The original E. purpurea plants were grown from seeds bought from the company of Plantation Products (Norton, MA, USA) [7]. Seeds were sterilized with 0.1% mercury bichloride for 30 s, 10% sodium hypochlorite for 60 s, and then washed with sterile water for three times [8]. The aseptic seeds were inoculated on Murashige and Skoog (MS) medium with half-strengthmineral salts, 0.01 mg・L⁻¹ Naphthalene Acetic Acid (NAA) and 30 g・L⁻¹ sucrose. The in vitro plants were propagated by inducing axillary buds [9] with medium consisting of full MS mineral salts, 0.5 mg・L⁻¹ BA, 0.01 mg・L⁻¹ NAA, and 30 g・L⁻¹ sucrose. After tests, the petioles cut from the cloned seedlings A9, A13, and A54, with different hyperhydricity ratios were selected for use in the present study.

・ Preparation of the Medium

Each glass jar was filled with 40 mL of medium and then covered with a polycarbonate screw cap. The medium for testing hyperhydricity contained Murashige and Skoog (MS) basal elements [10], 3% sucrose, 0.4 mg・L⁻¹ BA and 0.01 mg・L⁻¹ NAA [11]. The medium for inducing adventitious bud formation from explants contained MS basal elements, 3% sucrose, 0.2, and 0.5 mg・L⁻¹ BA, 0.01 mg・L⁻¹ NAA and 15 g∙L⁻¹ man- nitol, or 15 g∙L⁻¹ sorbitol. All the media used were adjusted to a pH value of 6.0, gelled with 0.45% agar prior to autoclaving at 1.4 kg・cm⁻² for 20 min.

・ Maintenance of the Cultures

The cultures were stored in a culture room at 25˚C - 27˚C. The cultures used for seed germination were first kept in darkness for 10 days, and all the other cultures were kept under a photoperiod of 12 h light (about 50 μmol・m⁻²・s⁻¹)/12h darkness.

・ Data Collection and Analysis

Data for adventitious bud formation were recorded after 40 days of culture. For the regeneration rate of adventitious buds, at least 1 cm in height, without hyperhydricity [12] and at least two visual leaves were considered as valid. For the height of buds, every bud was picked out from the bottle and the length from rhizome to leaf tip was measured. Statistical analysis of the data was carried out by using the SPSS 19.0 software. The results were expressed as mean ± SD (standard deviation). Data were analyzed by one-way ANOVA followed by Duncan’s multiple range tests or the independent sample t-test using Statistical Package for Social Sciences 19.0 (SPSS, Chicago, IL, USA). A p-value of less than 0.05 was considered statistically significant.

Three groups of petiole explants from different seeds were inoculated on the same medium consisting of full MS mineral salts, 0.5 mg∙L⁻¹ BA, 0.01 mg∙L⁻¹ NAA, and 30 g∙L⁻¹ sucrose. Results were showed in Table 1. Petiole explants from seed clone A9, A13, and A54 were tested. All three groups of explants share similar regeneration efficiency (3.29, 2.92, and 3.04 shoots per explant), but hyperhydricity ratio in clones A9 and A3 were higher.

Results showed in Figure 1 and Table 2 indicate that in MS medium with 0.2 mg∙L⁻¹ BA and 0.01 mg・L⁻¹ NAA, adding 15 g∙L⁻¹ mannitol increased the regeneration rate of explants of A54, with low hyperhydricity ratio from 1.77 to 2.37 shoots per explant. In MS medium with 0.2, and 0.5 mg∙L⁻¹ BA, and 0.01 mg・L⁻¹ NAA, adding 15 g∙L⁻¹ sorbitol decreased the regeneration rates for all the clones. But, supplementing 15 g∙L⁻¹ mannitol or 15 g∙L⁻¹ sorbitol to any of the media decreased the mean value and standard deviation of shoot height. In other words, both of them increased the uniformity.

Results showed in Figure 2 indicate that in MS medium with 0.3 mg∙L⁻¹ BA and 0.01 mg・L⁻¹ NAA, adding 15 g∙L⁻¹ mannitol increased the hyperhydricity rate, the regeneration rate, and the uniformity of height of shoots in clone A13, which was with higher hyperhydricity ratio. Meanwhile, it decreased the mean height of shoots.

Another group of explants with high hyperhydricity ratio were collected from seed clone A13. The explants were inoculated on MS medium with 0.2, and 0.5 mg∙L⁻¹ BA, 0.01 mg∙L⁻¹ NAA, and 15 g∙L⁻¹ mannitol or 15 g∙L⁻¹ sorbitol. The hyperhydricity ratios increased with the addition of mannitol or sorbitol. Some leave of shoots in medium

^*Values are expressed as mean ± SE. Data in the same column followed by different letters are significantly different by Duncan’s test at P < 0.05 level.

*Values are expressed as mean ± SE. Data in the same column followed by different letters are significantly different by Duncan’s test at P < 0.05 level.

with mannitol turned red, other leaves of shoots in medium with sorbitol were small, twist, and red (Figure 3).