One of the original purposes of FishBase was to make available to researchers some of the wealth of available data on various aspects of the biology of fishes.
However, before such data can be analyzed, an overview of their key features is necessary, and for this, we provide numerous ‘active graphs’, constructed on demand by FishBase from records in one or several of its tables after a graph button has been pressed.
These graphs presently come in four different forms:
as pie charts (e.g., for diet composition data);
as time series (e.g., of nominal FAO catches);
as plots of mathematical functions (for length/weight relationships and von Bertalanffy growth curves);
as frequency distributions of important variables;
as bivariate plots of a few records pertaining to a (group of) species, superimposed (in red) on yellow dots representing all other species for which FishBase has records; and
as 2D or 3D graphs illustrating interactive routines.
Items (1-4) do not require much comment, except to point out that we will continue to try to improve their design, based on concepts from Tufte (1983).
Item (5) is an idea first introduced in FishBase 96, which we are quite proud of as it resolved through simple graphs, in one fell swoop, a number of problems associated with the numeric records that they illustrate:
the records (in red) for a given species or group are accessible in a bivariate context, and hence their magnitude can be directly visualized;
the backdrop formed by the other species (in yellow) allows a direct evaluation of whether the red records are relatively high, or low, or average;
patterns in the data can be detected visually, thus encouraging hypotheses formulation and further analyses; and
outliers (yellow or red) can be immediately detected and, if found to be correct, used to generate further hypotheses.
FishBase features at least one, and often more graphs for most of its tables and forms.
Thus, while this profusion of graphs makes the data in FishBase much more visible, we have also been able to develop a new role for FishBase, that of presenting the data that test major hypotheses concerning the biology of fish, or the status of fisheries.
Examples of new graphs testing previously formulated hypotheses are our plot of frequency distribution of predator vs. prey size (see Fig. 41), which tests an important theory of Ursin (1973), or our plot of DNA contents per fish cell vs. the aspect ratio of their caudal fin (Fig. 58), which presents a first direct test of Hinegardner’s hypothesis on the DNA content of fish cells (Hinegardner 1968; Cavalier-Smith 1991).
Examples of graphs illustrating newly discovered relationships are our plots of trophic levels of fishery catches vs. time (see Fig. 4). These graphs were recently presented in the primary literature (Pauly et al. 1998), and had a huge media impact, as they illustrate extremely worrying trends (see e.g., Holmes 1998; Stevens 1998).
Fig. 4. Trend of the mean trophic level of fishery landings in FAO area 27 (N.E. Atlantic). Note the steady
decline, indicating a gradual transition from large piscivorous to small planktivorous fishes and invertebrates in
the landings.
The graphs, often constructed from data in several FishBase tables are not always straightforward to interpret, as was noted in a review of FishBase (Wootton 1997). Therefore, we introduced boxes (see Box 2) as devices for explaining the theory behind a given graph, and the related table entries, and to explain how the graph was built, and/or should be interpreted. These boxes, which may be viewed as miniature papers, are authored, and FishBase collaborators are welcome to contribute material for additional boxes (and/or the related graphs) for future publication, following the examples provided here.
Many chapters of FishBase 2000: Concepts, Design and Data Sources include boxes presenting material relevant to, but not part of, the main narrative.
The use of boxes to present such material is to provide details on data selection, algorithms, assumptions and implications, especially of the data used for the construction of graphs, and to provide backgrounds for, and first interpretations of these graphs.
Boxes are authored, and we invite FishBase collaborators and users with interpretative comments on tables or procedures to submit them in form of boxes to be included in future releases of the FishBase book.
Daniel Pauly and Rainer Froese
Here again, suggestions from FishBase users or collaborators are welcome, as are offers to jointly develop new routines.
The Internet version of FishBase contains already many of these graphs, accessible either under the respective tables attached to a species, or under the ‘Information by Family’ section if you click on the Graphs radio button.
I thank FishBase programmers Portia Bonilla, Alice Laborte and Ma. Josephine France Rius for their patience when implementing even my most outlandish design ideas, and Felimon ‘Nonong’ Gayanilo, Jr. for the first interactive graphs in FishBase.
Cavalier-Smith, T. 1991. Coevolution of vertebrate genome, cell and nuclear sizes, p. 51-85. In G. Ghiara et al. (eds.) Symposium on the evolution of terrestrial vertebrate, selected symposia and monographs. U.Z. 1, 4 Mucchi, Modena.
Hinegardner, R. 1968. Evolution of cellular DNA content in teleost fishes. Am. Nat. 102:517-523.
Holmes, B. 1998. The rape of the sea: fishing fleets are rampaging their way down the marine food chain. New Scientist 157(2121):4.
Pauly, D., V. Christensen, J. Dalsgaard, R. Froese and F. Torres, Jr. 1998. Fishing down the food webs. Science 279:860-863
Stevens, W.K. 1998. Man moves down the marine food chain creating havoc. New York Times (10 February 1998):B12.
Tufte, E.R. 1983. The visual display of qualitative information. Graphics Press, Cheshire, Connecticut. 197 p.
Ursin, E. 1973. On the prey preference of cod and dab. Medd. Danm. Fisk. Havunders. N.S. 7:85-98.
Wootton, R.J. 1997. Review of FishBase 96. J. Fish Biol. 50(3):684-685.
Daniel Pauly