Fruit and seed morphology of 10 taxa belonging to the genus Iberis (Brassicaceae) in Turkey were investigated with stereo microscopy (SM) and scanning electron microscopy (SEM), in order to understand and check their diagnostic significance at generic and spesific level. Fruits and seeds of Iberis species were photographed and examined to evaluate different characteristics including fruit size, shape, color, indumentum, surface ornamentation, style length, sinus width, wing width, septum length, and also seed size, shape, color, surface ornamentation, cell types, anticlinal and periclinal cell wall features. In addition, cluster analysis and principal coordinates analysis (PCoA) were performed to reveal the similarities between the related taxa in terms of seed and fruit morphology. According to fruit morphology, Iberis taxa in Turkey are divided into three main groups particularly by fruit shapes, those are ovate, ovate-obovate and ovate-orbicular. In the present study SM photographs of both immature and mature fruits of relevant species were also provided for the first time. Style length, which is one of the most important features between some Iberis taxa was taken into consideration in the study. There are three types of seed shapes, those are inequilateral obovate, broadly obovate and inequilateral elliptic. The fruit size limits of I. simplex were also been expanded. The UPGMA tree topology of cluster analysis based on a total of 20 fruit and seed characters presents the current taxonomical classification of the taxa in Turkey, which is updated in this study. An identification key for the genus Iberis in Turkey was provided in the light of current data.

Plant specimens used for this study were collected from different regions of Turkey. The author carried out field studies in many parts of Turkey between 2017-2020; and also some collections of other researchers and some herbarium specimens were taken into consideration and listed in Table 1 and Table 3. Relevant literature (Boissier 1867, Hedge 1965, Vural et al. 2012, Oskay 2017, Yıldırımlı 2018) were used for the identification of Iberis taxa. I. umbellata, a cultivated species which was reported only from İstanbul, Turkey (Mutlu 2012) had not been collected and/or evaluated. I. attica and I. spruneri are treated as two separate species instead of I. carnosa.

The author studied and measured the material collected in the field listed in Table 1 and Table 3 and the specimens are deposited in Hacettepe University Herbarium (HUB). These materials are studied together with type and representative material of collections of Iberis from 24 different herbaria worldwide, BM, E, G, K, L, U, LD, LINN, P, WAG, W, WU, AEF, ANK, AIBU, DUOF, EGE, GAZI, HUB, ISTE, ISTF, ISTO, KNYA and MARE (acronyms according to Thiers 2021+).

Morphological measurements were made using an ocular micrometer on stereoscopic binocular (Leica Zoom 2000), with the exception of the larger ones (> 10 mm), which were taken using a standard ruler. For SM photographs of the fruits, two images were taken for each taxa, representing their immature and mature appearances. The fruit and seed samples that best reflect the characteristics of the related species among the populations were photographed. For SM photographs of the seeds, one image was taken for each taxa, representing their general appearance. SM photographs were taken by Dino-Lite and Dino-Eye, the versatile digital microscope using DinoCapture 2.0, version 1.5.40.A programme. In the light of new morphological data from recently collected specimens, a current identification key is prepared for the fruit and seed characters of the genus Iberis in Turkey.

Scanning electron microscope (SEM) analysis

For SEM analysis of the seeds, totally two images were taken for each taxa, representing the general appearance of the seeds of all collected Iberis taxa and the ornamentation of the testa surface. For SEM analysis of the fruits, one image was taken for each taxa, representing the surface ornamentations of the fruit. SEM studies were carried out in both in Hacettepe University Geological Engineering Department and in Hacettepe University Advanced Technologies Application and Research Center (HUNITEK). For SEM analysis, a Zeiss EVO 50 EP (Low Vacuum [~100 pa] VPSE) (in Geo. Eng. Dept) and a Tescan GAIA 3 (Low Vacuum, under 30kV and < 25 nm resolution) (in HUNITEK) scanning electron microscopes were used. Samples for SEM were prepared by directly mounting fruits and seeds onto clean stubs. After coated with gold by using EMS250-333, micrographs of seeds were taken with SEM, connected to a Bruker-Axs Quantax X-Flash 3001 EDS. Taxa, whose fruits and seeds were photographed by scanning electron microscope, their voucher numbers and localities are listed in Table 1, in alphabetical order. 

Multivariative morphometric analysis:

To make measurements for the present study, 12 characters for fruit and 8 characters for seed of Iberis taxa were determined and listed. This dataset including both quantitative measurements and qualitative characteristics were also used for the multivariative morphometric analysis. Measurements were made on 10 individuals representing the species selected from different populations of the species for each species. For the species with a weak population, narrow distribution or few individuals, all available individuals were measured. Totally 298 individuals from 59 different Iberis populations from Turkey were studied (Table 3). PAST (PAleontological STatistics) ver. 4.03 programme was used for the analysis (2001). For the cluster analysis, totally 20 morphological characters  were used and examined. Each qualitative character examined in data set were defined and coded as binary or multiple characters to be converted to numerical data (Table 2). Quantitative characters (FL, FW, STL, SIW, WW, SPL, SL, SW) were measured for each individual, mean values and standard deviations were calculated for each voucher specimen, and a matrix data was generated (Table 5, Supporting data). Cluster analysis was carried out by SAHN (Sequential Agglomerative Hierarchical Nested Cluster Analysis) and UPGMA (Unweighted Pair Group Method with Arithmetic Average) method. Gower similarity index was used to examine both quantitative, binary and nominal data as multiple characters. Besides, relevant morphological characters were also coded in a matrix (Table 2, Table 5) used as input for the principal coordinates analysis (PCoA) which tries to represent (dis)similiarities between mixed variables (Gower 1971, Rohlf 1972, Hammer, et al. 2001).