10 MEMOIRS OF THE CARNEGIE MUSEUM. 



four teeth. But of this latter species I have but few specimens, unsatisfactorily 

 preserved, and the result is doubtful. The larger number of the species have seven 

 premaxillary teeth. The number of species having six or eight teeth are nearly 

 equally matched. It must be borne in mind that the number of specimens examined 

 has not been so great that we can be sure that all variations have been observed. 

 The maxillary teeth recorded are as follows : 



Number of teeth: 0,1, 2, 3 , 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 

 Number of species: 47 19 10 75. 5 2113244322 1 1 1 3 



Four of the species occur in two counts, six occur in three, four in four, one in 

 five, one in seven, and one in twelve (from nine to twenty in Aphyocharax dentatus) . 

 The prevailing number is two teeth, which are found in nineteen species, as shown 

 by the table; which also shows that the number of species having from twelve to 

 fourteen and twenty teeth is relatively greater than those having from seven to 

 eleven, and from fifteen to nineteen teeth. 



The Relationship op the Cheirodontinje. 



Are the Cheirodontinse a homogeneous group with a common ancestry, or 

 are they dwarfs of various other subfamilies? The most of them form a homo- 

 geneous group, divisible, however, into a number of minor groups. Doubt arises 

 as to Grundulus, which has only conic teeth, Paragoniates, Leptagoniates, and 

 Phanagoniates with a posterior dorsal, and Mixobrycon, which has tetragonopterid 

 teeth and cheeks. Certain other characters, notably the peculiar scaling in the 

 caudal of the males of Compsura and Odontostilbe hastata, also suggest relationship 

 to another subfamily, the Glandulocaudinse. 



The unicuspid teeth of Grundulus suggest relationship with the Characinse, 

 as some of the Characinse with partially tricuspid teeth, Oligosargus exodon, and 

 Bramocharax suggest relationship with the Cheirodontinse. The general shape and 

 backward position of the dorsal of Paragoniates as well as the peculiar scales in the 

 tail of the male Compsura and Odontostilbe recall the Glandulocaudinse. The 

 heavy teeth and armature of the cheeks of Mixobrycon suggest Hyphessobrycon of 

 the Tetragonopterinse and so does the tooth out of line with the rest in Megalam- 

 phodus micropterus. However, a double row of teeth has several timesbeen evolved 

 in the Characinse from a single row; or a single row from a double row. I have 

 pointed out such cases in Indiana University Studies No. 20, and will have occasion 

 to point out others in the monographs on the Chalcininse and Gasteropelicinse. 

 In fact if it were not for other considerations, the single series of teeth in the Cheiro- 

 dontinse would be no more sufficient to segregate them from the TetragonoiDterinse 



