Table 1.

Number of sample replicates

Data set	Samples	Tests	ML soft.	P-values			Min. conf. interval
				Average	Lowest	Highest	Lower	Upper
Buckley	100	100	RAxML	0.261	0.140	0.312	0.114	0.185
Buckley	500	100	RAxML	0.263	0.219	0.264	0.171	0.211
Sullivan	100	100	RAxML	0.411	0.290	0.510	0.256	0.327
Sullivan	500	100	RAxML	0.401	0.344	0.458	0.329	0.360
Dixon	100	100	RAxML	0.092	0.030*	0.160	0.017	0.051
Dixon	500	100	RAxML	0.095	0.065	0.132	0.056	0.073

Notes: 100 SOWH tests were performed for three data sets with a sample size of 100, and 100 tests were performed with a sample size of 500. P-values for the Dixon data set at a sample size of 100 vary from 0.030 to 0.169, indicating repeated SOWH tests at this sample size could result in different outcomes using a significance level of 0.05. The minimum confidence interval is 0.017–0.051, indicating that all p-values which fall below 0.05 are accompanied by a confidence interval which spans the significance level, therefore more sampling is required. At a sample size of 500, all p-values and all confidence intervals fall entirely above the confidence level (the minimum interval is 0.056–0.073), indicating a sufficient sample size. * indicates p-value less than 0.05.