Morphological and chemical evidence of Teucrium × rohlenae K. Malý (Lamiaceae), a new hybrid in Croatia

Introduction

The genus Teucrium L. is represented by approximately 250 species whose centre of the diversity is in the Mediterranean region (Harley et al. 2004). The genus includes shrubs, subshrubs, and herbaceous perennials, rarely annual or biennial plants, mostly growing on open exposed rocky grounds, rock crevices and screes. Based on the characteristics of the calyx and life forms the genus Teucrium is separated into seven sections. The section Polium (Mill.) Schreb. comprises 19 species, where T. montanum L. and T. polium L. belong to the same phylogenetic lineage (Tutin and Wood 1972, Salmaki et al. 2016). As a result of the wide distribution of these two species, chromosome numbers vary greatly within this area. T. montanum is diploid or tetraploid with chromosome numbers 2n = 13, 22, 26, 26+0–2B, 26+0–7B, 30, 60. Chromosome numbers for T. polium are 2n = 26, 26+3–4B, 52, 78, and it can be found as diploid, tetraploid and hexaploid (Ranjbar et al. 2018).

These two taxa are semi-woody, evergreen small shrubs. T. polium is erect up to 45 cm high, whereas T. montanum is usually prostrate, growing up to 30 cm. The most prominent morphological differences of this species are observed in the indumentum and leaf shape. Teucrium polium is distinguished by a very dense indumentum composed of glandular and non-glandular multicellular, uniseriate, branched hairs, whereas T. montanum has a less dense indumentum composed of glandular and non-glandular multicellular, uniseriate and unbranched hairs (Jurišić Grubešić et al. 2007, Lakušić et al. 2010). Teucrium polium is characterised by a dense, hairy calyx, in contrast to the calyx of T. montanum whose calyx has fewer hairs. The calyx teeth of T. polium are moderately obtuse, while T. montanum has acute often setaceous teeth (Maurer 1967, Diklić 1974).

These two species are found in sympatry in certain parts of their distribution, with several hybridogenous taxa being described as: T. × castrense Verg., T. × bogoutdinovae Melinkov and T. × rohlenae K.Malý. The hybridogenous taxon between T. montanum and T. polium first described was collected in July 1907 and June 1908 near the town of Castres in south France by Verguin (Verguin 1908). Verguin named that plant T. × castrense, with T. polium ssp. polium Briquet and T. montanum L. as supposed parental species. Fifty years later, Maurer found individuals with morphological characteristics intermediate between the species T. montanum and T. polium on dunes near Lignano, North Italy (Maurer 1967). This was the first finding of T. × castrense in Italy (Maurer 1967). Another hybridogenous taxon has been described by Rohlena in the area near Kotor in Montenegro (Rohlena 1922). However, according to the Shenzhen Code and Art. H.2.1. (Turland et al. 2018) it was invalidly described as T. montanum × polium (and Malý validly renamed it as T. × rohlenae (Malý 1950). Subsequently, T. × bogoutdinovae from Moldova was described (Melnikov 2014). The species T. reuticum Bogoutdinova and T. polium were listed as parents for T. × bogoutdinove which was described on a rocky terrace of the River Reut (Bogoutdinova 1991). Given that T. reuticum is considered a heterotypic synonym of T. montanum (Eur+Med 2006), T. × bogoutdinove can also be treated as a hybrid between T. montanum and T. polium.

We conducted several field trips, during June 2018 in Bisko near Trilj in Croatia (Fig. 1), where T. montanum and T. polium grow sympatrically. There we found individuals with morphological characteristics intermediate between these two species, which a represent hybridogenous taxon.

Fig. 1.Known occurrences of Teucrium × rohlenae K.Malý based on literature data (A) and recently discovered populations (B, C). A – Kotor in Montenegro, B – Matokit, and C – Trilj, both in Croatia.

The aim of our study is to compare morpho-anatomical and phytochemical characteristics of parents and their putative hybrid and to determine the most important characters of differentiation.

Materials and methods

Because of the small population sizes of T. montanum and the putative hybrid in their narrow hybridization zone (Bisko near Trilj, Croatia), and their presumed endangered status, destructive sampling was limited to a minimum, which resulted in a relatively limited number of analyzed individuals. A total of 29 specimens were selected and scored for analyses. Those include 10 specimens of T. polium, 13 specimens of T. montanum and six specimens of their putative hybrid. Specimens were deposited in the Herbarium of the Institute of Botany and Botanical Garden of the Faculty of Biology, University of Belgrade (BEOU – herbarium code follows Thiers 2019).

Analysed samples are as follows:

  1. montanum
    • CROATIA, Trilj, Bisko (43.579400° N, 16.695995° E), pseudomacchia, leg.: D. Lakušić, B. Lakušić, M. Zbiljić, 30.06.2018, (BEOU – 54028)
      1. polium ssp. capitatum
    • CROATIA, Trilj, Bisko (43.579400° N, 16.695995° E), pseudomacchia, leg.: D. Lakušić, B. Lakušić, M. Zbiljić, 30.06.2018, (BEOU – 54027)
      1. Putative hybrid "T. montanum × polium"
    • CROATIA, Trilj, Bisko (43.579400° N, 16.695995° E), pseudomacchia, leg.: D. Lakušić, B. Lakušić, M. Zbiljić, 30.06.2018, (Teucrium × rohlenae K.Malý, BEOU – 54029)
    • CROATIA, Dalmatinska Zagora, Vrgorac, Mt. Matokit, Zavojane, Zekulić-Roč (43.244743° N, 17.264857° E), rocky ground pasture, leg.: M. Vukojević, 13.06.2015, (sub. T. montanum L., ZAGR – 40074).

For morphological analysis samples of stem with leaves (without terminal inflorescence), leaves, bracts, calyx and flowers from terminal inflorescences were stored in solution of glycerol and ethanol (1:1). Anatomical sections of leaves were preserved on permanent slides, prepared by the standard method for light microscopy. Cross-sections of leaves were cut on a Reichert sliding microtome (up to 15 µm thick). The sections were stained with safranin (1%, w/v, in 50% ethanol) and alcian blue (1%, w/v, aqueous). All slides were mounted in Canada balsam after dehydration (Lakušić et al. 2010). Morphometric analysis included 21 morphological and 23 anatomical characters.

Plant material (3-4 g) was extracted with n-hexane (1:10) for 24 hours at room temperature. The extraction was repeated, and the extracts were united and concentrated under the reduced pressure.

The qualitative and quantitative analysis of hexane extracts was done using the GC and the GC/MS method. The GC analysis was performed on an Agilent 6890N GC system with a 5975 MSD and FID detectors. The column was HP-5 MS (30 m × 0.25 mm, 0.25 μm film thickness). Two μL were injected and the injector temperature was 200 °C with a 10:1 split ratio. Helium was used as a carrier gas (1.0 mL min-1, constant flow mode). The column temperature was linearly programmed (60–280 °C, rate 3 °C min-1, 280 °C for 5 min). The transfer line was heated at 250 °C and the FID detector at 300 °C. EI mass spectra (70 eV) were obtained in the m/z range of 35–550. The retention indices (Kovat's retention index, RI) of essential oil components were experimentally determined relative to two series of n-alkanes (C8-C20 and C21-C40). Their spectra were obtained under the same chromatographic conditions. Identification of compounds was based on comparison of their retention indices (RI) and mass spectra with those from authentic samples and/or the NIST AMDIS software, Wiley, the Adams database and available literature (Adams 2007). Relative percentages of the identified compounds were computed on the basis of the peak areas obtained by FID detector.

The chemical analysis was performed on all compounds (28) of hexane extracts or only on terpene (17) compounds.

In order to describe the variability and significance of morphoanatomical and phytochemical differentiation of analysed groups Principal component analyses (PCA) and Canonical discriminant analysis (CDA) were used, as well as an UPGMA (unweighted pair group method with arithmetic mean) clustering analysis based on Mahalanobis distances for mopho-anatomical data and Pearson’s distances for chemical data in the UPGMA clustering method. Discriminant function analysis (DFA) was used to estimate the contribution of individual characters to the overall discrimination. Statistical analyses were performed using Statistica v.8.0 (StatSoft 2007).

Results

Multivariate analysis of the morpho-anatomical characters

Principal component analysis (PCA) and Canonical discriminant analysis (CDA) performed on morpho-anatomical characters revealed clear separation of T. polium and T. montanum, as well as clear intermediate position of the putative hybrid (Fig. 2A,B). Both analyses revealed a clear separation of three groups along the x-axis, whereas CDA revealed clear separation of putative hybrid along both discriminant axes.

Fig. 2.Multivariate analysis of the morpho-anatomical data for the Teucrium montanum, T. polium and the putative hybrid in their narrow hybridization zone in Bisko near Trilj (Croatia). A – Principal component analysis (PCA), B – Canonical discriminant analysis (CDA).

Cluster analysis performed on all morpho-anatomical characters classified T. polium and the putative hybrid in the first cluster, while T. montanum belongs to the second cluster (Fig. 3A). On the other hand, the cluster analysis based on leaf anatomy revealed the hybrid individuals are closer to T. montanum (Fig. 3B).

Fig. 3.Cluster (UPGMA) analysis of the morpho-anatomical data for the Teucrium montanum, T. polium and putative hybrid in their narrow hybridization zone in Bisko near Trilj (Croatia). A – for all morpho-anatomical characters, B – only for the anatomical characters.

Characters that predominantly contribute to the distinction among the groups are: coverage of adaxial indumentum, thickness of cuticle, number of teeth on the edge of a leaf, bract length, stem length, frequency of lateral tuft on stem nodes, distance between calyx base and tooth base, length of the narrow part of the calyx teeth and number of terminal inflorescences (Tab. 1).

Anatomy Wilks' Lambda Partial Lambda F-remove (2.8) P-level
Radius of the central nerve 0.001401 0.990041 0.040235 0.96076
Coverage of adaxial indumentum 0.001605 0.864346 0.627779 0.55815
Number of capitate hairs C 0.001599 0.867574 0.610556 0.56654
Thickness of adaxial epidermal cells 0.001537 0.90247 0.432283 0.66333
Thickness of cuticula 0.001675 0.828184 0.829843 0.47044
Morphology
Leaf surface 0.002156 0.643504 2.215968 0.17148
Leaf base width 0.001557 0.890808 0.490304 0.62971
Number of teeth on edge of leaf 0.00197 0.704347 1.679023 0.24612
Bract length 0.001626 0.853201 0.688230 0.52991
Average width of bract 0.001747 0.794168 1.036716 0.39779
Stem length 0.001829 0.758577 1.273031 0.33113
Frequency of lateral tuft on stem's nodes 0.001786 0.776835 1.149100 0.36418
Average length of first three internodes 0.002502 0.554452 3.214335 0.09451
Distance between calyx base and tooth base 0.001631 0.850866 0.701094 0.52414
Length of narrow part of tooth 0.001512 0.917384 0.360227 0.70828
Width of tooth 0.001532 0.905726 0.416345 0.67296
Number of flowers in terminal inflorescence 0.001944 0.71378 1.603966 0.25957
Number of terminal inflorescences 0.001763 0.787023 1.082446 0.38366
Tab. 1.Discriminant functional analysis summary of the 18 most important morpho-anatomical characters for 29 specimens of Teucrium montanum, T. polium and the putative hybrid in their narrow hybridization zone in Bisko near Trilj (Croatia). Invariable and highly correlated characters were excluded from analysis.

Chemical analysis of volatile compounds

The chemical analysis of the aerial parts’ volatile compounds revealed that total hydrocarbons (saturated and unsaturated) were the main compounds in all investigated samples (82.9–90.7%). Sesquiterpene hydrocarbons and oxygenated sesquiterpenes were present in amounts that did not exceed 10.2% and monoterpene hydrocarbons only about 1% (Tab. 2).

Compund RI T. montanum putative hybrid T. polium
alpha-Thujene 922 0.39 t t
alpha-Pinene 929 0.18 0.29 t
Sabinene 968 0.40 0.50 t
beta-Pinene 972 0.13 0.16 0.18
beta-Bourbonene 1381 0.49 0.25 t
(E)-Caryophyllene 1416 t 0.97 2.32
(Z)-beta-Farnesene 1439 0.90 0.00 0.00
alpha-Humulene 1450 0.00 t 0.18
(E)-beta-Farnesene 1452 1.01 0.00 t
Germacrene D 1478 3.74 2.46 1.42
Bicyclogermacrene 1493 0.36 0.30 0.20
alfa-Bisabolene 1498 0.30 t 0.00
beta-Bisabolene 1505 t 0.00 0.27
cis-Sesquisabinene hydrate 1540 1.81 0.00 0.00
trans-Sesquisabinene hydrate 1576 0.48 0.00 0.00
epi-alpha-Cadinol 1637 0.51 0.00 0.00
Hexahydrofarnesyl acetone 1838 0.62 t 0.30
Tetracosane 2397 t t 0.10
Pentacosene 2489 1.22 2.93 2.74
Pentacosane 2509 0.00 33.56 46.53
Hexacosane 2589 0.31 0.26 0.34
Heptacosane 2689 3.93 2.88 2.61
Octacosane 2789 0.93 0.57 0.70
Nonacosane 2890 17.45 13.43 9.20
Triacontane 2990 3.12 1.47 1.09
Untriacontene 3092 48.43 20.61 11.40
Untriacontane 3125 0.16 11.94 12.32
Dotriacontane 3187 7.31 3.09 2.13
Percentage of identified compounds 94.18 95.67 94.12
Monoterpene hydrocarbons 1.10 0.95 0.18
Sesquiterpene hydrocarbons 6.80 3.98 4.48
Oxygenated sesquiterpenes 3.42 0.00 0.30
Saturated hydrocarbons 33.21 67.20 75.02
Unsaturated hydrocarbons 49.65 23.54 14.14
Tab. 2.The chemical composition (%) of the volatiles of Teucrium montanum, T. polium and the putative hybrid in their narrow hybridization zone in Bisko near Trilj (Croatia). RI – retention index, t – trace (< 0.1%).

The chemical compositions of the extracts of the studied groups were quite similar. Saturated and unsaturated hydrocarbons represented the main compounds (33.2-75.0% and 14.1-49.6%). Saturated hydrocarbons containing 25, 29 and 31 C atoms, pentacosane (33.6% and 46.5%), nonacosane (9.2-17.5), untriacontane (0.2-12.3%) as well as unsaturated untriacontene (11.4-48.4%) were dominant.

The cluster analysis of all compounds of hexane extracts (Fig. 4A) revealed a similarity in the composition of the volatiles of the aerial parts of T. polium and the putative hybrids. The extracts of T. polium and the putative hybrid were characterised by a pentacosane (46.5 and 33.6%), compound that was not present in the T. montanum volatiles. Unsaturated untriacontene (48.4%) was the main compound in the T. montanum extract.

Fig. 4.Cluster (UPGMA) analysis of the chemical data of volatile compounds for the Teucrium montanum, T. polium and putative hybrid in their narrow hybridization zone in Bisko near Trilj (Croatia). A – for all compounds of hexane extracts (28), B – only for the terpene compounds (17).

On the other hand different results were obtained when only terpene compounds were used for cluster analysis (Fig. 4B). There was a certain similarity between the composition of monoterpenes and sesquiterpenes in extracts of T. montanum and the putative hybrid, mainly represented by higher germacrene D content (3.7% and 2.5%).

Discussion

Intermediate position of putative hybrid

According to the leaf shape and the edge of a leaf with one or more teeth, majority of the hybrid individuals resemble T. polium (Fig. 5). Indumentum of the leaf of the hybrid is less dense than T. polium leaf indumentum, but more dense than indumentum of leaf T. montanum. In addition, non-glandular uniseriate branched hairs that dominate in T. polium and are absent in T. montanum are present in the putative hybrid. At the same time, the hybrid individuals contain many of the non-glandular uniseriate unbranched trichomes present in T. montanum but absent in T. polium (Jurišić Grubešić et al. 2007, Lakušić et al. 2010, Fig. 5).

Fig. 5.Morphological characteristic of inflorescence, leaves and calyx and anatomical characteristic of leaf cross section of Teucrium montanum (m), putative hybrid (×) and T. polium (p) in their narrow hybridization zone in Bisko near Trilj (Croatia). 1 – inflorescence, 2 – leaves, 3 – abaxial leaf side, 4 – adaxial leaf side, 5 – calyx, 6 – leaf cross section: m6 – only non-glandular uniseriate unbranched trichomes in T. montanum, ×6 – both branched and uniseriate nonglandular trichomes in putative hybrid, p6 – only uniseriate branched nonglandular trichomes on both side of leaf in T. polium.

Also, some hybrid individuals have short dense tufts of lateral shoots on the stem nodes, features typical for T. polium, but absent in T. montanum (Fig. 5).

Important indicators of the transition in morphological characters of the putative hybrid are the colour and number of flowers and the consistency of the inflorescences. Teucrium polium has a compact inflorescence with a large number of small white flowers (between 13 and 51, on average 30); T. montanum has a lax inflorescence with a small number of large pale yellow flowers (between 6 and 16, on average 11), and the hybrid individuals has compact inflorescence with a small number of white flowers (between 7 and 20, on average 13) (Fig. 5). Calyx morphology clearly shows intermediate forms. In terms of calyx dimensions (calyx tube length, calyx tooth length and width) the putative hybrids are between the parents: smaller than T. montanum and larger than T. polium. Additionally, the putative hybrid possesses setae on the top of the acutiform setaceous calyx teeth that are shorter than the setae of T. montanum, while setae are missing in T. polium (Fig. 5).

Finally, and in its general habitus, the putative hybrid has a transitional character. Namely, flowering stems of hybrid individuals are erect, as in T. polium, while lateral vegetative shoots are decumbent, as in T. montanum.

A previous work showed that Teucrium species from the Balkan Peninsula usually contain small quantities of essential oil (Kovačević et al. 2001). Sesquiterpenes were the main compounds in the essential oils of the aerial parts of T. montanum and T. polium. In the investigated samples from Montenegro germacrene D (15.0%), α-pinene (12.4) and β-eudesmol (10.1%) were the most abundant in T. montanum, while β-pinene (19.8%) and germacrene D (11.9%) in the oil of T. polium (Kovačević et al. 2001). The essential oil from the aerial parts of T. montanum from Croatia contained germacrene D (17.2%), β-pinene (12.3%) and β-caryophyllene (7.1%), while β-caryophyllene (52.0%) was dominant in the oil of T. polium (Bezić et al. 2011). By contrast, germacrene D (31%) was dominant in the essential oil of T. polium from Serbia, while δ-cadinene (8.1%) and β-caryophyllene (5.1%) were the main compounds in T. montanum (Radulović et al. 2012). Furthermore, the most recent study of the variability of essential oil of different populations of T. montanum from the central and south Balkan Peninsula (14 populations from Serbia, Greece and Albania), revealed extremely large differences in the chemical composition of the essential oils of the aerial parts of T. montanum (Marčetić et al. 2018). This study showed that the composition of essential oils was quite variable but the main compounds in almost all oils were germacrene D (trace-45.5%), sabinene (trace-23.1%), α-pinene (trace-20.7%), limonene (trace-20.4%), (E)-caryophyllene (2.9-14.5%), γ-cadinene (trace-13.8%) and δ-cadinene (trace-12.0%).

These results show that the composition of the essential oils of the two species is not clearly different, given that they contain similar compounds and similar pattern of the variability. It indicates that it is not possible to distinguish them on the basis of the chemical composition of essential oils alone, which is in accordance with the results of several recent studies that showed that variation in the composition of essential oils within a species appears to be the rule rather than the exception, and that the geographical distribution of different types of essential oils of plants is strongly co-related with the environmental conditions including climate, geological, pedological and phytosociological characteristics of the habitats (Kuštrak et al. 1984, Franz 1993, Jug-Dujaković et al. 2012, Lakušić et al. 2012).

However, together with data of morphological and anatomical analysis, the data of the composition of the analysed essential oils also support the assumption of the hybridogenous origin of intermediate individuals.

Taxonomical treatment and distribution

Brief descriptions of hybridogenous taxa T. × castrense and T. × rohlenae largely overlap with features of collected hybrid individuals (Verguin 1908, Rohlena 1922, Malý 1950, Maurer 1967). Considering that the subspecies T. polium ssp. polium, one parent of T. × castrense, is distributed only in the western parts of the Mediterranean (Tutin and Wood 1972, Euro+Med 2006) we have excluded the possibility that hybridogenic individuals from Croatia belong to these taxa.

At the same time, given that T. polium ssp. capitatum (L.) Archangeli and T. polium ssp. vincentinum (Rouy) D.Wood are distributed in Croatia (Nikolić 2019), and that the plant from Trilj belongs to the subspecies T. polium ssp. capitatum also distributed in Montenegro (Euro+Med 2006) we suggest the hybrid be included in taxon T. × rohlenae, originally found and described in area near Kotor in Montenegro (Rohlena 1922, Malý 1950).

We have analysed ZAGR herbarium specimens of T. polium and T. montanum collected in Croatia, and we found specimens that perfectly match the appearance of the hybrid T. × rohlenae. These specimens were determined as T. montanum and were collected in 2015 on Mt Matokit near Vrgorac in Dalmatinska zagora.

Even though T. × rohlenae has been registered only in three locations (Montenegro: Kotor, Croatia: Trilj and Matokit, Fig. 1), it is likely that its distribution is much wider, so we can expect hybrid individuals in areas where T. montanum and T. polium grow sympatrically.

Acknowledgements

This work was supported by the Ministry of Education and Science of the Republic of Serbia (Grants No. 451-03-68/2020-14/200178 to D. Lakušić and 451-03-68/2020-14/200161).

References

  2. Bezić N, Vuko E, Dunkić V, Ruščić M, Blažević I, Burčul F. Antiphytoviral activity of sesquiterpene-rich essential oils from four Croatian Teucrium species.. Molecules. 2011; 16:8119-29. DOI | PubMed
  3. Bogoutdinova TR. A new species of the genus Teucrium (Lamiaceae) from Moldova.. Botanicheskii Zhurnal (Leningrad). 1991; 76:745-6.
  4. Publisher Full Text
  7. Jug-Dujaković M, Ristić M, Pljevljakušić D, Dajić-Stevanović Z, Liber Z, Hančević K. High diversity of indigenous populations of Dalmatian sage (Salvia officinalis L.) in essential-oil composition.. Chem Biodivers. 2012; 9:2309-23. DOI | PubMed
  8. Jurišić Grubešić R, Vladimir-Knežević S, Kremer D, Kalođera Z, Vuković J. Trichome micromorphology in Teucrium (Lamiaceae) species growing in Croatia.. Biologia (Bratisl). 2007; 62:148-56. DOI
  9. Kovačević N, Lakušić B, Ristić M. Composition of the essential oils of seven Teucrium species from Serbia and Montenegro.. J Essent Oil Res. 2001; 13:163-5. DOI
  10. Kuštrak D, Kuftinec J, Blažević N. Yields and composition of sage oils from different regions of the Yugoslavian Adriatic coast.. J Nat Prod. 1984; 47:520-4. DOI
  11. Lakušić DV, Ristić M, Slavkovska V, Šinžar-Sekulić J, Lakušić B. Environmental-related variations of the composition of the essential oils of rosemary (Rosmarinus officinalis L.) from the Balkan Peninsula.. Chem Biodivers. 2012; 9:1286-302. DOI | PubMed
  12. Lakušić B, Stevanović B, Jančić R, Lakušić D. Habitat-related adaptations in morphology and anatomy of Teucrium (Lamiaceae) species from the Balkan peninsula (Serbia and Montenegro).. Flora. 2010; 205:633-46. DOI
  14. Marčetić M, Zbiljić M, Lakušić D, Lakušić B. Variability of essential oil of different populations of Teucrium montanum L. (Lamiaceae) from Balkan Peninsula.. Bot Serb. 2018; 42:133.
  15. Maurer W. Der Bastard Teucrium × castrense = Teucrium montanum × polium bei Lignano in Italien.. Phyton (Horn, Austria). 1967; 12:96-101.
  16. Melnikov D. A new nothospecies of the genus Teucrium L. (Lamiaceae) from Republic of Moldova.. Novitates Systematicae Plantarum Vascularium. 2014; 45:81-3.
  18. Radulović N, Dekić M, Joksović M, Vukičević R. Chemotaxonomy of Serbian Teucrium species inferred from essential oil chemical composition: the case of Teucrium scordium L. ssp. scordioides.. Chem Biodivers. 2012; 9:106-22. DOI | PubMed
  19. Ranjbar M, Mahmoudi C, Nazari H. An overview of chromosomal criteria and biogeography in the genus Teucrium (Lamiaceae).. Caryologia. 2018; 71:63-79. DOI
  20. Rohlena J. Additamenta ad floram dalmaticam.. Acta Botanica Bohemica. 1922; 1:28-34.
  21. Salmaki YS. Phylogeny of non-monophyletic Teucrium (Lamiaceae: Ajugoideae): implications for character evolution and taxonomy.. Taxon. 2016; 65:805-22. DOI
  22. Publisher Full Text
  23. Publisher Full Text
  26. Verguin ML. Un Teucrium hybride nouveau de la section Polium Benth.. Bulletin de la Société Botanique de France. 1908; 55:607-11. DOI