All data were subjected to one-way anal-ysis of variance using the general linear models procedure of SAS software (SAS Institute, 1996). Statements of significancewere based on P < 0.05 unless otherwise stated, and significant differences between means were separated by the leastsignificant difference test.