We compared the leaf morphology and anatomy of the putative rheophytic ecotype of Viola mandshurica W. Becker var. ikedaeana (W. Becker ex Taken.) F. Maek. and its closely related variety, V. mandshurica var. mandshurica. We showed that the leaf of the rheophytic ecotype of V. mandshurica var. ikedaeana was narrower than that of V. mandshurica var. mandshurica. Moreover, the leaf thickness and guard cell size of the rheophytic ecotype of V. mandshurica var. ikedaeana were significantly larger than those of V. mandshurica var. mandshurica. We further showed that leaves of the rheophytic ecotype of V. mandshurica var. ikedaeana contained fewer cells than did those of V. mandshurica var. mandshurica. Our results suggest that the narrower leaves of V. mandshurica var. ikedaeana are caused by a decrease in the number of cells. A narrower leaf may enable the rheophytic ecotype of V. mandshurica var. ikedaeana to resist the strong flow of water that occurs after heavy rainfall, while a thicker leaf may enhance tolerance to desiccation and high- intensity light.
Irregular flooding can be an important stress factor for plants that do not inherently possess or are not able to develop traits enabling survival under submerged conditions. Spring and summer floods, which occur after strong rainfall [
Viola mandshurica W.Becker var. ikedaeana (W. Becker ex Taken.) F. Maek. is distributed in mountains on the Japanese mainland (western Honshu, Shikoku, and Kyushu), and also in Korea and Taiwan [19,20]. In addition, we newly found the variety on riversides (
All samples of V. mandshurica var. mandshurica and the
rheophytic ecotype of V. mandshurica var. ikedaeana examined in this study were collected from the field. We collected samples from 6 localities (2 localities for V. mandshurica var. mandshurica and 4 localities for the rheophytic ecotype of V. mandshurica var. ikedaeana). The collection localities are shown in
For morphological analysis, individuals were measured for the following continuous macromorphological leaf variables: 1) length and width of the leaf blade; 2) leaf thickness; and 3) angle of the leaf base. Measurements were made using a digimatic calliper (CD-15CXR; Mitutoyo, Kanagawa, Japan) and a digimatic outside micrometer (MDC-SB; Mitutoyo, Kanagawa, Japan). The leaf size was calculated by using the following formula: (leaf length × leaf width)/2. The leaf index was calculated as the ratio of the leaf length to the leaf width, according to [
For anatomical analysis, fully expanded leaves were collected from each individual. To count the number of cells on the blade, the adaxial surface of the leaves was peeled
Locality no. corresponds to that given in
off by using Suzuki’s Universal Micro-Printing (SUMP) method. The number of epidermal cells was calculated by using following formula: leaf size/cell size. Replicas of each leaf (1 cm2) were prepared to determine the stomatal density (number per mm2) and to measure the epidermal cell size of 10 cells per leaf. These copied SUMP images were examined once for each individual, with a light microscope (CX41; Olympus Co., Tokyo, Japan). Statistical analyses were performed using Tukey’s honestly significant difference (HSD) test and the Steel-Dwass test to compare the characteristics of V. mandshurica var. mandshurica and the rheophytic ecotype of V. mandshurica var. ikedaeana.
The leaf lengths of V. mandshurica var. mandshurica (2 populations—Asakura and Noichi) and V. mandshurica var. ikedaeana (4 populations—Towa, Nakahira, Nanato and Oboke) were 34.77 ± 0.97 mm, 59.79 ± 1.16 mm, 31.13 ± 1.99 mm, 31.65 ± 1.84 mm, 32.94 ± 1.62 mm, and 35.62 ± 1.68 mm, respectively; the leaf widths were 21.73 ± 2.14 mm, 29.93 ± 1.57 mm, 8.35 ± 0.37 mm, 10.77 ± 0.55 mm, 10.64 ± 0.31 mm, and 11.95 ± 0.51 mm, respectively; and the average angles of the leaf base were 156.49 ± 9.64 degrees, 180.94 ± 1.55 degrees, 72.97 ± 2.67 degrees, 125.72 ± 10.00 degrees, 103.91 ± 3.08 degrees, and 105.47 ± 3.85 degrees, respectively (
The leaf index of V. mandshurica var. mandshurica (1.60 ± 0.07 [Asakura], 2.04 ± 0.05 [Noichi]) was significantly lower than that of the rheophytic ecotype of V. mandshurica var. ikedaeana (3.70 ± 0.12 [Towa], 2.94 ± 0.08 [Nakahira], 2.34 ± 0.12 [Nanato], 3.27 ± 0.10 [Oboke]).
The cell sizes of V. mandshurica var. mandshurica (2 populations—Asakura and Noichi) and V. mandshurica var. ikedaeana (4 populations—Towa, Nakahira, Nanato and Oboke) were 6585.97 ± 362.66 µm2, 8241.96 ± 279.70 µm2, 5241.65 ± 117.68 µm2, 6346.67 ± 271.50 µm2, 6252.58 ± 145.25 µm2, and 6911.77 ± 650.03 µm2. The cell size did not differ significantly between V. mandshurica var. mandshurica and the rheophytic ecotype of V. mandshurica var. ikedaeana (
All measurements are represented as mean ± standard deviation. Columns marked by different letters showing significant differences according to Tukey’s HSD test (p < 0.05); 1Nonparametric pairwise comparison was conducted by using the Steel-Dwass test.
on the leaf size and cell size, we calculated the number of epidermal cells per leaf. The cell numbers for V. mandshurica var. mandshurica and the rheophytic ecotype of V. mandshurica var. ikedaeana were 73544.4 ± 7440.3, 84750.8 ± 6098.7, 32302.3 ± 4208.3, 30373.9 ± 2768.2, 39203.6 ± 2677.5, and 39797.0 ± 3319.9, respectively. Leaves of V. mandshurica var. mandshurica contained a significantly higher number of cells than did those of the rheophytic ecotype of V. mandshurica var. ikedaeana. The stomatal density of V. mandshurica var. mandshurica (122.88 ± 4.64 and 158.02 ± 4.42) did not differ significantly from that of the rheophytic ecotype of V. mandshurica var. ikedaeana (121.27 ± 3.74, 172.99 ± 10.41, 144.10 ± 6.03, and 173.02 ± 8.83). By contrast, the stomatal size (250.39 ± 4.02 µm2 [Towa], 281.95 ± 8.40 µm2 [Nakahira], 285.19 ± 6.56 µm2 [Nanato], and 266.67 ± 3.51 µm2 [Oboke]) of the rheophytic ecotype of V. mandshurica var. ikedaeana was significantly larger than that of V. var. mandshurica (122.45 ± 5.51 µm2 [Asakura] and 132.81 ± 4.91 µm2 [Noichi]).
The speciation of inland rheophyte species appears to be associated with stenophyllization, and rheophytes have been shown as scattered among various taxa, from bryophytes to angiosperms [
Our results indicate that a decrease in the number of cells in a leaf contributes to stenophyllization of the rheophytic ecotype of V. mandshurica var. ikedaeana. This anatomical mechanism of stenophyllization is similar to that previously reported for R. ripense [
Our morphological data for the rheophytic ecotype of V. mandshurica var. ikedaeana indicate that the angle of the leaf base is strongly correlated with the length and width of a leaf. Moreover, the decreasing angle of the leaf base results in a lanceolate leaf, indicating that stenophyllization of leaves of the rheophytic ecotype of V. mandshurica var. ikedaeana occurred with the transition from triangularand oblong-lanceolate leaves to hastate leaves. The general tendency for plants growing closer to streams to have lanceolate leaves may be caused by selection pressures of habitats along a gradually decreasing flooding frequency, from streambed to inland [
In the present study, we further demonstrated that the leaf thickness and stomatal size of the rheophytic ecotype of V. mandshurica var. ikedaeana were significantly larger than those of V. mandshurica var. mandshurica. Previous studies of F. japonicum var. luchuense and R. ripense showed that individuals growing on riverbanks had larger leaf thicknesses than did those of inland plants [9,14,23]. Taken together, these findings indicate that increased leaf thickness is a general tendency among rheophytes—V. mandshurica var. ikedaeana, R. ripense, and F. japonicum var. luchuense show a similar pattern of differentiation but are widely separated phylogenetically. Moreover, leaf thickness and stomatal size are positively correlated with the mean solar radiation during leaf expansion [25,26]. In general, stomata are the main gates for gaseous exchange of leaves [27-29]. The guard cells that surround the stomata contain chloroplasts, and are therefore able to increase their sugar concentration; this, in turn, causes water absorption and swelling of the cells [30,31]. In addition, stomatal conductance depends on leaf characteristics such as size, number, and stomatal density [32-34]. The rheophytic ecotype of V. mandshurica var. ikedaeana grows mainly in riverside habitats—on sunny, moist rocks and on riverbanks—in southern Shikoku, Japan; therefore, the condition of high irradiance along the riverside may lead to increased leaf thickness and stomatal size.
In summary, we have analysed the evolution of the rheophytic ecotype of V. mandshurica var. ikedaeana, by using morphological and anatomical data. The results provide an unbiased interpretation of the rheophytic evolution of angiosperms. Our data clearly indicate the rheophytic process of V. mandshurica var. ikedaeana, but do not provide a definitive interpretation of the rheophytic pattern of adaptation. In the Violaceae, for example, [
We thank Ogawa K, Matsuyama K, Yoshimi Y, Yokoyama N, Isomoto S, Miyata H, Tsuchiya Y, Muroi M, Kakimoto N, Kumekawa Y, Nishimura C, Uemoto C, and Inoue S for their help during our research. This study was partly supported by the River Fund in charge of the Foundation of River and Watershed Environment Management (FOREM), Japan and a Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan (to JY and TF).