MULLIDAE OF THE WESTERN NORTH ATLANTIC 



405 



Measurements were made and plotted on dis- 

 tance from Ih-st to second dorsal fins and caudal 

 concavity, but the graphs are not included because 

 of the hi<,'h degree of variability and nnml)er of 

 factors that influence these dimensions. 



ABBREVIATIONS AND TERMINOLOGY 



I have followed Berry (1959) in terminology 

 of the developmental stages. 



Larval. — Defined by Bern- as, "the stage of de- 

 velopment commencing with hatching and ending 

 with the completion of formation of the adult 

 complement of fin rays." 



Only the larvae of MvUtis auratus are consid- 

 ered here. In M. auratus the secondary rays of 

 the caudal fin are the last rays to ossify (shown on 

 stained material). This ossification occurs at 10 

 or 11 mm. 



Juvenile. — Beriy (1959) defined this stage as, 

 "commencing with the end of the lai-val stage and 

 terminating with the attainment of sexual matu- 

 rity." All of my specimens of P. ma-eulatus (small- 

 est, 18.5 mm.), U. parvus (smallest, 20.8 mm.), 

 and Mulloidlchthys martmicus (smallest, 82.5 

 mm.) have reached the juvenile stage. 



Adult. — Defined by Beriy (1959) as, "commenc- 

 ing with the attainment of sexual maturity." I 

 did not determine at what size this occurs on any 

 of the species under consideration. 



I have made no attempt to identify the chro- 

 matophores either histologically (Gordon, 1959) 

 or chemically. I suspect, however, that in the goat- 

 fishes some of the chromatophores, particularly 

 those in the fins, remain migratory in much larger 

 specimens than has been shown in some other fishes 

 (Parker, 1948). 



In discussions of development of the rays, the 

 terms "branching" and "segmentation" refer only 

 to primary branching and segmentation in a ray. 

 "Branching is complete" or "segmentation is com- 

 plete" refer only to the fact that all rays that are 

 to become branched or segmented have done so at 

 least once. 



Some of the specimens examined were collected 

 in locations that are not widely known. For these 

 I have supplied approximate latitudes and longi- 

 tudes. 



A few of the small specimens had no data re- 

 garding method of capture and were taken at sta- 



643096 0—62 2 



tions listed in tlie vessel station lists as "longline," 

 "hook and line," or "drift" stations. Since it is 

 customaiy for participants of a cruise to use a 

 dip net at stations and since it is improbable that 

 a specimen under 40 nun. should have been caught 

 by these other means, I have designated them as 

 (D?) specimens. The abbreviations used in refer- 

 ring to the different collections are given on page 

 455. The Gill cruises made three different types 

 of stations that I will refer to: Regular (Reg.), 

 Special (Spc), and Standard (Std.) (Andei-son, 

 Gehringer, and Cohen, 1956). 



The methods of capture are indicated as fol- 

 lows: trawl (T), dip net (D), stomach contents 

 (SC), market procured (M), and rotenone (R). 



Key 



A. 



to mullidae of the western north 

 Atlantic 



Material 10.0-82.5 mm." 



Minute spine at origin of spinous dorsal 



(total of 8 spines) (fig. 28). 



B. Lower-limb gill rakers 11-16; over 



17 mm., palato-vomerine tooth patch 



(fig. 31) ; no opercular spine (fig. 



33) Mullus auratus. 



BB. Lower-limb gill rakers 17-24; no 

 palato-vomerine tooth patch; over 

 22 mm., opercular spine (fig. 23) 



Pseudupene us macula tus. 



AA. No minute spine at origin of spinous dorsal 

 (total of 7 spines) (fig. 37). 

 Lower-limb gill rakers, 16-19: cerato- 

 branchial bone gill rakers, 10-12 ; no oper- 

 cular spine (fig. 38) Upeneus parvus. 



As no MuZloidichthys martinieus below 82.5 mm. 

 have been examined or described, they are omitted 

 from this key. Between 20 and 80 mm., they 

 could be expected to be devoid of a visible minute 

 spine at the origin of the spinous dorsal (fig. 40) 

 and to have a lower-limb gill raker count of 19-23 

 and ceratobranchial Ixme gill raker I'ount of 13-10 

 (allowing for formation of one additional raker 

 between 20 and 80 mm.). As the opercular spine 

 of P. jnaculatus is developed at 22 mm., the same 

 trend might be exjiected in M. nuirtinicu.'i. a 

 closely related species. 



2 Based upon 30 sppclmens of hi. auratun. 4S /'. mnculalim. and 

 40 U. parvus. 



