为了研究水田退耕初期种子库的特征，在两个不同水位条件下通过幼苗萌发法对洪湖的濒湖退耕田种子库的结构和季节动态进行了研究。结果表明，种子库季节变化明显，种子库的密度于12月达到最大值2.61×105 ind.·m-2，6月时最小，为2.90×104ind.·m-2。而萌发的物种数则相反，6月时最大，12月时最小。Simpson 和Shannon-Wiener多样性指数及Shannon均匀度指数表现出较为一致的季节变化规律，在6月最高，12月和3月最低。该种子库类型属于持续性种子库。种子库与地表植被群落种相似性系数变化较大。0～5 cm和 5～10 cm土层中的种子库萌发的物种数差异不显著，但在数量上有明显分层现象，上层种子库占总数的78%。One-way ANOVA分析表明，萌发时的水位条件可显著影响种子库萌发的种类和数量。结合种子库研究的结果，讨论了种子库在退田还湖后湿地植被恢复过程中的重要性。
Germination tests of two water regimes were used to evaluate traits, mainly
community structure, of the soil seed bank of Honghu wetland withdrawn from
long-term rice culture.
One-way ANOVA was employed by means of diversity, richness and evenness
indices, and seed density. To evaluate the role of seed bank in restoration of wetland
communities, a soil core was sampled in June 2000, December 2000, March 2001 and
May 2001. Furthermore, the delamination of the seed bank was conducted by comparing
seed densities at different depths in the soil in the March 2001 sample. The results
showed that: 1) seed density in different seasons varied a lot, from 2.90×10 4
ind.·m-2 in June 2000 to 2.61×10 5 ind.·m-2 in December 2000. The size of the seed
bank in Honghu wetland is the largest among all the types of seed banks. The seed
bank belonged to the IV type of the Thompson and Grime model of soil seed banks of
different habitats. 2) For the community of the soil seed bank of Honghu wetland,
Simpson and Shannon-Wiener diversity indices, and Shannon evenness index were
maximal in June 2000, and minimal in December 2000. 3) The number of species had
no significant difference between soil depths of 0-5 cm and 5-10 cm. The seed density
in the upper layer accounted for 78% of the total seed number of the two layers,
which indicated that it was creditable to study only the upper level (0-5 cm) in
this kind of research. 4) Seed germination was affected by water depths in both number
of species and number of individuals of seedlings. 5) Degrees of similarity between
species composition of soil seed banks and that of the wetland vegetation during
December, March and May were 33.3%, 20.8% and 15.4%, respectively. They were
considerably high compared to those of some other types of habitats. 6) Seeds of
typical wetland species were scarce in all sampling periods, suggesting that in order
to restore the vegetation of the disturbed wetland more quickly, external seed banks
with abundant wetland species need to be introduced rather than only depending on
the original ones.