HANNIBAL: CALIFORNIAN I'EESHWATEE MOLLTJSCA. 117 



ridges remaining, particularly on the posterior portion of the shell, 

 -while the broad hinge and the sub-quadrate shape seem to have 

 suffered little change.' 



In the next stage, of which Migranaja littoralis, Elliptio crassidens, 

 and Lampsilis luteolus may be taken as representatives, the sculpturing 

 lias been carried back to the umboues, the shells become posteriorly 

 elongated and the hinge-area much reduced in breadth. Forms such 

 as Margaritana inargaritifera falcata, Stropitus edentulus, and Alasmi- 

 donta marginata represent the next condition, in which the laterals 

 have become obsolete, leaving the pseudo-cardinals as tubercles, while 

 in typical Anodonta the shell has become elongate-elliptical, the 

 umbones decidedly anterior, and the hinge simple, edentate, and 

 gently curved. 



This series of changes is not characteristic of any one group of 

 iN'aiades, but may be observed with more or less modifications in all, 

 irrespective of the special line of descent. In other words the 

 entire superfamiiy exhibits an instance of parallel development, and 

 the dissimilarities of the species as one sees them are due to the 

 different stages in the series reached by that species and the degree 

 of unequal parallelism developed by that particular line of descent 

 compared with the whole. Since specialization without differentiation 

 is not verj- valuable in classification, the shell characters become of 

 secondary importance in the recognition of groups more comprehensive 

 than genera. 



In the anatomy, liowever, one finds a sound foundation on which 

 the higher divisions may be built. Here the specialization which 

 takes place tends to ultimately result in the production of a parasitic 

 larva (unlike any stage in the develoj)ment of the race) which must 

 pass through a complete metamorpliosis before reaching the adolescent 

 condition, and a specialized marsupium for its development. This 

 reaches its culmination only in such genera as Spatha, Anodonta, and 

 Proptera, but all the higher forms show indications of approaching it. 

 It is evident that parallelism likewise exists in the anatomical 

 modifications, in that all the groups modify towards a certain definite 

 end, yet the means bv which this is accomplished are diverse, and 

 the resulting marsupia and larvte are not homologous but strictly 

 analogous. 



Margaritana (Margaritaninfe) represents in its anatomical features, 

 independent of the modifications of the shell, the oldest and least 

 differentiated type of Naiad structure. The gill-septa, which in the 

 more specialized groups (except the Hyriinoe) coalesce completely 

 to form the water-tubes, are present on the inner faces of each gill- 

 lamella. These are never united, but instead regularly beaded with 

 low-rounded papillse lying in rows at right angles to the septa. At 

 occasional and irregular intervals, perhaps once to a septum, its 

 development becomes more pronounced, and one or more of the 



^ That this shape is really primitive is evident from the inspection of the young 

 stages of almost any Unioid ; cf. pi. v, fig. 7, or Lefevre & Curtis, Journ. 

 Exp. ZooL, ix, pi. iv, fig. 29, 1910. 



