48 U.S. NATIONAL MUSEUM BULLETIN 271 



of a group as well as the possibility that the observer will handle 

 undescribed genera and families. The need to cross-check diagnoses and 

 test keys to genera in several families is a part of the identification 

 process. 



The textual diagnoses, descriptions and relationships of the 53 

 famiUes are presented in alphabetical sequence. The Hyalidae, Hyal- 

 llidae, and TaUtridae are combined into the superfamily Talitroidea 

 for ease of identification. Each family has a written key to its genera 

 and each family is accompanied by several sets of figures illustrating : 

 (1) the characters of the generic keys, and (2) the variation in lateral 

 aspects and appendages of the several genera. 



Perhaps the key to success in using the diagrammatic keys is the 

 need to memorize the plan of the basic gammaridean and to keep an 

 account of generalized characters, especially the generalization of 

 heads, feeble and strong gnathopods, anterior coxae, and elongation of 

 telsons and pereopods. Thus, Synopiidae are extremely difficult to 

 recognize because the subtle shapes of their heads are primary to their 

 identification. Acanthonotozomatidae, Paramphithoidae, and Stego- 

 cephalidae are recognized mainly by their anterior coxae. 



Written keys. — If the Diagrammatic Key to Families does not 

 produce results one may utilize the written key to families (pp. 106-116) . 

 The written key is not as satisfactory as the Diagrammatic Key be- 

 cause the observer must repeatedly make "yes or no" decisions on 

 individual characters and no accomodation for undescribed genera and 

 families can be made. Simplicity of the keys requires occasional 

 repetition of taxa in order to account for morphological diversity. 



Extraordinarily long keys are occasionally divided into sections in 

 order to improve their usefulness. The establishment of subgroups in 

 this manner is believed to be of assistance in retaining conceptual 

 images while the observer is proceeding through the maze of an 

 extremely complicated key. 



Diagnoses. — Familial diagnoses are compatible with the Diagram- 

 matic Key to Families and are not mutually exclusive, for the diagnoses 

 only differentiate the families from the basic gammaridean. Inter- 

 related famihes are Hsted as a part of the diagnoses and are discussed 

 under "Relationships." These discussions are not necessarily repeated 

 in all possible places and the reader may have to turn from the 

 Synopiidae to the Liljeborgiidae, for example, in order to find a dis- 

 cussion of the interrelationships between the two families. 



Genera are diagnosed in as brief and consistent form as possible, in 

 direct relationship to the keys. Where two genera are mutually dis- 

 tinguished by a character of no taxonomic importance to other genera 

 of a family, the alternatives of the character are italicized and the 

 related genus is stated. Such character alternatives are omitted from 



