g6 Memoir Sears Foundation for Marine Research 



5b. Snout not flattened dorsoventrally, its dorsal contour convex; 



teeth along rear part of lower jaw small and inconspicuous ; pre- 



maxillary bones close together in front; mesethmoid (proeth- 



moid) bone not paired. Clupeoidea, Part 3, p, 148. 



4 b. Adipose fin between rayed dorsal fin and caudal present in most, but 



lacking in a few (see 6 b). 



6 a. Swim bladder connected to oesophagus throughout life by 

 an open pneumatic duct (physostomic) ; premaxillary bones 

 well developed; jaws with well developed teeth except in the 

 Coregonidae, which are either toothless or have minute teeth; 

 eyes of ordinary type; oviducts incomplete (p. 455), so far as 

 is known. Salmonoidea, Part 3, p. 455. 



6 b. Swim bladder not connected to oesophagus during late 

 stages of growth (physoclystic);^* premaxillary bones small or 

 lacking; jaws toothless; eyes tubular in some but not in others; 

 oviducts incomplete in some,^" perhaps complete in others. 



Argentinoidea,^^ Part 4. 



3 b. Luminescent organs (photophores) present on head, on body, on tail 



sectors, or on all three; their arrangement in longitudinal rows on sides 



extending anterior to anus as well as along base of anal fin ; no tubular 



papilla on shoulder. Stomiatoidea (including Suborders Gymno- 



photodermi and Lepidophotodermi^^), Part 4. 



Families. The Isospondyli plus the Iniomi include about 55 families on a world- 

 wide basis, about 45 of which are known to occur in the western North Atlantic. At 

 first glance, this may seem an appalling number among which to choose when one has 

 to run down a given specimen to its proper family, the more so because there is no 

 one conspicuous character or combination of characters visible from the outside on 

 which anyone but a professional taxonomist can rely to tell him whether his fish be- 

 longs among the Isospondyli or among the Iniomi. The task, however, is not as dif- 

 ficult as one might expect, for nearly all of the western Atlantic families (or subdivisions 

 of families) of the two Orders combined are separable, one from another, by features 

 that are visible from the outside, that are susceptible of rather precise definition, and 

 that are mutually exclusive, or nearly so. 



However, in numerous instances a family, because of the diverse characters of 

 its genera and species, has been split into two or more parts. Hence the same family 



29. Cuvier and Valenciennes seem to have been the first to record this fact {ov Argentina (14: 411). Trewavas' interpre- 

 tation of the ventral "sole" of Opisthoproctus as a swim bladder is now known to have been erroneous (5J: 610). 



30. This is the case for Argentina, Opisthoproctus, and Dolichopteryx (Part 4). For Macropinna, however, Chapman 

 reported that a "fine duct" extends from the posterior end of each ovary, but he could not determine whether these 

 ducts open into the urinary bladder or into the rectum (5: 293). 



31. For a detailed comparison of the Argentinoidea and Salmonoidea, see Argentinoidea, Part 4. 



32. Proposed by Parr, 33: 15-17. 



