548 ■ MEMOIRS OF THE CARNEGIE MUSEUM. 



structure may be prevailingly reticulate {berthoe, enno), or prevailinglj' septiform 

 (gratus, decephis). 



4. D. paranensis has the marsupium in the middle half of the gill, and in- 

 terrupted septiform structure is present. 



5. In eight species, hooks are known to be present on the glochidia. Of 

 these six belong to the chilensis-grou\), one to the charruanus-group (piceus), one 

 to the lacteolus-group {mogyniirim). Glochidia without hooks (but possiblj' im- 

 mature) were found in charrvavus and bcrtho' (charruamis- and elUpticus-gvoup?.). 

 Two species have margined glochidia, hosemani of the hylceits-grovip, and hildce 

 of the charruanus-group. 



For the present, these facts do not furnish any clue as to the relationship of 

 the species or groups, they even are rather confusing, partly upsetting the divisions 

 arrived at by the study of the shell. But they should be carefully recorded, 

 because additional material may throw more light on the problem. 



Genus CASTALINA Von Ihering (1891). 



Von Ihering, 1891, p. 478; 1893, p. 73; Simpson, 1900, p. 865; 1914, p. 1204. 



Type-species. — C. martensi Von Ihering (designated by Simpson). 



The chief characters of this genus are found in the general shape of the shell, 

 which is subtriangular or subquadrate, with a well developed posterior ridge and 

 a sub truncated posterior slope, which, however, is somewhat elevated in the middle. 

 In addition, the beaks are rather elevated, the interdentum is well developed, form- 

 ing a rather deep beak-cavity. The hinge-teeth often are provided with parallel 

 ridges. 



The anatomy is similar to that of Diplodon. However, there is a tendency to 

 close the branchial opening in front, yet this is not always the case, so that this 

 character is variable, not only sjiecifically, but also individually. 



The genus has been well treated by Von Ihering (1893), and a key for the 

 species has been given (p. 83). Simpson (1914 p. 1205) also gives a key, but C. 

 undosa should be excluded; it is a Castalia. 



29. Castalina nehringi Von Ihering (1893)."^ 



Diagram of soft parts: text-fig. 2, p. 456; Anatomy of gills: PL XLVII, fig. 2; Section 

 of gills: Plate XLVIII, fig. 5; Glochidium: text-fig. 41, p. 469. 



Type-locality. — Rio Piracicaba, Sao Paulo, Brazil. 



^- Of both C. nehringi and martensi the specific names were first mentioned by Von Ilierins in 1891 

 (p. 477), but as nomina nnda without descriptions. 



