MARINE GAMMARIDEAN AMPHIPODA 23 



cause of vastly modified mouthparts or of so-called pygidization, the 

 solidification of the urosome and its appendages by segmental coales- 

 cence or loss of uropodal rami and peduncles. Of course, other per- 

 acarids mthout lateral shield may also have tube-spinning glands 

 (tanaidaceans) and the primitive amphipod may have been in the 

 glandular line. There are sufficient intergrades in telsonic morphology 

 between isaeids and gammarids to make this suggestion very attrac- 

 tive. Several isaeids are almost perfect replicas of the basic gammarid- 

 ean except for their fleshy telsonic nobs. One has to balance at least 

 two alternatives, whether the coalesced telson represents a full segment 

 lost in phylogeny or whether the lobed telson of gammarids represents 

 a pair of appendages long lost. This problem is fundamental to other 

 crustacean orders and my impression is that the uncleft, solid telson 

 is the most common, thus suggesting that it is either more primitive 

 or at least more successful functionally. The telson in Amphipoda is 

 not a conservative feature by any means, and our understanding of it 

 as an evolutionary marker will not be clarified until we understand its 

 function. Several cycles of morphological development, whether by 

 advancement or regression are apparent. If the fleshy isaeid telson 

 becomes ''falsely" lobed in advanced species one can almost imagine 

 the small step necessary to convert it into one of the poorly lobed or 

 entire "subfieshy" members of the Gammaridae. Extreme fiattening, 

 full clefting, and elongation then occur in other gammarids and 

 various derived families, but coalescence of the lobes and shortening 

 occur again and again. In some haustoriids the telsonic lobes become 

 fully disjunct basaUy, each lobe appearing as a vestigial appendage. 



The fully blown pygidization of Gammaridae, Isaeidae, and most 

 gammarideans, in the sense of having the last three pairs of abdominal 

 appendages formed into relatively inflexible, posteriorly directed 

 uropods, obstructs our detection of an ancestor in any other living 

 order of Peracarida, where only the final pair of appendages is formed 

 into uropods. The reduction or loss of uropods, pleopods, and most 

 of the abdomen in CaprelHdea is clearly a secondary development. 

 Several good intergrades occur in this procession from Gammaridea 

 to Caprellidea in such taxa as the Podoceridae, Caprogammaridae 

 and Cercops, a caprelfidean. Most CaprelHdea further have thoracic 

 somite 2 (free segment 1 of other Amphipoda) coalesced mth the 

 head, have a reduction in pereopods 1-2, gills and brood plates. If 

 the lateral shield or its functional substitute by means of tubicoly, 

 serve as partial protection for brood and gills in Gammaridea, then 

 CaprelHdea, with their complete loss of lateral shield must have 

 some other protection, perhaps reflective of their habitats or behavior. 

 The brood lamellae of caprelHds seem to be more strongly cornified 

 than those of gammarideans. The lateral shield has also been suggested 



