Errors in Length-weight Parameters at FishBase.org

Background: FishBase.org is an on-line database of fish related data that has been cited over 1500 times in the fisheries literature. Length-weight relationships in fish traditionally employ the model, W(L) = aL b , where L is length and W is weight. Parameters a and b are catalogued by FishBase for a large number of sources and species. FishBase.org detects outliers in a plot of log(a) vs. b to identify dubious length-weight parameters. Materials and Methods: To investigate possible errors, length-weight parameters from FishBase.org were used to graph length-weight curves for six different species: channel catfish (Ictalurus punctatus), black crappie (Pomoxis nigromacalatus), largemouth bass (Micropterus salmoides), rainbow trout (Oncorhynchus mykiss), flathead catfish (Pylodictis olivaris), and lake trout (Salvelinus namaycush) along with the standard weight curves (Anderson and Neumann 1996, Bister et al. 2000). Parameters noted as “doubtful” by FishBase were excluded. For each species, variations in curves were noted, and the minimum and maximum predicted weights for a 30 cm long fish were compared with each other and with the standard weight for that length. For lake trout, additional comparisons were made between the parameters and study details reported in FishBase.org for 6 of 8 length-weight relationships and those reported in the reference (Carlander 1969) for those 6 relationships. Results: In all species studied, minimum and maximum curves produced with the length-weight parameters at FishBase.org are notably different from each other, and in many cases predict weights that are clearly absurd. For example, one set of parameters predicts a 30 cm rainbow trout weighing 44 g. For 30 cm length, the range of weights (relative to the standard weight) for each species are: channel catfish (31.4% to 193.1%), black crappie (54.0% to 149.0%), largemouth bass (28.8% to 130.4%), rainbow trout (14.9% to 113.4%), flathead catfish (29.3% to 250.7%), and lake trout (44.0% to 152.7%). Ten of the twelve extreme curves reference two sources (Carlander 1969 and Carlander 1977). These two sources are used for a total of 100 different species at FishBase.org. In the case of lake trout, comparing the length-weight table at FishBase.org and the cited source (Carlander 1969) revealed that while 5 of 6 total length measurements were incorrectly reported as fork lengths by FishBase.org, all parameters accurately reflected the source. Comparing the length-weight relationships of the source (Carlander 1969) with the table of weights in different length ranges reveals the length-weight parameters in the source are clearly in error. However, FishBase.org also neglects to specify clearly distinguished subspecies and/or phenotypes such as siscowet and humper lake trout. Conclusion: Length-weight tables at FishBase.org are not generally reliable and the on-line database contains dubious parameters. Assurance of quality probably will require a systematic review with more careful and comprehensive methods than those currently employed.