2.6. Data analysis
The non-intercept sigmoid function as described in TableCurve® 2D (2002) was fitted on the cumulative germination/emergence percentage to determine the time to 50% germination/emergence (T50) (Jami Al-Ahmadi and Kafi, 2007): , where a is the maximum germination/emergence percentage, b is the turning point, c is slope of the line, x is the time (h) and y is the germination/emergence %. T50 germination/emergence was calculated and subjected to an appropriate analysis of variance (ANOVA) using SAS® statistical software version 9.2 (SAS, 1999).
The rate of germination/emergence was defined as the reciprocal of the time taken for half the population to germinate/emerge (1/T50). The optimum temperature (Topt) was determined by fitting a simple piece-wise linear model (broken-stick regression) using a non-linear (NLIN) procedure with SAS between temperature and rate of germination/emergence (1/T50) for each species separately. The rate of germination/emergence increased linearly with temperature from a minimum (Tmin) to a sharply defined optimum (Topt) beyond which the rate decreased linearly with temperature. The maximum temperature (Tmax) was taken as the highest temperature where 50% germination/emergence was reached.
Analysis of variance (ANOVA) was used to test for temperature and light treatment effects. Treatment means were separated using Fisher's protected least significant difference test at the 1% instead of the 5% level of significance (Snedecor and Cochran, 1980), because of heterogeneity of variances. For light experiment 2, treatment means were separated at the 5% level of significance. The data were analysed using the statistical program GenStat® (Payne et al., 2007).