Orr and Matarese; Revison of the genus Lepidopsetta Gill, 1862 



543 



of species were generated by using the geographic infor- 

 mation system Arc/Info and Arc View (Environmental Sys- 

 tems Research Institute, 1996). Tables were produced with 

 Arc View and Systat software (SPSS, Inc, 1996). Densities 

 were mapped by dividing the geographic area into contig- 

 uous square polygons, 625 km- each. Catches per unit of 

 effort (CPUEs) from all tows (including catches) within 

 each square polygon were averaged, yielding a mean den- 

 sity. Polygons were shaded according to their mean density, 

 or left unshaded when no tows occurred within a polygon. 



Adult statistical analyses 



The software programs S-Plus (Statistical Sciences, 1993) 

 or Statgraphics Plus 2,1 ( Manugistics, 1997) were used 

 in statistical analyses performed on juveniles and adults. 

 Unless otherwise indicated, tests were considered signifi- 



cant at P < 0.05, as adjusted by the Bonferroni correction 

 (Sokal and Rohlf 1995). For all characters, Bartlett's test 

 of homogeneity of variances (Sokal and Rohlf 1995) was 

 used to determine the appropriateness of an ANOVA to 

 test for differences between species. When Bartlett's test 

 found significant differences in variances, pairwise compar- 

 isons were made. Pairs with significant differences in vari- 

 ances were then tested for differences in medians by using 

 the nonparametric Kruskal-Wallis test (Sokal and Rohlf 

 1995). The following meristic characters did not differ sig- 

 nificantly in variance among species pairs and were sub- 

 jected to an ANOVA: gill rakers of the upper and lower 

 first arch and lower second arch; dorsal- and anal-fin rays; 

 ocular-side and blind-side pectoral-fin rays; posterior and 

 anterior /\DB pores; interorbital scales; cheek scales; scales 

 above, below, and total around the lateral line: ocular-side 

 and blind-side suborbital pores; supraorbital pores. 



