Premium
Predicting Whole‐Fish Nitrogen Content from Fish Wet Weight Using Regression Analysis
Author(s) -
Ramseyer Laurel J.
Publication year - 2002
Publication title -
north american journal of aquaculture
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.432
H-Index - 41
eISSN - 1548-8454
pISSN - 1522-2055
DOI - 10.1577/1548-8454(2002)064<0195:pwfncf>2.0.co;2
Subject(s) - fish <actinopterygii> , biology , linear regression , regression analysis , body weight , regression , standard error , zoology , mathematics , statistics , fishery , endocrinology
Regression analysis was examined as a method of estimating whole‐body nitrogen (N) from whole‐fish wet weight for a broad range of fish species. Whole‐body N and corresponding wet weight data were collected from the literature for 60 fish species and 6 hybrids. Log 10 ‐transformed data were used to regress whole‐body N on fish wet weight for each species. At least 98% of the variation in whole‐body N was explained by wet weight for 60 of the 66 species and hybrids ( P < 0.001). Regression parameters were similar for most species of fish and were often statistically indistinguishable ( P < 0.05) for species within a genus or other taxonomic grouping. Paired t ‐tests showed no significant differences ( P > 0.05) between whole‐fish N values determined chemically and those predicted by linear and broken‐line regression equations. A linear regression equation with combined data for all species and hybrids was log 10 (fish N) = 1.03 × log 10 (fish wet weight) − 1.65 ( n = 2,811; r 2 = 0.996; SE Y · X (the standard error of the Y estimate) = 0.054; P < 0.0001). Estimation of whole‐fish N using regression equations presents a useful alternative to direct chemical analysis for selected applications.