of 

 three feathers each ; so that at first each tract looks as if made up of 

 rows of three (or four) feathers alternating with single feathers. This 

 tendency to a 3.1.3 arrangement, however, disappears in the more pos- 

 terior parts of the tracts, the four feathers of each row there standing in 

 a direct line with one another. The two arms unite to form the " handle " 

 at about three quarters their entire length ; after the junction the 

 tract narrows rather rapidly towards the tail. The dorso-lumbar fork is 

 throughout quite distinct from the lumbar feathering, which is very weak 

 and diffuse. 



In all the truly Platycercine* forms that I have examined namely 

 Platycercus eocimius and pennantii, Psephotus hcematogaster (four specimens) 

 and P. hcematonotus, Pyrrhulopsis splendens and P. personata, Cyano- 

 rhamphus auriceps and 0. novce-zealandice the disposition of the outer 

 pectoral tract and dorso-lumbar fork resembles essentially that of 

 Lathamus. In all the outer pectoral is a distinct, more closely feathered, 

 and rather narrowish tract, clearly separated throughout from the main 

 part. In Cyanorhamphus this tract is distinctly hook-like, dilated at the 

 end. In all the same lengtht, and uniformity in strength and width, of 

 P. Z. S. 1879, the arms of the dorso-lumbar tract is observable, the inclosed space being 

 p. 170. O f about the same width as either of the tracts inclosing it, no tendency 

 to a dilatation of the arms at their junction (though there is some in front) 

 being present, and the rows of feathers in front having a more or less 

 clear 3.1.3 arrangement. The lumbar feathering is always very weak ; 

 so that the boundaries of the dorso-lumbar fork are very clearly defined. 

 Lathamus, however, differs from the above-mentioned forms a little by 

 its longer and not so widely divaricated scapular fork, and by the greater 

 breadth of its inferior tract on the sternum, thereby causing a correspond- 

 ing diminution in the breadth of the carinal space. The general agree- 

 ment, however, of the pterylosis in the two types will, I think, at once 

 be evident from the figure of Lathamus (PI. I. figs. 1, 2), and that of 

 Platycercus pennantii (PL I. figs. 3, 4), which I have represented next 

 to it for the sake of comparison. 



If now we turn to the TrichoglossinseJ (see PI. I. figs. 5, 6), in 

 which so many naturalists have included Lathamus, we shall find im- 

 portant and well-marked differences in the two tracts mentioned above, 



* I. e. excluding Aprosmictus, Potyteles, Euphema, Pezoporus, &c. 



t In Pe. pennanti and in the two species of Pyrrhulopsis I counted fourteen, in 

 C. auriceps thirteen, in Ps. hcematonotus thirteen, and in Ps. hcematogaster eleven rows 

 of feathers in the arms of this tract to their junction. 



| Of these I have examined the pterylosis in Eos rvhra, Trichoglossi ornatus, hesma- 

 todes, swainsoni, concinnus (two specimens), and pusillus, and Coriphilus fringilloceus. 



