COHEN and RVSSO VARIATION IN FOL'RBEARD ROCKLING 



concordance in patterns of variation. Differences 

 between and similarities among samples are 

 summarized in Figure 4 and discussed below for 

 meristics and color pattern. Differences in morph- 

 ometric characters are so slight that we do not 

 further consider them. 



Gulf of Mexico and Southern Atlantic samples 

 are quite similar, although at this time the two 

 might be semi-isolated from each other. The 

 clockwise loop current system in the Gulf of 

 Mexico provides a possible pathway for the disper- 

 sal of young, pelagic stage rocklings out of the gulf; 

 there is no present avenue for recruitment into the 

 Gulf of Mexico. If the single rockling taken off the 

 Florida Keys represents more than a stray, then 

 perhaps Gulf of Mexico and Southern Atlantic 

 populations are continuous; otherwise, the north 

 Gulf-northeast Florida distribution pattern is 

 similar to that noted first in fishes by Ginsburg 

 (19521. Although E. cimhnus seems rare in the 

 Gulf of Mexico its occurrence at two widely sepa- 

 rated localities, with a collection of 16 individuals 

 from one of them, indicates that the .species is estab- 

 lished there. Although pelagic stages have not yet 

 been taken from the Gulf of Mexico or Southern 

 Atlantic areas, it seems reasonable to assume that 

 they occur there and are available for dispersal in 

 the Gulf Stream system. 



Rocklings from the Intermediate area are in- 

 deed intermediate between adjacent populations 

 to the north and south in degree of pigmentation 



and in dorsal and anal ray counts. Furthermore, 

 for two of these characters, color and number of 

 anal fin rays, the standard deviation is larger than 

 in other populations, implying that recruits from 

 different spawning populations are entering the 

 area or that the characters are genetic and variabil- 

 ity is being maintained during spawning in the 

 Intermediate area. For two characters, numbers of 

 vertebrae and pectoral fin rays. Intermediate and 

 Northern Atlantic fish are nearly identical and 

 differ from Southern Atlantic and Iceland sam- 

 ples. These characters must be determined or 

 mediated differently than are color pattern and 

 dorsal and anal fin ray counts. Gill raker count 

 appears to reflect still a third method of character 

 determination as the means are different on the 

 two sides of the Atlantic. Although pelagic early 

 stages have not been taken in the Intermediate 

 area, they may be available for dispersal to the 

 northeast by means of the Gulf Stream and to the 

 southwest in coastal currents that parallel the 

 Gulf Stream. Such dispersal patterns would help 

 to account for the occurrence of dark-colored rock- 

 lings in the north and light-colored ones in the 

 south. 



Rocklings from the Northern Atlantic area 

 more closely resemble fish from Europe and Ice- 

 land in degree of pigmentation and number of 

 vertebrae than they do their immediate neighbors 

 to the south. Conversely they are closer to other 

 North American samples in numbers of pectoral 



CHARACTER 



Gulf 



S. Atl. 



GEOGRAPHICAL AREA 



Intermed . 



N. Atl. 



Iceland 



Europe 



Color 

 Anal Rays 

 Dorsal Rays 

 Vertebrae 

 Pectoral Rays 

 Gill Rakers 



6.2 



40.2 



47.0 



51.4 



17.2 



6. 7 



40.8 



47.9 



51.! 



17.2 



3.8 



42.8 



43.5 



1.9 



1.) 



42.6 



49.2 



FIGL'RE 4. — Summary of means of character states for Enchelyopu^ cimbnus from six geographical areas. Heavy lines are drawn 

 around entries that are discussed in the te.xt as separate entries and that illustrate overall lack of convergence m character states. 



97 



