The Compleat ELEFAN, Version 1.1 : Post-Release Blues FELIMON GAYANILO,JR. DANIEL PAULY ICLARM Abstract This article discusses differences between the newly released Compleat ELEFAN Software Package (Vers. 1.1) (Gayanilo and Pauly 1989, Fishbyte 7(2): 20-21) and its predecessor (Vers. 1.0). Also, a number of ''bugs'' found in Version 1.1 are documented; this led to release 1.11, which costs U5$75. Introduction For various reasons, notably lack of space, the two-page contribution in Vol. 7(2) of Fishbyte in which we announced Version 1.1 of the Compleat ELEFAN software package failed to describe some of its differences with the previous version 1.0. As these differences may create confusion among users, (e.g., by leading to different estimates of .some parameters), we shall list the most important of these differences. Then, we shall deal with "bugs". The following text doesn't really make for fascinating reading but unfortunately, it is unavoidable. Differences between Versions 1.0 and 1.1 The differences between the two releases are presented by program, as they appear on the main menu. "Now" refers to Version 1.1 ELEFANO • The data entry routine of Version 1.1 automatically identifies and stores, for any file, the smallest and largest size classes with nonzero frequencies, as required for particular routines, especially those related to VPA. It is still possible to expand the size range covered by a file, by adjusting - using the edit routine the smallest and/ or largest midI ength (s). • Length-at-age data may now be entered and edited (the file this created can be used only within ELEFAN V). April 1990 Fishbyte 8(1):47 - 49 • An option for an ASCII disk output has been incorporated in the PRINT ROUTINE of ELEFAN O. The ASCII file saved on disk can be imported to a wordprocessor or an electronic spreadsheet (e.g., LOTUS 1-2-3). ELEFANI • The seasonally-oscillating form of the von Bertalanffy growth function proposed by Pauly and Gaschiitz (1979) has been replaced as growth model for ELEFAN I to V by the related model suggested by Hoenig and Choudary Hanumara (1982) and Somers (1988) (see also Hoenig and Choudary Hanumara, Fishbyte, this issue). The differences between the growth curves generated by these two models are very small, generally on the third decimal place or less when dealing with fish length expressed in em. • The routine for computation of the Explained Sum of Peaks (ESP) was modified. The modification concerns the relationshsips between four quantities: - the current asymptotic length (L oo ) - the mid length of the largest length class (L max ) - the upper limit of the largest length class included in the file (L'max) - the largest fish included in the computation of ESP (L t ). In Version 1.0, the computation of ESP stopped, when L oo < L max , as soon as L t 0.95 .. L oo • Also, the computation stopped when L t > L max ' Thus, in both cases, L'max was not considered. In Version 1.1, the computation of ESP stops when L oo < L'max' as soon as L t 0.95 .. LOO' Also, the computation stops when L t > L'max' These modifications - which render the ESP computation routine consistent with our initial intentions - will not affect growth parameter estimates when L oo > L'max' but may lead to changes when L oo < L'max' particularly when 47 Gayanilo, F.C. Jr. and D. Pauly. 1990. The Compleat ELEFAN, Version 1.1: Post-release blues. Fishbyte, Newsletter of the Network of Tropical Fisheries Scientists 8(1): 47-49.