Figure S1 Relationships between cross-sectional area and watershed area (derived from a 30-m digital elevation model in GIS) for spring-fed and surface runoff-fed streams used in the analysis of ecosystem size effects on food chain length. While there is a strong relationship between drainage area and cross-sectional area for surface runoff-fed streams (R2 = 0.96, P < 0.001), there is no relationship for spring-fed streams (R2 = 0.10, P = 0.678). Consequently, the association between the two ecosystem size descriptors is weak in the pooled dataset (R2 = 0.09, P = 0.247).

Figure S2 Scatter plots of mean predator (grey circles = fish, white circles = predatory invertebrates) d13C and d15N signatures as a function of mean Deleatidium values. The left figure illustrates that Deleatidium characterises well basal carbon in the 16 streams (linear regression: R2 = 0.70, P < 0.001). The right figure illustrates that mean predator d15N increases steadily with mean Deleatidium d15N values, and that predatory invertebrates and fishes differ from primary consumers in a different manner (linear regression for fishes: R2 = 0.55, P = 0.002).

Table S1 Criteria scored in each stream under the Pfankuch (1975) river disturbance assessment method, as applied in New Zealand rivers (Collier 1992).

Table S2 Factor loadings from a principal components analysis (PCA) of disturbance and environmental stability variables [i.e., stage height and temperature coefficients of variation, CV(H) and CV(T), and Pfankuch disturbance scores]. The first principal component (PC1) was taken as our multivariate index of disturbance (after Death and Winterbourn 1994), which in this instance accounted for 67% of the variance in the three variables. To shift PC1 into a positive domain, we added 10 to the values for each site prior to using them in tests of FCL controls and proximate structural mechanisms.

Table S3 Summary statistics for cross-sectional area measurements taken in study streams between September 2004 and August 2006.

Table S4 Occurrence of fish species in sampled streams. A bold-faced capital ‘X’ denotes the species holding the highest trophic position at that site. Sites 8 and 15 did not contain fish. Site numbers correspond (sequentially) to sites named in Table 1.

Table S5 Occurrence of predatory invertebrate taxa at sampled sites. Note that MTP was estimated for predatory invertebrates at sites 8 (Fan, top = Hydrobiosis sp.) and 15 (Tussock, top = Stenoperla sp.) only. Site numbers correspond to sites named in Table 1.

Table S6 Model-selection results for distinguishing between ecosystem size and disturbance as drivers of proximate structuring mechanisms for food chain length variation. K is the number of estimated parameters (including the residual error term), n is the number of data points, R2 is the coefficient of determination, AICc is Akaike’s Information Criterion (with small sample size correction), and Di is the AICc difference between a given model and that with the lowest AICc value. Evidence ratios (i.e., wtop/wi) are relative to the top model in each set (i.e., that with Di = 0.0).

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.

ELE_1484_sm_fS1-S2-tS1-S6.doc191KSupporting 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.