MUSICK: COMPARISON OF THE HAKES 



1 U ftttlHW PIE 0-~0 



2 U tfuit NS. • • 



4 U e*uia * « 





S <i^ 



STANDARD LENGT- 



Figure 11. — Regression of head length on standard 

 length from three samples of Urophycis tenuis and one 

 of U. chiiss. 



Regression of Preanal Length on Standard Length 

 (Figure 12, Table 1) 



All regressions were linear. U. te}iui!< P.E.I, 

 was significantly different from U. tenuis N.E. 

 at the 5% level and from U. tenuis N.S. at the 

 1% level. There was no significant difference 

 between the latter two samples. U. chnss was 

 significantly different from U. tenuis N.E. and 

 U. tenuis N.S. at the 1% level and from U. tenuis 

 P.E.I, at the 5% level. Although U. chuss is 

 statistically different from all samples of U. 

 tenuis:, there is much overlap between them 

 and the character is not useful in species 

 identification. 



DISCUSSION 



There were significant differences between 

 L^ chuss and all samples of U. tenuis for all 

 meristic characters except the total number of 

 vertebrae. The ranges of the numbers of lateral 

 line scales of the two species did not overlap, 

 and this character can be used with confidence 

 to distinguish between them. 



U. chuss invariably had three gill rakers on 

 the epibranchial of the first gill arch, U. tenuis 

 had but two. The difference is diagnostic. Mujib 

 (1967) was in error when he reported that the 

 epibranchial of U. chuss was devoid of gill 

 rakers. 



The numbers of abdominal vertebrae for each 

 of the species showed little variability, most 

 U. toiuis having 16 and most U. chuss having 

 15. This character may be valuable in identify- 



STANDARD LENGTH 



Figure 12. — Regression of preanal length on standard 

 length from three samples of Urophycis tenuis and one 

 of U. chuss. 



ing collections of postlarval and juvenile hakes 

 in which the scales have not yet formed and 

 when the mean number of abdominal vertebrae 

 may be computed for an entire sample (Robert 

 Marak, pers. comm.) but should not be relied 

 upon for identification of single specimens. 



None of the meristic characters showed 

 statistically significant differences among the 

 samples of U. tenuis. Such differences may 

 exist but could have been masked by the group- 

 ing of several year classes in each sample 

 (Appendix). 



Morphometric analyses (Table 1) show that 

 U. tenuis N.E. and U. tenuis N.S. are quite 

 similar. U. tenuis P.E.I, differs from both of 

 the latter and, of the three, is the most similar 

 to U. chuss. This pattern suggests character 

 displacement in U. tenuis off New England 

 where it is sympatric with U. chuss. However, 

 final judgment on the existence of character 

 displacement should be withheld until small 

 specimens of U. toiuis from the Gulf of St. 

 Lawrence are available for examination. Mor- 

 phometric differences between U. tenuis in 

 the Gulf of St. Lawrence and those from Nova 

 Scotian and New England waters may have a 

 strong biological basis. The Gulf of St. Lawrence 

 group gathers in large spawning aggregations 

 during June in the Northumberland Strait 

 (Stephen Nepszy, pers. comm.). whereas the 

 Nova Scotia-New England group probably 

 spawns in the fall (Musick, 1969). Thus, the 

 two groups may be reproductively isolated. 



485 



