Introduction FultonÕs condition factor (K) is a well established parameter in fishery management. A fish population is dependent upon a variety of internal (genetics, stages of development) and external (e.g. food availability, water quality etc.) parameters. Variations of K between species are greatly influenced by the body shape, which is well described by the form factor (Froese, 2006). The aim of the paper was to report condition factors for 43 fish species from Croatian freshwaters and to determine the relationship between K and the form factor (a 3.0). Materials and methods Mean condition factors (K) for 279 fish populations (all total lengths in centimetre and weights in gram) were gathered from freshwaters throughout all regions of Croatia (Habekovicét al., 1993; Habekovic 1994a,b; Habekovicánd Pazr, 1998; Treer et al., 1995, 2000, 2003a,b, 2005, 2006, 2008b; Jakovlicánd Treer, 2001; Š prem et al., 2001, 2005; Bakota, 2002; Bakota et al., 2003; Piria et al., 2006; Prpa et al., 2007). For each species the mean K with respective 95% confidence limits (CL) was calculated. The regression between these means and the respective form factors (a 3.0) was calculated for 34 species according to Froese (2006): a 3:0 ¼ 10 log aÀSðbÀ3Þ ; where a and b are coefficients of LWRs and S is the regression slope of log a vs b. Following Froese (2006), the form factor from original data was calculated for 15 species with at least five LWRs with r 2 higher than 0.800 and with no outliers (Treer et al., 2008a). In addition, the form factor was calculated for 19 species that had length–weight relationships (LWRs) data in the FishBase (Froese and Pauly, 2008). Median a and b were used, as well as the mean slope for all fish species (S =)1.358). In the cases where median a was based on length other than the total length (Froese, 2006), the equation applied was: a TL ¼ a Ls ðTL/LsÞ Àb ; where a TL refers to total length, a Ls to the length type used in the FishBase, and TL ⁄ Ls is the average ratio of the two. Nine species were not included in the calculation. Six of these have no LWR data in the FishBase (Cobitis elongatoides, Aulopyge huegelli, Rutilus pigus, Chondrostoma nasus, Rhodeus sericeus, and Leuciscus souffia). The three other species had LWR data based on lengths other than total length (TL) and no length–length (LL) relationship to be transferred (Leuciscus svallize, Salmo obtusirostris, and Abramis sapa). Scientific names for each species were checked with the FishBase (Froese and Pauly, 2008). Results and discussion The values of mean condition factors, with their respective 95% confidence limits and form factors available for each species, are presented in Table 1 (form factors obtained from FishBase are marked with asterisks). The lowest mean condition factor (K = 0.5818) refers to the elongated C. elon-gatoides and the highest (K = 1.9440) to the deep-bodied Carassius carassius. These variations of K correspond with those obtained from 320 fish populations in Nigeria (Fafioye and Oluajo, 2005), which varied from 0.64 to 1.99. Although condition factors can vary considerably over different seasons, the geometric means for each species are relatively stable. These means of K depend mostly on the form of the fish, as shown in Fig. 1. The correlation between condition and form factors calculated from the original data of 15 populations (Fig. 1, solid circles) is highly significant (K = 0.1668 + 91.7354 a 3.0 ; r 2 = 0.966; P < 0.01; n = 15). The form factors calculated from the FishBase (Table 1, values indicated with asterisks) are not as accurate due to the following limitations: most (13 of 19 species) have fewer than five LWRs (many only 1 or 2); the outliers are not excluded; medians a and b are used instead of mean values; the mean slope for all fish species (S =)1.358) and not the specific slope for each species is used. However, in spite of these imperfections, when included (Fig. 1, open circles) the results showed significant correlation between condition and form factors (K = 0.6684 + 40.6545 a 3.0 ; r 2 = 0.505; P < 0.01; n = 34) and agreed entirely with the form factor concept of Froese (2006). The form factors multiplied by 100 (Table 1) from original data and most (10) of them from the FishBase, fall within the confidence limits of K; for three of them there were no CLs, as only one population gave K (Ctenopharingodon idella, Leucis-cus leuciscus, Oncorhynchus mykiss) and only seven form factors from FishBase fall outside the CLs. The reasons for this are: the small number of populations (<5) in calculating K (Gymnocephalus cernuus), in calculating the form factors (Ballerus ballerus, Scardinius erythrophthalmu, Tinca tinca), or in calculating both of them (Cottus gobio, Rutilus rubilio, C. carassius). Including more species in the calculation using the original data that satisfy recommendations by Froese (2006) (at least five LWRs with r 2 higher than 0.800, outliers excluded, form factors correlated to Ks obtained from at least five populations) would give even more accurate results.