SEM-Based Morphometric Analysis of Pollen Grains in Eight Brassicaceae Taxa from Iraq

The Brassicaceae (Cruciferae) family, known as the mustard family, is considered one of the most significant plant families globally. With numerous economically valuable species, such as Eruca sativa Mill., Raphanus raphanistrum subsp. sativus (L.) Schmalh and Brassica nigra (L.) W.D.J. Koch, while Rapistrum rugosum (L.) All., Sinapis arvensis L, Sisymbrium irio L., Lepidium draba L., and Sinapis alba L. are treated as wild plants or plants growing in disturbed habitats, with more than 338 genera and 3,709 species around the world. In Iraq, more than 75 genera and 18 species are mentioned in the flora. This family has the tendency to withstand ecological stress, making it resilient under global warming. A great importance for agriculture and food security; by doing research on this family, some species are used in pharmaceutical and industrial production [1, 2].

The species within this family have a high degree of similarity in vegetative and floral characters, especially in the cruciform yellow flowers and fruits of the siliqua or silicula type. Hence, external morphology can’t be relied on alone and may not be sufficient for accurate taxonomic delimitation, particularly among closely related genera such as Eruca, Raphanus, Rapistrum, Sinapis, Brassica nigra [3].

Pollen grain morphology has been used as a supportive and helpful character in solving taxonomical problems, especially by Scanning Electron Microscope (SEM), because it reveals the external ornamentation, for example, lumen and muri, which can’t be observed by a light microscope. Some studies mentioned the importance of pollen morphology to solve taxonomical issues at the level of family, genus, and species [4].

The selected species in this study represent a group of closely related and taxonomically important genera within local or regional flora. They all belong to the family Brassicaceae, but differ in genus, fruit type, economic or wild status, and degree of taxonomic affinity. Therefore, they provide a suitable model for evaluating the taxonomic worth of pollen characters in distinguishing species and genera within this family [5].

Although several regional studies have examined Brassicaceae pollen morphology, for example, Mathree and Aliwy [6] studied pollen grains and fruits of eleven Brassicaceae species in Iraq, including Eruca sativa, Sinapis arvensis, and Brassica nigra, but with only a light microscope. As well, Erden and Menemen [5] examined pollen grain morphology in several Brassicaceae taxa, including Sisymbrium irio and Lepidium draba. The family is generally considered stenopalynous, with pollen grains typically described as tricolpate and reticulate. As a result, the taxonomic value of pollen characters relies on detailed quantitative and micromorphological features, especially those observed using SEM. In the local flora, closely related or morphologically similar taxa include Eruca sativa Mill., Raphanus raphanistrum subsp. sativus (L.) Schmalh, Rapistrum rugosum (L.) All., Sinapis arvensis L., Sisymbrium irio L., Lepidium draba L., Brassica nigra (L.) W.D.J. Koch have not been thoroughly compared as a single palynological group.

1.1 Aim of the study

Therefore, the research gap is not the absence of palynological studies, but the lack of an integrated comparative SEM-based study on the selected taxa as one group, supported by quantitative morphometric measurements and exploratory multivariate analyses. This study aims to evaluate the usefulness of pollen micromorphological characters for distinguishing the examined Brassicaceae taxa.

3.1 Overview of pollen morphology

The morphological pattern of pollen grains is generally consistent within the examined taxa. The pollen grains of all taxa studied are monads, with radial symmetry, and are mainly isopolar. The aperture was tricolpate for all taxa and characterized by three elongated colpi. The general shape of the pollen grains ranged from 1.56 to 2.29, which gives them prolate and perprolate shapes, as determined by the P/E ratio [11, 12].

SEM showed that exine ornamentation was reticulate in the taxa studied (Figure 2). However, significant differences were identified in the size and arrangement of the lumina and in the thickness of the muri. These micromorphological variations facilitated the differentiation of certain taxa, especially among closely related genera. While the species shared common palynological features characteristic of Brassicaceae, variation in pollen size, shape, lumen diameter, and muri width offered additional characters with potential taxonomic significance.

3.2 Quantitative pollen characters

Quantitative pollen characteristics exhibited substantial variation within the examined specimens. S. arvensis demonstrated the greatest mean polar axis length (43.23 ± 1.69 µm), while L. draba had the lowest (22.61 ± 3.52 µm). S. arvensis also possessed the highest mean equatorial axis length (23.93 ± 3.38 µm), in contrast to S. irio, which had the lowest value (11.76 ± 1.76 µm). The highest P/E ratio was found in S. irio (2.29 ± 0.42), followed by B. nigra (2.21 ± 0.37). The lowest ratio was observed in R. raphanistrum (1.56 ± 0.28) and L. draba (1.60 ± 0.35), as shown in Table 2.

Exine ornamentation characteristics also varied substantially among the taxa. The largest mean lumen diameter was recorded in S. arvensis (1.81 ± 0.75 µm), followed by B. nigra (1.43 ± 0.52 µm), while L. draba had the smallest lumina (0.51 ± 0.17 µm). The greatest muri width was observed in S. arvensis (0.73 ± 0.13 µm), whereas L. draba had the lowest value (0.49 ± 0.08 µm). Notable within-taxon variation in lumen diameter was detected in several taxa, particularly S. arvensis, E. sativa, and B. nigra.

Table 2. Mean values and standard deviations (SD) of quantitative pollen morphological characters in the studied taxa

Note: P = polar axis; E = equatorial axis; SD = standard deviation; CV = coefficient of variation.

qq20260928-110805.png

Figure 2. The pollen morphology and exine ornamentation of (A) Raphanus raphanistrum, (B) Eruca sativa, (C) Sinapis alba, (D) Sisymbrium irio, (E) Rapistrum rugosum, (F) Sinapis arvensis, (G) Brassica nigra, and (H) Lepidium draba

3.3 Statistical analysis among taxa

For each taxon, descriptive statistics were calculated from the 12 measured pollen grains, including the mean and SD. These statistics were used to summarize grain-level morphometric variation within each examined plant specimen. Boxplots were used to visualize the distribution and variability of the measured pollen characters within the examined specimens (Figure 3).

6862e11f604562c20efddc7df87eb927.png

PCA was conducted as an exploratory multivariate analysis using standardized taxon-level mean values of the quantitative pollen characters. The input matrix comprised eight taxa and five quantitative variables. Hierarchical cluster analysis was similarly performed on the standardized taxon-level mean data using Euclidean distance and Unweighted Pair Group Method with Arithmetic Mean (UPGMA) clustering. Both multivariate analyses were interpreted as exploratory approaches for visualizing morphometric and phenetic similarity within the examined specimens.

3.4 Principal component analysis

PCA using the correlation matrix revealed that the first two principal components accounted for 94.28% of total quantitative pollen variation within the examined specimens (Table 3 and Figure 4). PC1 explained 64.22% of the total variance and exhibited strong positive loadings for polar axis length (0.967), equatorial axis length (0.924), lumen diameter (0.845), and muri width (0.841). Thus, PC1 primarily reflected variation in overall pollen size and dimensions of exine ornamentation.

Table 3. Component loadings of quantitative pollen morphological characters on the first two principal components

Note: PC1: Principal Componen 1; PC2: Principal Component 2.

image010.png

PC2 accounted for an additional 30.06% of the variance and was strongly associated with the P/E ratio (0.986), followed by lumen diameter (0.501). These results indicate that PC2 primarily represented variation in pollen shape, with a secondary contribution from lumen size. The negligible loading of the P/E ratio on PC1 (-0.004) further supports that pollen shape is largely independent of the size-related characteristics represented by PC1.

The exploratory PCA score plot revealed a distinct pattern of morphometric differentiation among the examined specimens (Figure 4). S. arvensis occupied the most positive position along PC1, consistent with its comparatively large polar and equatorial dimensions and with the quantitative characters contributing positively to this component. R. rugosum also showed a positive PC1 score, whereas L. draba was positioned on the negative side of PC1. Along PC2, S. irio and B. nigra were positioned on the positive side, consistent with their relatively high P/E ratios and the strong positive loading of P/E on PC2. In contrast, R. raphanistrum and L. draba showed negative PC2 scores, corresponding to comparatively lower P/E ratios. S. alba, E. sativa, and R. rugosum were positioned closer to the centre of the ordination space, indicating comparatively greater similarity in their combined quantitative pollen characters.

3.4 Hierarchical cluster analysis

This hierarchical cluster analysis revealed patterns of morphometric similarity among the eight taxa (Figure 5). The closest pair occurred between R. rugosum and S. alba, indicating the greatest similarity in the five quantitative pollen characters included in the analysis. R. raphanistrum subsequently joined this subgroup, followed by L. draba at a greater Euclidean distance. A second grouping was formed by S. irio and E. sativa, with B. nigra subsequently joining this cluster.

image011.png

Figure 5. Hierarchical cluster dendrogram of eight examined Brassicaceae taxa based on standardized mean values of five quantitative pollen morphological characters using Euclidean distance and UPGMA clustering

The two major groups merged at a higher distance level. In contrast, S. arvensis showed the greatest separation from all other examined specimens, reflecting its distinctive combination of comparatively large polar and equatorial dimensions and lumen diameter. The cophenetic correlation coefficient (r = 0.776) indicated a reasonably good correspondence between the dendrogram and the original distance matrix. Overall, the clustering pattern was broadly consistent with the exploratory PCA, particularly in showing the marked morphometric separation of S. arvensis from the remaining examined specimens.

The limited biological replication indicates that these pollen characters should not be interpreted as independent evidence of phylogenetic relationships or definitive taxonomic boundaries. The analysis used a selected set of quantitative SEM-derived features, and some detailed pollen characters may need further qualitative evaluation. As well, characters like lumen and muri measurements can change depending on where they are measured on the pollen surface and how the pollen grains are oriented in SEM images. For these reasons, these characters should be interpreted carefully and together with pollen size and shape characters when making taxonomic decisions.