Setaria (L.) P. Beauv. (Poaceae, Panicoideae, Paniceae) comprises 125 species and some subsections in tropical regions of the world. There are four species of this genus in Iran (Bor, 1970). Many species of Setaria are important forage and weeds. The only world wide cultivated cereal of this genus is S. italica (L.) P. Beauv. Species as S. palmifolia Koenig is an ornamental garden species in South America (Dekker, 2003).
Inflorescence features provide good diagnostic characters in different groups of grasses which are efficient in species and sub species levels (Keshavarzi et al., 2002, 2005a, b, 2007). In Setaria texture of upper lemma, awn color, ligule shape and etc. are used for differentiating taxa (Davis, 1965; Clayton, 1980; Bor, 1968 and 1970). Morphological used characters are some how overlapping in specific and intra-specific levels.
Setaria species in Iran have some differences in their anatomical structures. These are of diagnostic value (Keshavarzi and Seifali, 2005a). S. viridis shows a great variation in plant color. These are sometimes referred as different varieties (Alex et al., 1972). Isoenzyme studies show that these varieties are basically identical (Dekker, 2003).
In this study we make a vast field trips and gathered many accessions of Setaria. Diagnostic features for differentiating the weedy species (Non- cultivated) of this genus in Iran were revised. Characters were studied biometrically. The aim of this study was to prepare an efficient identification key for this genus in Iran.
MATERIALS AND METHODS
Thirty populations of three species of Setaria native to Iran, were studied from 2004-2005 for their morphological diversity. Details of localities and the voucher are presented in Table 1. Voucher specimens are deposited at Herbarium of Alzahra University (Tehran, Iran).
In total 20 qualitative and quantitative morphological characters were studied (Table 2). The variables were standardized for multivariate statistical analysis. In order to group the populations with morphological similarities, cluster analysis using UPGMA (un-weighted group with arithmetic mean) and WARD (minimum variance spherical clusters) as well as ordination based on Principal Component Analysis (PCA) were performed. The squared Euclidean distance was used as the dissimilarity coefficient in a cluster analysis of morphological data.
In order to determine the most variable morphological characters among the populations, factor analysis based on Principal Component Analysis (PCA) was performed. Statistical analyses were performed by using SPSS (Ver. 9).
Field studies show that in disturbed and managed area in Iran, this weedy genus has more than one species. More often S. viridis is accompanied by one of other two species in such places. This evidence is in accordance with Dekker (2003). These species are mainly differentiated from each other by spikelets fine features in many literatures (Bor, 1968, 1970; Davis, 1965).
|| Population locality and voucher details of studied Setaria
|| Studied qualitative and quantitative characters of Setaria
Our morphological studies show that Setaria species native to Iran have
no variation in the compactness of spike and presence of Palea. But there are
great variations in awn color, the size of glume and lemma, wrinkle and texture
of upper lemma, the presence of hairs in the base of leaf blade and sheath,
the direction of bristle prickles and spike color.
Observations indicate that Setaria species native to Iran are efficiently differentiated by some studied morphological features (Table 3). S. verticillata (L.) P. Beauv. by a unique feature as the direction of bristle prickle (antrorse) is clearly differentiated. There are some confusion between characters of S. verticillata and S. viridis in Iran.
Variance analysis of quantitative morphological characters show that in different
weedy species of Setaria native to Iran, leaf length to width, plant
height, spike and spikelet length, number of bristle prickles, minimum length
of bristles and the length of bristles to length of lower glume have significant
differences between species (Table 4). Anther length was varied
from 4.5 mm. in S. viridis populations to 1 mm. in S. verticillata
and S. glauca.
Results of quantitative morphological characters show that selected set of characters are capable of separation of Setaria Species in Iran. Cluster analysis and dendrograms by Ward method based on quantitative characters (Fig. 1) shows that S. viridis is differentiated from other two species efficiently.
In order to determine the most variable quantitative characters among studied
populations, factor analysis based on PCA was performed revealing that first
three factors comprise more than 92% of total variation in this genus in Iran.
In the first factor with about 66% of total variation, characters such as spike
length, maximum and minimum awn length and awn number in spikelets possessed
the highest positive correlation. In the second factor with about 16% of total
variation, characters such as width of spikelet and the length to width ratio
of leaf blade (Table 5) possessed the highest positive correlation.
|| Results of qualitative morphological features in Setaria
populations native to Iran
|| Descriptive statistical quantitative morphological data In
weedy Setaria species of Iran
|ns = Non- Significant, * Significant in 5%, ** significant
||Principal component analysis of morphological characters in
setaria population of Iran
Observing qualitative morphological features in weedy Setaria populations of Iran, it seems that leaf features especially the leaf base hairs, sheath hairs and ligule shape are capable of making an efficient separation. In the other hand some species as S. glauca are very distinct from others by wrinkled upper lemma. Some confusion is evident in S. verticillata and S. viridis which, due to the documented cases of Flora of Iraq, could be due to the hybridization between these two species. Certainly further studies are necessary to define the hybrids. We should evaluate the characters again and especially field observation is a necessity.
Due to results of quantitative and qualitative morphological characters in studied accessions of Setaria in Iran, the cluster analysis was done. The dendrograms are shown in Fig. 1 and 2. Cluster of quantitative analysis show that accessions of S. verticillata are grouped inside a cluster of S. glauca but Ward dendrograms of qualitative characters dose not show such thing. After merging these two kinds of features, we found that cluster of S. verticillata is clearly separated from other two species. So it is apparent that quantitative features are not good choice to be used alone for Setaria species differentiation in Iran. We should not rely on identification key mainly based on quantitative features in Setaria.
In Fig. 1 it is prominent that S. viridis comprise a unique cluster and no other grouping is shown. In the other hand S. glauca show a kind of grouping. One group comprises accessions of northern and southern Alborz Mountains and the other comprises mainly populations of Central Platue of Iran. Central part of Iran contains dry, desertified low lands, salty deserts and sand dunes.
||Ward cluster analysis of quantitative characters in setaria
species (e = S. viridis, g = S. glauca and v = S. verticillata)
Studying the Ward dendrogram, we found that selected set of qualitative characters
in this genus is capable of a fine separation and are diagnostic. In this genus
two clusters are separated. Setaria glauca is located in one separate
cluster. In S. glauca two groups are defined which have different status
of hairs in base of leaf blade. This grouping is not in congruence with the
grouping based on qualitative characters. Separated cluster are not limited
to definite geographical regions of Iran.
Other two weedy species are grouped in another cluster. In this cluster S.
verticillata comprises a separate position in which two populations of North
of Iran (Noor and Minodasht) are showing a close relationship and the population
of South of Iran (Bajgah, Fars) make a separate neighboring cluster.
||Ward Cluster analysis of Qualitative Morphological features
of Setaria in Iran. (vi stands for S. viridis, v for S. verticillata
and g= S. glauca)
In clusters of S. viridis again there is no geographical base for grouping.
Dendrogram topology reveals the complete separation of these three species in
Iran. After merging qualitative and quantitative characters cluster analysis
result in Fig. 3. In this cluster analysis, species are clearly
differentiated from each other. S. verticillata make a separate cluster
but it shows a close position to the S. viridis than S. glauca.
This final clustering dose not reflects the geographical pattern of distribution
of these three species in Iran.
Principal Component Analysis (PCA) shows this separate position too (Fig. 4). In this diagram S. verticillata comprises an almost compact group. The populations of S. glauca are widely scattered. Due to studied quantitative and qualitative characters in different species of Setaria in Iran, these are clearly differentiated from each other in PCA analysis. Anatomical studies of leaf blades provide efficient diagnostic features in weedy species of Setaria in Iran (Keshavarzi and Seifali, 2005a).
Due to qualitative morphological characters and the results of Factor analysis an identification key for weedy species of Setaria in Iran is provided:
||Ward Cluster analysis after merging Qualitative and Quantitative
Morphological features in Setaria of Iran. (e stands for S. viridis,
v for S. verticillata and g = S. glauca)
||PCA ordination of setaria species of Iran. (Codes as in Fig.
||Spikelet length more than 2 mm, upper lemma wrinkled, upper
glume shorter than upper lemma S. glauca
||Spikelet length less than 2 mm, upper lemma not wrinkled, glume and lemma
in same length
||Bristle prickles exclusively retrorse S. verticillata
||Bristle prickles exclusively antrorse S. viridis
This study shows the efficiency of morphological features and their diagnostic
value in weedy Setaria species native to Iran.