## Table 1 |
||||

Summary of the statistical regression models used to investigate effects of species
richness and species composition on respiratory activity under (A) constant and (B)
varying temperature regimes |
||||

A Constant temperature regime |
||||

Independent variables |
Significant terms |
L-ratio |
d.f. |
p-value |

1. SR |
SR | 12.7 | 6 | 0.013 |

2. SC |
SC | 36 | 32 | <0.0001 |

B Varying temperature regimes |
||||

Independent variable |
Significant terms |
L-ratio |
d.f. |
p-value |

1. SR, A, F |
SR × A | 30.1 | 18 | <0.0001 |

A × F | 5.66 | 21 | 0.017 | |

Importance of independent variables: | ||||

SR | 94.4 | 14 | <0.0001 | |

F | 8.34 | 20 | 0.0154 | |

A | 37.8 | 16 | <0.0001 | |

2. SC, A, F |
SC × A × F | 132 | 156 | <0.0001 |

Importance of independent variables: | ||||

SC | 1485 | 66 | <0.0001 | |

F | 338 | 124 | <0.0001 | |

A | 540 | 124 | <0.0001 |

For each model, the highest order terms are presented and, where multiple terms are present, the relative importance of individual terms (and their interactions) within a model can be compared using the likelihood ratios (L-ratios). Note that separate models were run to test the effects of species richness (SR) and species composition (SC). A: amplitude of temperature change, F: frequency of temperature fluctuation. d.f.: degrees of freedom.

Langenheder * et al.*

Langenheder * et al.* *BMC Ecology* 2012 **12**:14 doi:10.1186/1472-6785-12-14