SEARCH

SEARCH BY CITATION

Appendix S1a. Means (standard deviations) of survey lengths and eleven habitat variables measured for each of the 54 study streams. For disturbance we give the proportion of streams with visible evidence of riparian disturbance (mostly cattle). For riparian it is the proportion of streams with closed/partial canopy riparian zones rather than pasture/grassland. The final three columns ‘Sub. PC1-3’ summarize variation in the scores along the first three axes of a principal components analysis conducted on the correlation matrix of the percent substrate in each of four categories: fines, gravel, cobble and boulder.

Appendix S1b. Bivariate Pearson correlation coefficients among the 11 continuous variables measured in the 54 study streams. Those significant at the uncorrected α=0.05 are in bold. The final three columns ‘Sub. PC1-3’ summarize variation in the scores along the first three axes of a principal components analysis conducted on the correlation matrix of the percent substrate in each of four categories: fines, gravel, cobble and boulder.

Table S2a. Analysis of Covariance on the abundance of post-yearling trout captured per square metre surveyed for 38 streams where rainbow and brown trout occurred in allopatry. The table shows the final model selected following the forward and backward stepwise (entry/removal P=0.05) assessment of 25 terms: three class (species (rainbow, brown), disturbance (cattle/erosion, none), riparian (forest, grassland)), ten continuous environmental variables (watershed area, watershed high point, channel width, channel gradient, valley width index (=valley width/channel width), pool percentage, total fish cover, PC1-3 scores summarizing variation in substrate), and all twelve first order interaction terms involving species. The final model includes terms that were retained by both forward and backward stepwise models.

Table S2b. Analysis of Covariance on the abundance of post-yearling trout captured per square metre surveyed for 38 streams where rainbow and brown trout occurred in allopatry. The table shows the final model selected following the forward and backward stepwise (entry/removal P=0.05) assessment of 25 terms: three class (species (rainbow, brown), disturbance (cattle/erosion, none), riparian (forest, grassland)), ten continuous environmental variables (watershed area, watershed high point, channel width, channel gradient, valley width index (=valley width/channel width), pool percentage, total fish cover, PC1-3 scores summarizing variation in substrate), and all twelve first order interaction terms involving species. The final model selected by including terms that were retained by both forward and backward stepwise models.

Table S2c. Analysis of Covariance on the total number of post-yearling trout captured per square metre surveyed for 38 streams where rainbow and brown trout occurred in allopatry. The table shows the final model selected following the forward/backward stepwise (entry/removal P=0.05) assessment of 13 terms: three class (species (rainbow, brown), disturbance (cattle/erosion, none), riparian (forest, grassland)), four continuous (scores of four PC axes summarizing variation in reach level habitat and substrate), and all six first order interaction terms involving species. The same model was selected using a backward stepwise procedure, so this is also the final overall model.

Table S3a. Analysis of Covariance on the total number of post-yearling peladilla captured per square metre surveyed for 32 streams north of latitude 42.8° S. The table shows the final model selected following forward and backward stepwise (entry/removal P=0.05) assessment of 12 terms: two class (disturbance (cattle/erosion, none), riparian (forest, grassland)), and ten continuous environmental variables (watershed area, watershed high point, channel width, channel gradient, valley width index (=valley width/channel width), pool percentage, total fish cover, PC1-3 scores summarizing variation in substrate). The table shows the final model selected by including terms that were retained by both stepwise models.

Table S3b. Analysis of Covariance on the total number of post-yearling peladilla captured per square metre surveyed for 32 streams north of latitude 42.8° S. The table shows the final model selected following forward and backward stepwise (entry/removal P=0.05) assessment of 9 terms: two class (disturbance (cattle/erosion, none), riparian (forest, grassland)), and seven continuous (scores of four PC axes summarizing variation in reach level habitat, PC1-3 scores summarizing variation in substrate). No terms were retained by either stepwise procedure. The table shows the model including all the terms available to enter.

Table S3c. Analysis of Covariance on the total number of post-yearling peladilla captured per square metre surveyed for 32 streams north of latitude 42.8° S. The table shows the final model selected following forward and backward stepwise (entry/removal P=0.05) assessment of 6 terms: two class (disturbance (cattle/erosion, none), riparian (forest, grassland)) and four continuous (scores of four PC axes summarizing variation in reach level habitat and substrate) variables. No terms were retained by either stepwise procedure. The table shows the model including all the terms available to enter.

Table S4a. Analysis of Covariance on the total number of post-yearling peladilla captured per square metre surveyed for 21 streams north of latitude 42.8° S with post-yearling peladilla and rainbow trout. The table shows the final model selected following the forward stepwise (entry/removal P=0.05) assessment of 12 terms: two class (disturbance (cattle/erosion, none), riparian (forest, grassland)), and ten continuous environmental variables (watershed area, watershed high point, channel width, channel gradient, valley width index (=valley width/channel width), pool percentage, total fish cover, PC1-3 scores summarizing variation in substrate).

Table S4b. Analysis of Covariance on the total number of post-yearling peladilla captured per square metre surveyed for 21 streams north of latitude 42.8° S with post-yearling peladilla and rainbow trout. The table shows the final model selected following the backward stepwise (entry/removal P=0.05) assessment of 12 terms: two class (disturbance (cattle/erosion, none), riparian (forest, grassland)), and ten continuous environmental variables (watershed area, watershed high point, channel width, channel gradient, valley width index (=valley width/channel width), pool percentage, total fish cover, PC1-3 scores summarizing variation in substrate).

Table S4c. Analysis of Covariance on the total number of post-yearling peladilla captured per square metre surveyed for 21 streams north of latitude 42.8° S with post-yearling peladilla and rainbow trout. The step-wise models retained no term in common.

Table S4d. Analysis of Covariance on the total number of post-yearling rainbow captured per square metre surveyed for 21 streams north of latitude 42.8° S with post-yearling peladilla and rainbow trout. The table shows the final model selected following the forward stepwise (entry/removal P=0.05) assessment of 12 terms: two class (disturbance (cattle/erosion, none), riparian (forest, grassland)), and ten continuous environmental variables (watershed area, watershed high point, channel width, channel gradient, valley width index (=valley width/channel width), pool percentage, total fish cover, PC1-3 scores summarizing variation in substrate).

Table S4e. Analysis of Covariance on the total number of post-yearling rainbow captured per square metre surveyed for 21 streams north of latitude 42.8° S with post-yearling peladilla and rainbow trout. The table shows the final model selected following the backward (entry/removal P=0.05) assessment of 12 terms: two class (disturbance (cattle/erosion, none), riparian (forest, grassland)), and ten continuous environmental variables (watershed area, watershed high point, channel width, channel gradient, valley width index (=valley width/channel width), pool percentage, total fish cover, PC1-3 scores summarizing variation in substrate).

Table S4f. Analysis of Covariance on the abundance of post-yearling peladilla captured per square metre surveyed for 21 streams north of latitude 42.8° S with post-yearling peladilla and rainbow trout. The step-wise models retained no term in common.

Table S4g. Analysis of Covariance on the total number of post-yearling peladilla captured per square metre surveyed for 21 streams north of latitude 42.8° S with post-yearling peladilla and rainbow trout. The table shows the final model selected following forward and backward stepwise models (entry/removal P=0.05) assessment of 9 terms: two class (disturbance (cattle/erosion, none), riparian (forest, grassland)) and seven continuous (scores of four PC axes summarizing variation in reach level habitat, PC1-3 scores summarizing variation in substrate) variables. The same model was selected by both stepwise procedures.

Table S4h. Analysis of Covariance on the total number of post-yearling rainbow captured per square metre surveyed for 21 streams north of latitude 42.8° S with post-yearling peladilla and rainbow trout. The table shows the final model selected following forward and backward stepwise models (entry/removal P=0.05) assessment of 9 terms: two class (disturbance (cattle/erosion, none), riparian (forest, grassland)) and seven continuous (scores of four PC axes summarizing variation in reach level habitat, PC1-3 scores summarizing variation in substrate) variables. The same model was selected by both stepwise procedures.

Table S4i. Analysis of Covariance on the total number of post-yearling peladilla captured per square metre surveyed for 21 streams north of latitude 42.8° S with post-yearling peladilla and rainbow trout. The table shows the final model selected following forward and backward stepwise models (entry/removal P=0.05) assessment of 7 terms: two class (disturbance (cattle/erosion, none), riparian (forest, grassland)) and scores of five PC axes summarizing variation in reach level habitat and substrate. The same model was selected by both stepwise procedures.

Table S4j. Analysis of Covariance on the total number of post-yearling rainbow captured per square metre surveyed for 21 streams north of latitude 42.8° S with post-yearling peladilla and rainbow trout. The table shows the final model selected following forward and backward stepwise models (entry/removal P=0.05) assessment of 7 terms: two class (disturbance (cattle/erosion, none), riparian (forest, grassland)) and scores of five PC axes summarizing variation in reach level habitat and substrate. The same model was selected by both stepwise procedures.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
ACV_354_sm_appendix-s1.doc86KSupporting info item
ACV_354_sm_appendix-s2.doc54KSupporting info item
ACV_354_sm_appendix-s3.doc59KSupporting info item
ACV_354_sm_appendix-s4.doc112KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.