42 BULLETIN 36, UNITED STATES NATIONAL MUSEUM. 



I have observed it, is always clear plumbeous ftiay above. The color of 

 dead specimeus is very dark, bat I have ucver observed a greeiiisli tinge 

 iu the coloratiou. The types of T. catalania were lead-colored. We 

 might divide, therefore, the three species by their coloration as fol- 

 lows : 



1. Upper parts lead-colored. 



a. Belly spotted T. calalania. 



h. Belly unspotted T. tinsio. 



2. Upper i^arts dark sea greeu. 



Belly spotted T. ahusalam. 



The proportions of the body are much alike in T. catalania and T. 

 ahtisalam, as will be seen from the following comparative table: 



Mensurements. 



Total kngtli 



Tip of suout to eye 



Tip of snout to anterior base of doisal tin 



Height of dorsal tin (vertical) 



Distance between tbc extremiiics of the Hukes 



T. catalania 

 (from Macgil- 



lerray, re- 

 duced to me- 

 ters), ? . 



2.058 

 0.305 

 0.915 

 0. 203 

 0.559 



T. abusalam 

 (fromKiip- 



pell, reduced 



to meters), 



$ adult. 



1.949 

 0. 297 

 0.893 

 0.216 

 0.514 



The skeleton of T. catalania is unknown; for T. ahiisalam Etippell 

 gives the following vertebral formula: C, 7 ; D., 12 ; L., 16 ; Ca., 2G=G1. 



In two of the skeletons of T. tursio in the national collection, how- 

 ever, but twelve pairs of ribs are present, and in one of these, which 

 appears to be perfect, the total number of vertebrte is but sixty-one. 

 We can, therefore, make no separation on the basis of the total num- 

 ber of vertebra. 



From such indications as we possess, this species appears to be inter- 

 terniediate between T. tursio and T. catalania, but most closely" allied 

 with the latter. It would not bo a matter of surprise if it should prove 

 to be identical with T. catalania. 



Since the foregoing remarks were written Dr. W. Kobelt, director of 

 the Senckenberg Museum, has very obligingly sent me some measure- 

 ments of the type-skull iu the collections of that institution. From 

 these it appears that the beak is longer than in T. tursio, as already 

 stated, and is much narrower at the base than in cither T. tursio or T. 

 catalania. Its breadth at the middle is about as iu the latter species. 

 The breadth at the orbits, however, is less than iu either of the others. 

 I subjoin the measurements : 



Afeasiircmcnfs of the iijpe slnll of Titrsio2>s ahusalam, 9 . 



Centimeters. 



Total length '. 40.0 



Leugth of beak ■- -7. 



Breadth of beak before the notclies 9-2 



Breadth of beak at its middle ^>- 8 



Breadth of intermaxill;c at same jioint ^-4 



Breadth between ceiiter.s of orbits l^^- 7 



