Fig. 3. Graphical representations of the multiple regression analyses we performed on the morphological, gait characteristics and performance data in the 11 lacertid lizard species. (A) All potential relationships. Each arrow represents one analysis. Independent variables introduced into the same multiple regression are boxed. Arrows are drawn from the independent variables box to the dependent variable box. We performed two series of analyses: one that included the (contrasts of the) values for the four gait characteristics at maximal sprint speed, and the second on the (contrasts of the) slopes of each of the four gait characteristics against speed. (B) Significant relationships in the multiple regression analyses including the contrasts of the estimated values of the four gait characteristics at maximal sprint speed. (C) Significant relationships in the multiple regression analyses including the contrasts of the slopes of each gait characteristic against speed. By definition, the contrasts of the slopes of stride length and stride frequency against speed are inversely related so we never introduced them simultaneously into the analyses. Instead, we included only the contrasts of the slopes of stride frequency against speed in the analysis; including the contrasts of the slopes of stride length against speed gives the opposite result (r=0.56; P=0.075).