POLLEN MORPHOLOGY OF AGROPYRON GAERTNER IN TURKEY

Pollen morphology of Agropyron cristatum (L.) Gaertner. s.s. (subsp. incanum (Nábĕlek) Melderis and subsp. pectinatum (M. Bieb.) Tzvelev, latter including var. pectinatum and var. imbricatum (Roemer & Schultes) G. Beck) in Turkey has been studied by using light microscope and scanning electron microscope. The above-mentioned taxa are homogenous in both aperture type and exine ornamentation. Pollen grains are monoporate (rarely diporate in the case of var. imbricatum) having scabrate grouped exine surface. The scabra density and the height of scabrae as well as other morphological parameters such as annulus and operculum diameter are peculiar features for differentiation of taxa. Two different phenograms were created with the UPGMA (Unweighted Pair Group Method with Arithmetic mean) clustering technique using quantitative measurements of the pollen grains. Introduction The taxonomy of Agropyron Gaertner (Poaceae) was studied in different ways by various researchers. In a broad sense, it was once thought to be one of the largest genera encompassing more than 100 species in the tribe Triticeae Dumort (Dewey, 1983). Nevski (1934) treated Agropyron as a small genus consisting only of the species with keeled glumes. The remaining taxa were placed in Elytrigia Desv, Roegneria C. Koch and Elymus L. Agropyon has been restricted to the species with P genome composed of three ploidy levels (2x = 14, 4x = 28, 6x = 42) (Dewey and Asay, 1975; Melderis, 1978; Dewey, 1983; Assadi, 1995; Jensen et al., 2006). This narrow concept of Agropyron has been accepted by many authorities. The major Eurasian floras have followed the generic concept of Nevski (Tzvelev, 1976; Melderis et al., 1980; Melderis, 1985). Much confusion prevailed regarding the number of species included in this genus. Dewey and Pendse (1967) considered all the crested wheatgrasses, regardless of ploidy level, as a single breeding population. Tzvelev (1976) recognized 10 species in Agropyron and nine subspecies in Agropyron cristatum (L.) Gaertner. Agropyron cristatum is represented by two subspecies in Turkey, namely subsp. incanum (Nábĕlek) Melderis distributed in East Anatolia and subsp. pectinatum (M. Bieb.) Tzvelev distributed throughout Turkey. Morphologically, subsp. incanum is 1 The Ministry of Environment and Forest, Forest Tree Seeds and Tree Breeding Research Directorate, P.O. Box 11, 06560 Gazi, Ankara, Turkey. 2 Corresponding author. E-mail: ecabi@metu.edu.tr, ecabi2004@yahoo.com 3 Department of Biology, Faculty of Science, Gazi University, Ankara, Turkey.


Introduction
The taxonomy of Agropyron Gaertner (Poaceae) was studied in different ways by various researchers. In a broad sense, it was once thought to be one of the largest genera encompassing more than 100 species in the tribe Triticeae Dumort (Dewey, 1983). Nevski (1934) treated Agropyron as a small genus consisting only of the species with keeled glumes. The remaining taxa were placed in Elytrigia Desv, Roegneria C. Koch and Elymus L. Agropyon has been restricted to the species with P genome composed of three ploidy levels (2x = 14, 4x = 28, 6x = 42) (Dewey and Asay, 1975;Melderis, 1978;Dewey, 1983;Assadi, 1995;Jensen et al., 2006). This narrow concept of Agropyron has been accepted by many authorities. The major Eurasian floras have followed the generic concept of Nevski (Tzvelev, 1976;Melderis et al., 1980;Melderis, 1985).
Much confusion prevailed regarding the number of species included in this genus. Dewey and Pendse (1967) considered all the crested wheatgrasses, regardless of ploidy level, as a single breeding population. Tzvelev (1976) recognized 10 species in Agropyron and nine subspecies in Agropyron cristatum (L.) Gaertner.
The aim of this paper is to describe the palynological features of three taxa of the genus Agropyron available in Turkey and also to contribute to their taxonomy showing some differences between them regarding their exine sculptures.

Materials and Methods
Pollen samples from each of the three studied taxa were obtained from herbarium specimens listed in Table 1. The identifications of the specimens were made according to the Agropyron accounts given in Flora of Turkey (Melderis, 1985). For the light microscope study, the pollen grains were prepared following the Wodehouse (1935) and Erdtman (1952) methods. Morphological features of 30 pollen grains were measured using Leica DM 1000 and their microphotographs were taken by the Leica DFC280 camera attachment. The measurements included the following parameters: long axis of spheroidal pollen grains (A), short axis of spheroidal pollen grains (B), long axis of ellipsoidal pore, short axis of ellipsoidal pore, exine thickness, annulus diameter, A/B ratio indicating shape of a spheroidal pollen grain, operculum diameter, thickest part of intine, and intine thickness. Using the average values of pollen measurements, phenograms of the investigated taxa were produced for fresh (W, Wodehouse) and acetolysed (E, Erdtman) pollen grains based on the Gower General Similarity Coefficient. Gower's (1971) coefficient was chosen to generate a distance matrix. This distance matrix was used for cluster analysis with the help of UPGMA algorithm (Sneath and Sokal, 1973). Correlation values were obtained using SPSS version 11.0 (SPSS, 1999).
For scanning electron microscope (SEM) studies, pollen grains were put on stubs, sputter coated with gold and examined under Jeol JSM-6060LV SEM at the central laboratory of Middle East Technical University. The terminologies for pollen morphology were used in accordance with Wodehouse (1935), Faegri and Iversen (1989), and Chaturvedi et al. (1998).

Results and Discussion
The means and standard deviations of the measured pollen parameters of taxa are given in Table 2. All investigated taxa had heteropolar, monoporate (rarely diporate in the case of var. imbricatum) and spheroidal pollen grains. The pore was surrounded by an annulus and it was partly covered by an operculum. Pollen grains of Gramineae were classified as annulate or nonannulate and operculate or nonoperculate by Perveen (2000Perveen ( , 2006, Chaturvedi et al. (1998) and Salgado-Labouriau and Rinaldi (1990). They also rarely observed diporate pollen grains. In this study, we observed that all taxa were annulate and operculate (Figs 1-3).
Exine was thicker in subsp. pectinatum var. imbricatum and subsp. incanum than subsp. pectinatum var. pectinatum (Table 2). Typical measurements for exine thickness of Gramineae are (0.5-)0.85-1.10(-1.53) µm (W) and 1.02-1.61 µm (E) (Erdtman, 1943;Lewis et al., 1983;Salgado-Labourian and Rinaldi, 1990;Liu et al., 2004;Pehlivan et al., 2004).  Liu et al. (2004) found a positive correlation between pollen and aperture size and between aperture and annulus diameter. In our study, both acetolysed (E) and fresh pollen (W) measurements also showed that there are strong positive correlations between the long axis of pollen grain and pore, and short axis of pollen grain and pore (Table 3). Differences between measurements of fresh and acetolysed pollen grains resulted in the formation of different clusters. Although the two subspecies differentiated distinctively and the varieties of subsp. pectinatum formed a tight cluster with respect to their quantitative pollen data obtained from fresh pollen grains, var. pectinatum formed a different cluster due to their acetolysed pollen grains (Fig. 4). The phenogram obtained from measurements of fresh pollen grains is much more suitable for separating the taxa regarding their morphological pollen features in the genus Agropyron. Although acetolysed pollen grains give excellent topographic information, but due to the very process of acetolysis they get modified, thus do not represent actual size. The results indicate that the genus Agropyron has stenopalynous pollen grains, thus the value of pollen characters for taxonomic applications is limited. Faegri and Iversen (1975), Andersen (1978), and Perveen (2006) also indicated similar uniformity in pollen grains of Poaceae. The application of cluster analysis showed the possibility of using quantitative data based on fresh pollen grains for differentiating the taxa. The density of scabrae and the distance between the scabrae can be used as the most functional differentiating characters.