taxonID	type	description	language	source
03F187DCFFCB5151FFF9F96A8023E9B9.taxon	description	3.1 | Geometric morphometrics The shape changes associated with the first three between-group principal components (Figure 1 (b )) cumulatively explained 76.9 % of variance in the data (Table 1). PC 1 described enlargement of head height at length, whereas PC 3 was most strongly associated with reduced head and body depth. PC 2 was associated with dorso-ventral arching of the body and, as an artefact of positioning the flexible fish bodies for photographing, was removed mathematically before further analysis (see § 2.4.1). While individual loadings on PC 1 described a progressive, age-related shape change from a larger to smaller head size (Figure 2 (a )), loadings on PC 3 appeared to describe an increase in body depth of exercised as opposed to control groups in the later stages of the experiment (Figure 2 (a )). This was partly confirmed by the results from linear discriminant analysis (LDA): the coefficients from linear discriminants 1 and 2, cumulatively explaining 88.33 % of variance in the data (Table 1) were most strongly associated with PCs 1 and 3, respectively (Table 2). LDA achieved 85.19 % correct group classification and 62.96 % cross-validated correct classification for treatment groups across weeks (Figure 2 (b) and Table 3). Between-group Mahalanobis distances and permutated T 2 tests (Table S 2 in File S 1) confirmed a clear group separation along the age axis (weeks under treatment), associated with a decrease in head height / depth relative to lateral body depth (LD 1, PC 1; Figure 2 (b )). Although body depth appeared to be greater in the exercised as opposed to control groups towards the end of the experiment (LD 2, PC 3; Figure 2 (b )), the difference between same-aged exercised and control groups was not statistically significant (p> 0.05; Table S 2 in File S 1). 3.2 | Mass – length relationships and condition at length A comparison of log mass-log length relationships for the exercised v. control cohorts did not reveal a significant difference in the slopes (F 1,43 = 0.21; p> 0.05) or intercepts dependent on treatment (F 1,43 = 0.65; p> 0.05). Cohort-specific regressions of M and L S with the C 00 individuals (n = 6) as a common origin described a mass – length relationship in the form of M ¼ 0: 0143 L 3: 0242 for the exercised s cohort and M ¼ 0: 0146 L 3: 0044 for the control cohort (Figure 3). Theres fore, exercised individuals were slightly (albeit not significantly) heavier for their length as they grew larger (stronger positive deviation of b from the cubic 3) than control individuals. Consequently, condition-at-length, K, derived from the regression parameters for each cohort, differed across weeks (i. e., age; F 3,39 = 194.038, d. f. = 3, p <0.001) as well as treatments (F 1,39 = 864.99, d. f. = 1, p <0.001) (Figure 4). In addition, the pattern of variation in K across weeks differed between treatments as indicated by the statistically significant interaction term (F 3,39 = 85.803, d. f. = 3, p <0.001). Specifically, K was higher in exercised compared with control groups over the whole experimental period. Within treatments, K consistently increased for exercised individuals over the experimental period, while for control individuals, K only increased from week 4 to week 20, week 4 to week 32 and week 10 to week 32 (Figure 4). 3.3 | Caudal-fin size A CF: L S 2 did not differ between treatments at 20 weeks (mean S. E.; exercised = 0.029 0.002, control = 0.028 0.002; t = 0.291, d. f. = 9.36, p> 0.05) or at 32 weeks (exercised = 0.024 0.001, control = 0.024 0.001; t = − 0.147, d. f. = 8, p> 0.05) after treatment initiation. However, overall and irrespective of treatment, A CF: L S 2 ratios were significantly higher after 32 weeks compared with 20 weeks (Figure 5) since treatment initiation (20 weeks = 0.028 0.001, 32 weeks = 0.024 0.001; t = 2.999, d. f. = 14.778, p <0.01).	en	Fenkes, Miriam, Shiels, Holly A., Nudds, Robert L. (2018): Body shape and robustness response to water flow during development of brown trout Salmo trutta parr. Journal of Fish Biology 93 (2): 360-369, DOI: 10.1111/jfb.13772, URL: http://dx.doi.org/10.1111/jfb.13772
