Appendix B. Results of repeated-measures ANOVA on the effects of birds and shade on the abundance of Katharina under experimental devices.
Source of variation
df SS MS F P Between Subjects (average effect over time) Treatment
2 6.63 3.31 4.85 0.04
Block 4 9.29 2.32 3.40 0.07 Error 8 5.47 0.68 contrasts Birds [-H +B -S (EC) vs. -H -B -S] 1 0.21 0.21 0.31 0.59 Shade [-H -B -S vs. -H -B +S] 1 3.86 3.86 5.34 0.045 Within Subjects (change in effect over time) Time 4 5.31 1.33 6.92 0.0004 Time × Treatment 8 12.4 1.56 8.11 < 0.0001 Time × Block 16 2.10 0.13 0.68 0.79
Error 32 6.14 0.19 contrasts Time × bird effect 4 0.84 0.21 1.09 0.38
Time × shade effect 4 9.00 2.25 11.73 < 0.0001 Notes: Treatments are coded by the presence or absence of Hedophyllum (+ or -H), Bird access (+ or -B) and Shade (+ or -S). Control treatments are coded by the presence or absence of devices (Exclusion Controls = EC). Abundances were calculated for five periods: Summer 1998 (average of May, June, July, August), November 1998, Summer 1999, November 1999 and Summer 2000. Data were square-root transformed prior to analysis. As Mauchly’s criterion for homogeneity of variance-covariance matrices was nearly met (P = 0.03), P values from univariate analyses are presented with Huynh-Feldt adjusted probabilities. P values which are relevant to tested hypotheses and significant at the alpha = 0.05 level are in boldface type.