54 ME. W. S. EOWNTEEE ON THE 



The conditions in my specimens liave rarely permitted of accurate counting, and there 

 is reason to think that there is considerable variation even in individuals ; hut, 

 supplementing my own observations with those of previous writers, I may make the 

 following statement regarding the number of appendages commonly found in different 

 forms : — 



Erythrhnis 30 Cuvier & Valencieinies. 



Macrodon 60 ( + ) in 2 rows „ 



Lebiasina 5 or 6 „ 



Pyrrhulina 6 (5 right, 1 left) ,, 



Otlier Characinids . . 10 to 25, rising in Hydro- Bouleuger. 



cyon and Citharinus 



to 35 or 40. 



The appendages vary not only in number, but also in length and thickness, and in 

 their arrangement. Thus, according to my observations, those of Sydrocyon are large 

 and fall into three groups, two opposing one another at the pylorus, the third posterior 

 to these and forming a sort of fringe to the intestine for about one-sixth of its entire 

 length (2"5 cms. in li cms.). In Sarcodaces the appendages are short and numerous, 

 and form a fringe to the intestine for about the first ninth of its course (I'o cms. in 

 14 cms.). In Ichthyoborus they are numerous and long, but fine, and fringe the 

 intestine for nearly one-sixth of its length (1-5 to 2 cms. in 12 cms.). In Macrodon they 

 are continued for about one-fourth or one-fifth of the length of the intestine (3-1 cms. 

 in about 13 cms.). In Dlstichodus the extension of the csecal fringe is at its maximum 

 amongst the forms I have examined. In this fish, the apj)eudages at the pylorus are 

 long, numerous, and fine, and are thence continued in an abbreviated condition as 

 a fringe or sacculated border for about 9 cms., which, however, is only about 

 one-seventh of the length of the intestine — very long in this and other herbivorous 

 forms. In Salmimis thev are lar^re and numerous, extendinsf about 2 cms. alono? 

 the intestine. In Anacyrtm, Tetragonopteriis, and Alestes I have noted that the 

 appendages were large. In Alestes macrolepidotus 12 and 11 respectively Avere covmted 

 in two si:)ecimens. They apparently formed a single row, fringing the intestine for 2 cms. 

 In Leporinns they are long and set in two rows of 7 each. In Bryconaethiops I made 

 out two rows of rather long appendages — about 5 in each row. In Xenocharax they 

 appear to be much reduced in number, size, and area, being in fact quite inconspicuous. 

 In Anostomns they are rather long but not numerous ; in Citharinus numerous, long, 

 tine, and restricted. In the other forms I have examined the api^endages were in such 

 an unsatisfactory state of jireservation that nothing can be confidently stated about 

 them. 



Looking at the foregoing observations as a whole, I find myself unable to do more 

 than simply state the facts which have come under ray notice. I may here, however, 

 point out the curious circumstance that Amia on the one hand, and the Cyprinidte and 

 Siluridfe on the other — families to which we are especially led to look for affinities with 

 the Characinidte — are entirely devoid of pyloric appendages, whilst the Gymnotidse, as I 

 have verified, possess them. 



