Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://herba.msu.ru/shipunov/belomor/english/2008/nymph.pdf Äàòà èçìåíåíèÿ: Tue Sep 2 22:11:34 2008 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 05:55:59 2012 Êîäèðîâêà:

Nordic Journal of Botany 25: 329à338, 2007 doi: 10.1111/j.2007.0107-055X.00140.x, # The Authors. Journal compilation # Nordic Journal of Botany 2007 Subject Editor: Anne Krag Brysling. Accepted 25 March 2008

Morphological variation of Nymphaea (Nymphaeaceae) in European Russia
Polina A. Volkova and Alexey B. Shipunov
P. A. Volkova (avolkov@orc.ru), Biology Dept, Moscow State Univ., Vorobyevy Gory, RUà119899 Moscow, Russian Federation. à A. B. Shipunov, College of Natural Resources, PO box 441133, Univ. of Idaho, Moscow, ID 83844-1133, USA.

The morphological variability of the genus Nymphaea is described on the basis of detailed studies of fresh material from 49 populations in European Russia and southern Siberia. Macromorphological characters in the field (including analysis of leaf shape using geometrical morphometry), pollen size and exine sculpture have been analyzed. One polymorphic widespread species, N. candida, grows in most of European Russia. Nymphaea tetragona seems to be absent in the investigated waters and possibly in the whole central part of European Russia, whereas N. alba was found only in the Astrakhan Region (the delta of the River Volga). These three species are separated relatively well by several morphological characters in fresh plants. Nymphaea tetragona differs from N. alba and N. candida by the sculpture of the exine of the proximal part of the pollen grains, but the latter species do not seem to be differentiated by pollen characters. Size characters of Nymphaea leaves and flowers do not depend on organic content in water.

The genus Nymphaea (white water-lilies) is a taxonomically difficult group; different taxonomists distinguish from two (Uotila 2001) up to 12 species (Papchenkov 2003) of white water-lilies in the territory of European Russia. In addition, many species are believed to have numerous subspecies, chromosomal races, and forms of hybrid and artificial origin (Heslop-Harrison 1955, Uotila 2001, Papchenkov 2003). This ambiguous situation is caused by a high level of interspecific polymorphism (Heslop-Harrison 1955, Komarov 1970, Kupriyanova 1976) with poorly investigated nature, and intensive interspecific hybridization, as suggested by many authors (Heslop-Harrison 1955, Uotila 2001, Papchenkov 2003) but, however, still lacking strong evidence. Nymphaea species have high morphological plasticity. Size of leaves and flowers, and also some qualitative characters of flowers, are thought to be strongly dependent on hydrological (especially temperature) and edaphic conditions (Heslop-Harrison 1955, Kupriyanova 1976, Dubyna 1982). However, quantitative estimation of edaphic conditions was not performed in the above-mentioned studies. The size of different organs of Nymphaea species also depends on the age of the plant (Dubyna 1982). The most common opinion is that there are three Nymphaea species in Russia: N. alba L., N. candida Presl. and N. tetragona Georgi. Nymphaea alba occurs across all European Russia, N. candida grows both in European Russia and in Siberia, and N. tetragona occurs in Siberia, in the Russian far east and on the Kola Peninsula (Komarov 1970, Muntendam et al. 1996). Similar values of qualitative

diagnostic characters are found in all analyzed sources, so in theory, these characters should be sufficient to distinguish these three species (Table 1). Unfortunately, many important diagnostic characters (e.g. color and shape of the stigma, shape of cup base, etc.) change or disappear even after very careful pressing in herbaria (Lisitsyna 2003). Moreover, the differences in the leaf blade shape, which appear to be a frequently used diagnostic character (Komarov 1970, Lisitsyna et al. 1993, Uotila 2001), before and after herbarization are very similar to interspecific differences, i.e. leaves with leaf blade shape typical for N. alba obtain a shape more typical for N. tetragona after the herbarization (Volkova 2008). That is why it is so important to investigate the morphology of white waterlilies on fresh material (Uotila 2001). However, there is a belief that N. candida and N. alba can only be distinguished with certainty on the basis of size, shape and exine sculpture of pollen grains (Kupriyanova 1976, Muntendam et al. 1996, Uotila 2001). Kupriyanova (1976) showed that Nymphaea pollen grains in the European part of the USSR were characterized by high morphological stability and could be used for distinguishing species; this opinion is based mainly on the exine sculpture. Interspecific hybrids of Nymphaea are characterized by lower fertility (Heslop-Harrison 1955, Komarov 1970) and various morphological features of the pollen grains (Kupriyanova 1976). One should note that these investigations were conducted on small samples (1à3 flowers per species), while there is large variance of the palinomorphology even within one Nymphaea population (Volkova 2008), which lead us to 329


Table 1. Main diagnostic characters for distinguishing Nymphaea species in Russia (data from the literature). Character Shape of cup base Shape of filaments of inner stamens Sculpture of pollen grains exine Number of stigma lobes Shape of stigma disc Color of stigma disc Shape of the central stigma projection Flower diameter (cm) Ovary appearance Bud shape Leaf shape Shape of main leaf veins Length of the leaf (cm) Width of the leaf (cm)

N. alba
round linear baculums (7) 8à20 (23) almost flat (slightly concave) yellow short spherical (3) 5à15 (20) does not become narrower near stigma, covered up to the top with scars of fallen stamens oblongàovoid with obtuse top widely elliptic either roundedàovate or rounded almost straight (10) 15à30 (35) (8) 14à27 (35)

N. candida
roundedàquadrangular lanceolate verrucas 6à14 (20) strongly concave yellow, orange, red long conical (3) 5à11 (16) become narrower near stigma, is not covered up to the top with scars of fallen stamens oblongàovoid with acute top widely elliptic either roundedàovate or rounded bent along the full length (6) 12à26 (30) (8) 12à24 (30)

N. tetragona
quadrangular with prominent ribs oval granular (4) 5à10 (16) strongly concave yellow, red, purple long conical 3 (and less) à6 (10) become narrower near stigma, is not covered up to the top with scars of fallen stamens four-sided pyramid elliptic either roundedàovate or rounded bent only in the first third of the length (4) 5à9 (20) (3) 4à10 (16)

consider these studies as preliminar. Investigations by Muntendam et al. (1996) were carried out on SEMmicrographs. However, this method does not estimate the shape and size of intact pollen grains correctly because of their deformation in the high vacuum during scanning microscopy (Volkova 2008). In contrast to the previous investigations, Poddubnaya-Arnol'dy (1976) noticed that structure, size and shape of pollen grains can vary significantly within one species, although these are diagnostic characters. In the current study we aim to explore the morphological variation of the genus Nymphaea in European Russia and to clarify the taxonomic situation within this genus in the studied territory. To achieve our goals we used a complex approach: analyzing pollen size, exine sculpture and also macromorphological characters of fresh plants and their relation to organic matter content in the water on a large material.

population were deposited in the herbarium of Moscow State University (MW), Russia. We aimed to investigate no less than 15 plants per population. However, there were populations with fewer plants, so 344 plants were investigated in total. Six qualitative characters (shape of cup base, shape of filaments of inner stamens, color of stigmatic disc, shape of the central stigma projection, shape of main leaf veins, bud shape) and six quantitative (number of stigma lobes, diameter of the circle of outer stamens, length of the outer petal, length of the leaf, width of the leaf, position of the maximum width of the leaf) were observed on each plant (Table 1, Fig. 2). The amount of organic material in the water was estimated via the saprobity index, which was determined from a list of indicator species of diatoms and their abundance (Sladechek 1967). Analysis of the leaf shape The thin-plate spline (TPS) method of geometrical morphometry (Bookstein 1991, Adams et al. 2004, Shipunov and Bateman 2005) was used for investigations of the variability of leaf shape in Nymphaea. This method let us explore shapes directly, excluding the size factor, by the use of landmarks situated on the contours. In general, the contour of one fresh leaf (maximum leaf) from one plant per population was outlined. In addition, contours of all leaves from 11 plants from different populations were also obtained to investigate whether leaf shape depends on leaf position on the rhizome. Contours of N. candida and N. tetragona leaves from `Flora Nordica' (Uotila 2001) and from `Illustrated flora of northern US and Canada' (Britton and Brown 1913) were used for the reference as `anchors'. To digitize contours, we used 100 equally spaced landmarks (first landmark located at the base of the leaf). This algorithm of producing (pseudo)landmarks let us compare the shapes of objects without getting any information about the biological sense of the observed differences (Kores et al. 1993), a procedure that entirely fits our aim.

Material and methods
We treated geographically isolated groups of white waterlily plants as separate populations. Separate, well-delimited groups of leaves and flowers were treated as one plant (recognizing individual plants of water-lilies is often difficult due to the active branching of the underwater rhizomes and their frequent fragmentation). We investigated 44 populations of white water-lily in Karelia Republic, and the Moscow, Tver, Kaluga, Chelyabinsk, Lipetsk and Astrakhan regions of European Russia (Fig. 1). We also investigated three populations of N. tetragona on the shores of lake Bajkal (southern Siberia), situated close to the type location of this species (Krupkina 2001), and two populations of N. alba sensu lato from the same location for comparison with populations from European Russia. Data were collected from June to September in 2003à2005. Voucher specimens from each 330


Fig. 1. Collection sites for the investigated populations. Population numbers refer to Table 2. Investigated populations in southern Siberia (shore of the lake Bajkal) are not shown. 1 0Karelia republic (populations no. 20, 130à138, 206à211), 2 0Tver region (101à116), 3 0Moscow region (117, 118), 4 0 Moscow region (119), 5 0Kaluga region (128, 129), 6 0Lipetsk region (127), 7 0Chelyabinsk region (121), 8 0Astrakhan region (213à217).

The coordinates of the landmarks were written in the data file with the help of the screen digitizer tpsDig (Rohlf 2006). The coordinates of the consensus configuration, and also the values of the main, relative and partial warps, characterizing the degree of differences between the specimen and the consensus configuration were calculated with help of the program tpsRelw (Rohlf 2007), which realizes the idea of geometrical morphometry in prototype of principal component analysis. Original coordinates were normalized by the Procrustes fit method (a 00). We investigated the dependence of leaf shape on the leaf position on the rhizome with the help of tpsRegr (Rohlf 2005) software; the fit to the regression model was tested using a generalization of Goodall's (1991) f-test. Averaging of the blade shape of the leaves was done with the help of the program tpsSuper (Rohlf 2003). Data files were edited and converted with the auxiliary program tpsUtil (Rohlf 2000). Pollen morphology Our preliminary studies have shown that acetolysis treatment of pollen grains does not significantly change pollen

Fig. 2. Investigated quantitative characters of Nymphaea leaves and flowers. (A) leaf, AC 0width, DE 0length, BD 0position of the maximum width. (B) Flower, A 0diameter of the circle of outer stamens, B 0length of the outer petal.

size or the characters of exine sculpture (Volkova 2008). Thus, all the investigations of pollen with light microscopy were carried out on intact (unacetolysed) pollen from herbarized flowers. Pollen was measured using a MIKMED-1 light microscope (magnification 1.5 )15 ) 40) with ocular micrometer to within 1 mm. We measured no less than 10 pollen grains in polar view per flower (564 pollen grains were measured in total). Light microscopy was not sufficient to place most of the investigated samples into one of the traditionally distinguished exine-based pollen types (Kupriyanova 1976; Table 1). Therefore we continued by investigating the exine sculpture on the scanning electron microscope (Camscan S-2, accelerating voltage of 20 kV, hereafter SEM). Pollen of several specimens from herbarium (MW) from typical localities were also investigated under SEM for reference. Pollen grains were coated with gold-palladium arroy (approx. 25 nm thick). We made photos of the first 3à10 pollen grains that were seen in the appropriate position (polar view or equatorial view) using the software Scan Microcapture 2.20119, designed by A. V. Grigor'ev (421 pollen grains in total). 331


Pollen fertility was assessed on herbarized flowers by the acetocarmin staining method (Radford et al. 1974) using 100 pollen grains per flower. All unstained or faintly stained pollen grains were considered sterile (Table 2). Statistical data analysis Four morphotypes (hereafter `anchor specimens') which correspond to theoretical plants of N. alba s. str., N. candida, N. tetragona and N. )sundvikii (N. candida )N. tetragona), were `simulated', based on the common conceptions about morphology of these taxonomical units as described by Papchenkov (2003) (Table 1). We used multivariate analysis of variance, nonparametric correlation analysis, non-parametric Wilcoxon test and parametric Student test for detection of differences between and correlations among variables. The ShapiroàWilk's test was performed to test for normality. Kruskal non-metric multidimensional scaling (hereafter MDS, Ripley 1996) of similarity matrices computed with `daisy' metrics (Kaufman and Rousseeuw 1990) was used for the classification based on the morphology. This method was specially developed for data with mixed types of variables (both continuous and discrete), such as our data. Principal component analysis (hereafter PCA) was used for classification of the leaf shapes (because in this case we deal with continuous variables of relative warps matrix, Pavlinov 2001). Dichotomous recursive classifications (Ripley 1996) were used for calculations of the typical character values during the preparation of the diagnostic key. All calculations and graphs creation were made in the R environment for statistical computing (R Devel. Core Team 2005).

shape of the flower base, shape of filaments of inner stamens, number of stigma lobes, size of the leaf and flower size. This key is appropriate for distinguishing Nymphaea species in nature but not in the herbarium, which can be considered as a weakness of the key, caused, however, by the peculiarities of our object. Diagnostic key to N. alba, N. candida and N. teteragona (Fig. 4) 1. Cup base squared, with clear ribs (view peduncle). Filaments of inner stamens Plants small: length of outer petals B3 cm, leaf B9 cm .............................................. N. from the rounded. width of tetragona

- Cup base rounded or rounded-quadrangular, without clear ribs. Filaments of inner stamens oblong. Plants larger ........................................................................ 2 2. Cup base rounded (view from peduncle). Filaments of inner stamens linear. Leaves cover each other, raising above water. Number of stigma lobes 13. Length of the leaf 15 cm ............................................. N. alba - Cup base roundedàquadrangulate, clear edges never seen. Filaments of inner stamens linear or lanceolate. Leaves floating on the water surface. Number of stigma lobes B13, sometimes more but then the leaf B15 cm .............................................. N. candida

Macromorphology: shape of the leaf We can distinguish three poorly isolated plot zones as the result of the classification of leaves with PCA, based on their shape (relative warps matrix, GM data) (Fig. 5, 6). Zone A consists only of plants from the Astrakhan Region, whereas zone B consists of N. tetragona plants from the shore of Lake Bajkal. The Nymphaea tetragona `anchor' is situated in the central area of zone B. All other plants, including the N. candida `anchor', are situated between zones A and B. We found a significant dependence (generalized Goodall f-test: p B0.05) of leaf shape on the leaf position on the rhizome (distance from the apex) for 6 of 11 investigated plants. However, the nature of this dependence was not the same for different plants. Leaves of two plants became rounded basally with more divergent lobes while moving in the acropetal direction, whereas leaves of two other plants became elongated with less divergent lobes, and leaves of the other two plants did not change their shape. Dependence of sizes of leaves and flowers on the organic content in the water Differences in organic content between all investigated reservoirs and streams were not large. Values of saprobity index vary from 0.65 to 2.00, which corresponds to the second and the third class of water purity (Table 2). Most of the investigated reservoirs and streams are oligo- or betamesosaprobic (saprobity index 1.4à2.0) and all Karelian lakes with floating mats and one lake of this type in the

Results
Macromorphology: metric characters Classification of investigated plants by MDS based on their morphology put the N. candida `anchor' between the N. alba and N. tetragona `anchors', and N. )sundvikii between N. candida and N. tetragona. This result corresponds completely to the common conception of the morphology of these taxa (Uotila 2001, Papchenkov 2003) and serves as evidence of adequacy of our classification (Fig. 3a). One can see three fuzzy `clouds' of plants on the plot of the two first MDS dimensions. All plants from the Astrakhan region (populations 213à217) and plants from three seaside populations from the Karelia Republic (populations 130, 131 and 211, Table 2) are agglomerated around the N. alba `anchor'. These plants show typical N. alba macromorphology (Table 1) with the exception of the rounded-quadrangular shape of the flower base of the Karelian plants. All plants of N. tetragona from the shore of the Lake Bajkal (populations 201, 203 and 204) are concentrated around the N. tetragona `anchor'. All other investigated plants are concentrated around the N. candida `anchor' (Fig. 3a). Values of morphological characters typical for each of the groups distinguished were calculated and used for the diagnostic key to thew three European Russian Nymphaea species. We used such simple and reliable characters as 332


Table 2. Investigated populations of Nymphaea spp. Pop. number Name of stream or reservoir Geographic coordinates Saprobity index1 Proportion of fertile pollen grains (%)2 Class of water purity1 Type of the exine sculpture3

14. Karelia republic, Loukhi region 120 130 131 132 133 134 135 136 137 138 206 207 208 209 210 211 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 128 129 127 121 213 214 215 216 217 201 202 203 204 205
1 2 3 4

Lake Indigo Lake Tikhoe Lake Gamarbiya Peatbog near Lake Gremyakha Lake Sennoe Lake Evrika Lake Speloe Lake Kh Lake I Lake La Riv. Sinyaya Lake Verkhnyaya Pazhma, western part Lake Verkhnyaya Pazhma, eastern part Mouth of riv. Nol'ozyorskaya Head of riv. Nol'ozyorskaya Lake Tajnoe 2. Tver region, Udomlya environs Lake Lake Lake Lake Lake Lake Lake Lake Lake Lake Lake Lake Lake Lake Lake Lake 3à4. Pond Lake Pond Zaverkhov'e Matras Klin Borovno Belen'koe Golovets Perkhovo Moldino, eastern part Moldino, north-eastern part Moldino, central part Rogozno Turishino Soroka Volkovo Pisoshno Pochaevo Moscow region Sterlyazhij near Zvenigorod Sima near Zvenigorod near tourists base near Mozhajsk

66812.7? 66824.0? 66823.8? 66825.2? 66823.3? 66816.5? 66817.1? 66818.0? 66818.1? 66818.0? 66833.7? 66826.2? 66826.2? 66825.0? 66824.2? 66817.1?

N, N, N, N, N, N, N, N, N, N, N, N, N, N, N, N,

33815.9? 33831.2? 33830.5? 33829.1? 33816.6? 33829.5? 33827.4? 33802.7? 33802.6? 33802.9? 31815.0? 32820.0? 33820.2? 32827.0? 32829.0? 33812.4?

E E E E E E E E E E E E E E E E

0.68 0.8 0.75 ... 0.85 0.7 0.7 0.65 0.75 0.78 ... ... ... ... ... ... 1.7 1.4 1.24 1.4 1.4 1.87 1.4 1.6 1.6 ... ... ... 2 1.5 1.4 1.5 1.8 0.95 1.4 1.5 1.7 ... 1.9 ... ... ... ... ... E E E E E ... ... ... ... ...

... 80 ... 89 ... 100 93 38 98; 88 94 97 98; 93 82; 88; 99 99 ... 98 100 100 ... ... 100 95 ... ... ... 99 33 99 5 97 95 63 77; 87 ... 95 100 100 91 93 ... 98 95 97 98; 100 ... 80 ... 96 93

2 2 2 .. . 2 2 2 2 2 2 .. .. .. .. .. ..

. . . . . .

4 1 1 2 2 2 4 2 2 1 2 2 4 2 ... 1 4 2 .. . 2 4 3 2 1 4 2 4 2 4 4 4 2 4 2 2 and 3 1 1 2 2

57840.3?N, 57841.3?N, 57845.0?N, 57846.5?N, 57842.7?N, 57843.4?N, 57843.4?N, 57846.5?N, 57847.5?N, 57846.5?N, 57847.4?N, 57849.2?N, 57845.8?N, 57843.3?N, 57846.1?N, 57842.7?N,

34821.0?E 34820.3?E 34823.0?E 34825.0?E 35801.5?E 35805.0?E 35805.3?E 35814.5?E 35814.0?E 35814.0?E 35800.5?E 35806.3?E 34845.5?E 34844.7?E 34844.5?E 34847.0?E

3 2 2 2 2 3 2 3 3 .. . .. . .. . 3 3 2 3 3 2 2 3 3 .. . 3 .. .. .. .. .. .. .. .. .. .. . . . . . . . . . .

57842.3?N, 36841.6?E 57842.3?N, 36841.8?E 55840?N, 35855?E 54835.3?N, 35805.2?E 54835.0?N, 35805.2?E 52829.3?N, 39856.4?E 52845?N, 59830?E 46830?N, 46830?N, 46830?N, 46830?N, 46830?N, 51832.5? 51830.2? 51832.5? 51832.5? 51839.3? 48815?E 48815?E 48815?E 48815?E 48815?E 105806.3? 104808.5? 105819.2? 105808.5? 103842.2?

5. Kaluga region, Mosal'sk district Riv. Ressa, upper stream Riv. Ressa, lower stream 6. Lipetsk region Lake Mokhovoe 7. Chelyabinsk region, Bredy district Riv. Karaganka 8. Astrakhan region, delta of the Volga river Erik Finogenov Erik Volodarovskij Kultuk Pryamoj Lotosnyj, northern part Kultuk Pryamoj Lotosnyj, central part Krep' Blinovskaya 9. Southern Siberia, near the Lake Bajkal Lake near road Irkutsk*Ulan-Ude, 905 km Ponds of Bajkal pulp and paper plant Lake near road Irkutsk*Ulan-Ude, 202.5 km Peatbog Leshkovskoe Ponds near railway station Slyudyanka

... 1 1 1 1 3 2 3 3 2

N, N, N, N, N,

according to Sladechek 1967. data for several flowers from the same population are separated with semicolon. proximal part of the pollen grain was investigated, for exine types description see Results. Number of region corresponds to the Fig. 1.

333


Fig. 3. Multidimensional scaling of morphometric data for Nymphaea in European Russia. `Anchor specimens' are marked with black circles. (A) each plant is marked by the number of the region (Fig. 1, Table 2). (B) each plant is marked by the symbol, indicating the size of pollen grains: filled circles 0`small' pollen (maximum equatorial diameter is 32à40 mm, minimum 28à37 mm), filled triangles 0`large pollen' (44à52 mm, 38à50 mm, correspondingly), open circles 0unknown pollen size.

Moscow region are oligokseno- or xenobetasaprobic (saprobity index 0.65à1.00) according to Sladechek (1967). We did not reveal any significant correlations between the size characteristics of the plants and the saprobity index or any significant differences in sizes of plants. Pollen morphology Values of maximum and minimum equatorial diameters have a bimodal distribution. Almost all investigated populations clearly belong either to the group with `small' (maximum equatorial diameter is 32à40 mm, minimum 334

Fig. 4. Qualitative characters, Nymphaea species in European base (view from peduncle, scale (B) N. candida, (C) N. tetragona inner stamen (scale bar length candida, (F) N. tetragona.

used for separation of three Russia. (A)à(C) shape of flowers bar length is 2 cm). (A) N. alba, . (D)à(F) shape of the filament of is 5 mm), (D) N. alba, (E) N.

28à37 mm) or `large' (44à52 mm, 38à50 mm, correspondingly) pollen grains. Values of maximum to minimum equatorial diameter ratio have a unimodal distribution with median 1.08 and quartile range 1.04à1.10. Plants of N. alba and N. tetragona morphotypes have `small' pollen. Plants of the N. candida morphotype have either `small' or `large' pollen (Fig. 3b); it is not possible to distinguish separate morphotypes in N. candida according to the pollen size.


Fig. 5. Principal component analysis of relative warps matrix (geometrical morphometry data) for leaf contours of Nymphaea in European Russia. Each plant is marked by the number of the region (Fig. 1, Table 2). `Anchor specimens' are marked with black circles. Plants from Astrakhan region (`A zone') and from Siberia (`B zone') are separated.

Pollen fertility in different plants from one population did not vary more than 15%. Plants from a vast majority of the populations have highly fertile pollen (more than 75% of fertile pollen grains), whereas one population (116) from the Tver region has deferred pollen fertility (63%), two populations, from Karelia (136) and from the Tver region (111), have low pollen fertility (33à38%) and one population from the Tver region (113) has almost sterile pollen (5% of fertile pollen grains) (Table 2). This last population produced very few (10à20 pollen grains per anther) and deformed pollen. The macromorphology of plants from this population was also quite unique and combines characters of N. alba (linear filaments of inner stamens, yellow and flat stigma) and N. candida (conic central projection of stigma). Visual analysis of SEM-micrographs revealed that exine sculpture is not homogenous on the whole surface of the pollen grain. Sculpture of the distal part of pollen grains does not demonstrate any discrete patterns, consisting of verrucas of different size that become larger near the equator. The exine sculpture of the proximal part of the pollen grains is very diverse. Hereafter we will describe the exine sculpture of the proximal part of the pollen grain as `exine sculpture' because this part is the most informative for distinguishing different pollen morphotypes in Nymphaea. Four main types of exine sculpture can be distinguished with some transitions between them: 1. sparse (7à9 sculpture elements per 100 mm2 of pollen grain surface) dense groups and single verrucas combined with single bacula of 2à4 mm length (Fig. 7a); 2. quite dense (22à38 sculpture elements per 100 mm2 of pollen grain surface) verrucas and bacula of 1à5 mm length (Fig. 7b);

Fig. 6. Averaged contours of leafs for three groups of leaves with different shape (the size was not taken into account). (A) plants from Astrakhan region (Fig. 5, Zone A), (B) group of N. candida `anchor' (Fig. 5, central region), (C) plants from Siberia (Fig. 5, `B zone').

3. very dense, frequently merging, different sized verrucas varying from almost flat to evidently prominent (Fig. 7c); 335


Fig. 7. Scanning electron micrographs of Nymphaea pollen grains, view from the proximal pole. Different types of the exine are shown. Population numbers refer to Table 2. (A) exine sculpture of the first type (population number 128), (B) exine sculpture of the second type, variant with long baculums (population number 137), (C) exine sculpture of the third type, variant with prominent verrucas (population number 106), (D) exine sculpture of the fourth type (population number 105).

4. combined type, characterized by different combinations of verrucas and bacula of various density (13à33 sculpture elements per 100 mm2 of pollen grain surface) (Fig. 7d). Exine sculpture frequently varies considerably within one population. The population from one pond in the Moscow region (population 119) appear especially interesting in this respect. Near the southern shore of the pond there are plants with N. candida morphology (Table 1) and exine sculpture of the second type. In the middle of the pond there are plants with some morphological characters of N. tetragona (bent along the full length main veins of the leaf, strongly invaginated stigma, Table 1) and exine sculpture of the third type. Three investigated populations in the large Lake Moldino in Tver region (108à110) have various exine sculpture (Table 2) and very similar macromorphology (Fig. 3). We found the third type of exine sculpture in all herbarium specimens of N. tetragona and in all populations of N. tetragona collected by us near Lake Bajkal (201, 203 and 204). This type of exine sculpture was also found in one lake in the Tver region (population 106, Table 2). Plants in this population differed from the N. candida 336

morphotype by linear filaments of inner stamens, main veins which are bent along the full length of the leaf and strongly invaginated stigma (Table 1). All other populations have pollen of the first, second, and forth type, irrespectively of plant macromorphology or geography (Table 2).

Discussion
Dependence of Nymphaea size characters on the organic content in the water Contrary to the widespread opinion (Heslop-Harrison 1955, Komarov 1970, Kupriyanova 1976, Dubyna 1982), there was no significant dependence of Nymphaea size characters on content of organic matter in the water. On the one hand, we can suppose that the range of organic content in the investigated reservoirs was not large enough for detecting such a dependence. On the other hand, special quantitative investigations of organic content were not carried out. Small northern lakes with floating mats are traditionally classified as oligotrophic contrary to mesotrophic lakes of the middle Russia: central part of European Russia (Kupriyanova 1976). Our quantitative investigations


do not demonstrate any considerable differences in organic content for the investigated lakes. Consequently, the size characters of Nymphaea are not so clearly caused by organic content in the water as was thought before. Some notes on Nymphaea taxonomy in European Russia Our data allowed us to distinguish three Nymphaea morphotypes in European Russia and southern Siberia, which correspond to the literature descriptions of N. alba, N. candida and N. tetragona. These species can be separated by several macromorphological characters of living plants, as indicated in the diagnostic key given above. These species differ also by the shape of fresh leaves, but these differences are not clear so we do not recommend the use of leaf shape as a diagnostic character. Our investigations did not reveal any definite dependence of leaf shape on leaf position on the rhizome. Therefore, one can analyze leaf shape without taking into account the leaf position. Pollen morphology, in our opinion, has much less diagnostic value than was thought before (Komarov 1970, Kupriyanova 1976, Dubyna 1982, Uotila 2001). Sizes of pollen grains for N. alba, N. candida and N. tetragona overlap considerably; moreover, this character show large variation within populations. Only N. tetragona can be distinguished on the base of exine sculpture while N. alba and N. candida do not differ on this character, contrary to the findings of researchers who worked with small samples (Kupriyanova 1976, Muntendam et al. 1996). We found plants with typical N. alba morphology only in the most southern part of European Russia (the delta of the Volga River, Astrakhan region), plants with typical N. tetragona macromorphology were found only in southern Siberia. Our investigations do not support reports about N. alba and N. tetragona in middle Russia (Lisitsyna et al. 1993). According to our data, the variation of Nymphaea morphotypes in middle Russia can not be dissected into discrete groups. Separate marginal morphotypes of the morphological continuum, existing in middle Russia, can be interpreted as N. alba and N. tetragona. Specimens that have intermediate morphology between an imaginary center of continuum (typical N. candida) and its marginal morphotypes can also be interpreted as hybrids N. alba )N. candida 0N. ) borealis and N. tetragona )N. candida 0N. )sundvikii. This approach was driven to its logical end by Papchenkov (2003). However, this interpretation does not correspond with generally accepted species definitions (Grant 1981). Distinguishing species on the basis of small differences in morphology is not appropriate, in particular not for plants with a prevalence of vegetative reproduction over sexual (Elven et al. 2004), such us water-lilies (Heslop-Harrison 1955, Dubyna 1982). We think that in middle Russia only N. candida exists, a very polymorphic species. Our data about high morphological variability of N. candida agree with the data of Aleksandrova et al. (1996) for Lipetsk region of European Russia and contradict the data of Dubyna (1982) for the Ukraine.

Our point of view is supported by reports on the existence of hybrids in the absence of parental species (Uotila 2001, Papchenkov 2003). As far as we know, these hybrids were distinguished only on the basis of morphology so plants with unusual combinations of morphological characters lacking hybrid origin can easily be interpreted as `hybrids'. Our studies of pollen fertility show that plants with unusual combinations of morphological characters almost always have highly fertile pollen and can therefore hardly be considered as hybrids (Heslop-Harrison 1955, Komarov 1970). Division of the investigated populations of N. candida on the basis of the pollen size corresponds to their division on the basis of the macromorphological character set which probably means that two chromosomal races of N. candida exist in middle Russia (Uotila 2001). This assumption is in agreement with data on the correlation between pollen size and ploidy level (Poddubnaya-Arnol'di 1976) and on various chromosomal numbers known for European N. candida (Heslop-Harrison 1955, Dubyna 1982, Krupkina 2001). To test this hypothesis in the future, it is essential to estimate the ploidy level. However, these estimations will be troubled by high chromosomal numbers (up to more then 100) and small sizes of the chromosomes (Heslop-Harrison 1955). It is also essential to carry out DNA analysis for additional evidence about species diversity in the genus Nymphaea in European Russia (Muntendam et al. 1996), which, as far as we know, has not been done for the European Nymphaea species on a sufficient material. Investigations of Nymphaea phylogeny based on the chloroplast trnTàtrnF region had not enough resolution to solve the relationships in the N. alba, N. candida and N. tetragona group (Borsch et al. 2007). More sensitive methods of indirect DNA analysis i.e. AFLP fingerprinting has been used successfully to show hybridization between N. alba and N. candida but only in few localities in Germany and Sweden (Werner and Hellwig 2006), while N. tetragona remains unstudied. Our study let us make the following conclusions: one polymorphic widespread species, N. candida, grows in most of European Russia. Nymphaea tetragona seems to be absent in the investigated waters and possibly in the whole central part of European Russia, whereas N. alba was found only in the Astrakhan Region (the delta of the River Volga). These three species are separated relatively well by several morphological characters in fresh plants. Nymphaea tetragona differs from N. alba and N. candida by the sculpture of the exine of the proximal part of the pollen grains, but the latter species do not seem to be differentiated by pollen characters. Size characters of Nymphaea leaves and flowers do not depend on the organic content of the water.
Acknowledgements à We are grateful to everybody who helped us in field collection of the material in various ways: all members of summer practices of Moscow South-West High School (1543), led by Dr S. M. Glagolev; E. Chibelyov (natural-historical reserve `Arkaim'), Dr K. I. Aleksandrova (Lipetsk Univ.), Dr V. I. Sutula (Bajkal Zapovednik) and Dr A. K. Gorbunov (Astrakhan Zapovednik). We are also grateful to Dr S. V. Polevova (Moscow State Univ.) for consultations on palynology, Dr O. V. Anisimova (Moscow State Univ.) for estimation of the organic content in the water, Dr I. Ya. Pavlinov (Moscow Zool. Museum) for consultations on geometric morphometry, Dr D. D. Sokoloff (Moscow

337


State Univ.) and Dr V. G. Papchenkov (Inst. for Biology of Inland Waters) for valuable comments on the paper, J. A. Dragon (Middlebury College, USA) for English improvement and Dr S. R. Mayorov (Moscow State Univ.) for all his support.

References
Adams, D. C. et al. 2004. Geometric morphometrics: ten years of progress following the `revolution'. [Online] Bhttp://life.bio. sunysb.edu/morph/review/review.html . Aleksandrova, K. I. et al. [ ] (eds) 1996. Flora of Lipetsk region. à Argus, Moscow, in Russian. Bookstein, F. L. 1991. Morphometric tools for landmarks data. à Cambridge Univ. Press. Britton, N. and Brown, A. 1913. An illustrated flora of northern US, Canada and British possessions (2nd ed.). à C. Scribners son. Borsch, T. et al. 2007. Phylogeny of Nymphaea (Nymphaeaceae): evidence from substitutions and microstructural changes in the chloroplast trnTàtrnF region à Int. J. Plant Sci. 168: 639à671. Dubyna, D. V. [ ] 1982. Nymphaeaceae of the Ukraine. à Naukova Dumka, Kiev, in Russian. Elven, R. et al. 2004. An operational species concept for the panarctic flora [Online] Bhttp://binran.spb.ru/projects/paf/ papers/sp_concept/sp_concept.htm . Goodall, C. 1991. Procrustes methods in the statistical analysis of shape à J. R. Stat. Soc. 53: 285à339. Grant, V. 1981. Plant speciation. à Columbia Univ. Press. Heslop-Harrisson, J. 1955. Nymphaea à J. Ecol. 43: 719à734. Kaufman, L.and Rousseeuw, P. J. 1990. Finding groups in data: an introduction to cluster analysis. à Wiley. Komarov, V. L. [ ] 1970. Nymphaea L. à In: Komarov, V. L. (ed.), Flora of the USSR. Vol. 7. Israel Program for Scientific Translators, pp. 8à12. Kores, P. J. et al. 1993. Morphomeric variation in three species of Cyrtostylis (Orchidaceae) à Syst. Bot. 18: 274à282. Krupkina, L. I. [ ] 2001. Nymphaeceae Salisb. à In: Flora of the Eastern Europe. Vol. 10: 25à30, Nauka, Saint Petersburg, in Russian. Kupriyanova, L. A. [ ] 1976. Pollen morphology of Nymphaea species in the European part of the USSR. à Bot. Zhurn. 61: 1558à1563, in Russian. Lisitsyna, L. I. [ ] 2003. Herbarization of aquatic plants, arrangement of the collections. à In: Papchenkov, V. G. et al. (eds), Aquatic botany: methodology, methods: materials of the school of aquatic botany. OAO `Rybinskij dom pechati', Rybinsk, pp. 49à56, in Russian.

Lisitsyna, L. I. et al. [ ] 1993. Nymphaea L. à In: Flora of the reservoirs of Volga basin. à Gidrometeoizdat, Saint-Petersburg, pp. 136à137, in Russian. Muntendam, J. B. et al. 1996. Morphometric patterns in the Nymphaea albaàcandida complex à Acta Bot. Neerl. 45: 279à 302. Papchenkov, V. G. [ ] 2003. On the determination of difficult groups of aquatic plants and their hybrids. à In: Papchenkov et al. (eds), Aquatic botany: methodology, methods: materials of the school of aquatic botany. OAO `Rybinskij dom pechati, Rybinsk, pp. 82à91, in Russian. Pavlinov, I. Y. 2001. Geometric morphometrics, a new analytical approach to comparison of digitized images. à In: Information technology in biodiversity research. Abstracts of the 2nd Int. Symp. Russian Acad. Sci., St. Petersburg, pp. 44à64. Poddubnaya-Arnol'di, V. A. [ ] 1976. Cytoembryology of the angiosperms. à Nauka, Moscow, in Russian. Radford, A. E. et al. 1974. Vascular plant systematics à Harper and Row. R Development Core Team. 2005. R: a language and environment for statistical computing. à R Found. Stat. Comp., Vienna. Ripley, B. D. 1996. Pattern recognition and neural networks. à Cambridge Univ. Press. Rohlf, F. J. 2000. tpsUtil*TPS utility program [Online] Bhttp://life.bio.sunysb.edu/morph/ . Rohlf, F. J. 2003. tpsSuper. Ver. 1.12 [Online] Bhttp://life.bio. sunysb.edu/morph/ . Rohlf, F. J. 2005. tpsRegr. Ver. 1.31 [Online] Bhttp://life.bio. sunysb.edu/morph/ . Rohlf, F. J. 2006: tpsDig. Ver. 2.10 [Online] Bhttp://life.bio. sunysb.edu/morph/ . Rohlf, F. J. 2007. tpsRelw: relative warps. Ver. 1.45 [Online] Bhttp://life.bio.sunysb.edu/morph/ . Shipunov, A. B. and Bateman R. M. 2005. Geometric morphometrics as a tool for understanding Dactylorhiza (Orchidaceae) diversity in European Russia à Biol. J. Linn. Soc. 85: 1à12. Sladechek, V. 1967. System of water quality from the biological point of view. à Arch. Hydrobiol. 7:1à218. Uotila, P. 2001. Nymphaea L. à In: Jonsell, B. (ed.), Flora Nordica. Vol. 2. Bergius Foundation, Stockholm, pp. 216à 221. Volkova, P. A. [ ] 2008. Changes of size and shape of pollen grains and leaves after traditional methods of treatment of fresh plant material with Nymphaea species (Nymphaeaceae) as an example. à Bot. Zhurn. 93:145à152, in Russian. Werner, K and Hellwig, F. 2006. Hybridization between Nymphaea alba and Nymphaea candida investigated by AFLP fingerprinting and morphological data. à In: Abstracts of 17th Symp. Biodiv. Evol. Biol., Bonn, p. 227.

338