Towards a taxonomic revision of Calendula L. (Asteraceae)
in the Iberian Peninsula
Ana Carla Gonçalves1, Sofia Nora2, Sílvia Castro3, João Loureiro3, Helena Oliveira1, Conceição Santos4, Paulo Silveira1*
1 Department of Biology & CESAM, University of Aveiro, 3810 - 193 Aveiro, Portugal, 2 Dpto. Biología Vegetal y Ecología. Universidad de Sevilla. 41012, Seville, Spain. 3 CFE & Department of Life Sciences, University of Coimbra, 3000-456 Coimbra, Portugal and 3 Faculty of Science, University of Porto, 4169-007 Porto, Portugal
Authors are very grateful to several herbaria (ABH, AL (ENSA), ARAN, B, BC (IBB), BCN, BM, BONN, BR, C, COI, E, ELVE, FI, G, GAT, GDA (GDAC),
HAL, JACA, JAEN, JE, K, LD, LEB, LISI, M, MA, MACB, MAF, MARK, MGC, MPU, O, P, PO, RAB, RNG, SALA, SANT, SEV, TFMC, US, VAL and W)
allowed the study and loaned specimens of Calendula. Thanks are also due to Portuguese Foundation for Science and Technology and
European Social Fund for financing the work of SC with the Starting grant IF/01267/2013, the work of HO with the scholarship
SFRH/BPD/48853/2008 and the work of AG with the scholarship SFRH/BD/51464/2011.
Results and Discussion
•Morphological analysis for each species
Analysis of variance showed significant (P<0.001)variation among all characters, except bialate achenes’ length, trialate achenes’ length and width
(Fig. 2). Although these characters are not statistically different, other qualitative characters (e.g. shape of wings) support the differences between
PCAs axis accounted for 48%of the total of the variation (Fig. 3). PCAs axis are mostly influenced by the achenes’ variables. PCA 1 x 2 using all taxa
shows four major groupings within Calendula. The C. suffruticosa group reveals to have a high variability between specimens.
•Statistical analysis for C. suffruticosa group
Analysis of variance and PCA were performed on C. suffruticosa comprising two groups: one with arachnoid-pubescent indumentum (GROUP 1) and
another including those with non-arachnoid-pubescent indumentum (GROUP 2). Both analysis of variance showed significant differences among all
characters (p<0.001), except for the ligula length/involucre length ratio.
GROUP 1. PCAs axis accounted for 63%of total of variance. PCA 1 x 2 are mostly influenced by bialate and sub-cymbiform achenes characters (Fig. 4A).
In this projection Cs_cin and Cs_vej appeared to form a group, but when we observed the projection of PCA 2 x 3, these two groups separated, due to
sub-cymbiform achenes characters and thickness of basal leaves (Fig. 4B).
GROUP 2. PCAs axis accounted for 55.2% of total of variance. The trialate and cymbiform achenes loaded heavily on the first component. Although
Cs_lus and Cs_tri formed a single group, they presented other morphological characters and geographical barriers that support the separation of these
two taxa (Fig. 5).
Fig. 3 –Principal component analysis for species
Fig. 2 –Boxplot representing two examples of analysis of variance for species (different letters reveal a
significant differences at P<0.001.)
Fig. 4 –PCA for C. suffruticosa with arachnoid-pubescent indumentum Fig. 5 –PCA for C. suffruticosa with non-arachnoid-pubescent indumentum
•Taxonomic treatments, chromosome counts and flow cytometry
The chromosome number and genome size support the unification of the C. incana and C. suffruticosa groups, as proposed by Meikle (Nora et al.
2013). However, morphological characters, especially the achenes, support the recognition of some of the subspecies proposed by Ohle, and the
description of two new taxa (Table 1) (Gonçalves et al. submitted).
A, et al. (submitted) Phytotaxa
CC, et al. (1974). Israel J Bot 23:1201-1969
J, et al. (2007). Ann Bot 100:875-888
(1976). In: Flora Europaea 4: 206-207
S, et al. (2013)Pl Syst Evol 299:853-864
T(1946). Bot Not 4: 471-506
H (1974). Feddes Repert 85:245-283
H (1975 a). Feddes Repert 86: 1-17
H (1975 b). Feddes Repert 86:525-541
Table 1 –Relationship between different taxonomic treatments, chromosome counts and flow cytometry results for each taxon.
A B C
D E F
Fig. 1 –Heterocarpy in Calendula L.. A –C. suffruticosa, B –C. tripterocarpa, C –C. officinalis, D-F –C. arvensis
(D and E –same population from Alcalá Guadaira, Sevilla).
Mean SD CV
aC. arvensis aC. arvensis Carv 44 5,69 0,65 11,4%
bC. tripterocarpa bC. tripterocarpa Ctrip 30 3,44 0,06 1,7%
1C. officinalis cC. officinalis 1 + c C. officinalis Coff 32 2,98 0,08 2,7%
C. suffruticosa group C. suffruticosa group C. suffruticosa group Csuff
C. suffruticosa subsp. suffruticosa
C. suffruticosa subsp. carbonelii 2 + e d? C. suffruticosa subsp. carbonelii Cs_car 32 3,18 0,08 2,5%
C. suffruticosa subsp. greuterii 3 + e d? C. suffruticosa subsp. greuterii Cs_greu 32 3,36 0,10 3,0%
C. suffruticosa subsp. lusitanica eC. suffruticosa subsp. lusitanica 4 + e C. suffruticosa subsp. lusitanica Cs_lus 32 3,38 0,08 2,4%
4 + g e d? C. suffruticosa subsp. A** Cs_tri 32 3,23 0,10 3,0%
5? + f? C. suffruticosa subsp. B* Cs_mar 32 3,09 0,05 1,5%
C. suffruticosa subsp. C** Cs_vej 32 3,22 0,07 2,2%
C. incana group
C. incana subsp. incana var. incana fC. suffruticosa subsp. tomentosa 5 + f C. suffruticosa subsp. tomentosa* Cs_tom 32 3,31 0,08 2,4%
C. incana subsp. algarbiensis var. cinerea 6 + f C. suffruticosa subsp. cinerea* Cs_cin 32 3,09 0,11 3,4%
C. incana subsp. algarbiensis var. algarbiensis gC. suffruticosa subsp. algarbiensis 7 8 9 + g C. suffruticosa subsp. algarbiensis Cs_alg 32 3,18 0,14 4,4%
C. incana subsp. algarbiensis var. prostrata
C. incana subsp. microphylla
* nomenclatural changes; ** new taxa described; same colour = same taxa in diferent treatments
Genone size (2C/pg)
Calendula L. has its main distribution in the Mediterranean region (Norlindh
1946). It is one of the most taxonomically complex genus of the Asteraceae
family. Although this genus has been subject to several studies (e.g. Heyn et al.
1974; Ohle 1974, 1975a, b; Meikle 1976), no agreement has been reached on its
taxonomy and evolutive relationships.
The genus Calendula L. presents evidence of hybridization events, several
chromosome numbers have been reported (2n = 14 to 88), different genome
sizes and ploidy levels, a large morphological variability (e.g., life cycle, growth
form, leaf form, reproductive strategies, and heterocarpy).
A single capitulum can produce up to 7 types of achenes (Fig. 1), with different
sizes, colours, shapes and with presence/absence of dispersal structures (e.g.
wings, spines). Different combinations can be found among and within taxa.
The aims of this study were:
▪To assess the morphological variability within and among species of the genus;
▪Redefine the relationships and evolution of Calendula species;
▪Support taxonomic decisions for a revision of this genus.
▪Plant material was based on 42 field locations from Iberian Peninsula,
observation of up to 4200 vouchers from 45 herbaria (e.g. BM, K, MA) and
▪Morphological analysis was performed by using 24 quantitative characters
(e.g. plant height and diameter of the head) and 23 qualitative characters (e.g.
life form, colour of florets) from material collected in the field. Statistical
analysis included analysis of variance to test the significance of each
quantitative characters (One-way ANOVA + ttest and Kruskal-Wallis + Mann-
Whitney test), and principal component analysis (PCA) was used to test and
evaluate the contribution of each character.
▪Chromosome counts were made using the squash technique in root tips and
flower buds: pre-treatment for 12hin cold water, fixation for 24hin
ethanol:glacial acetic acid (3:1) and staining with alcoholic hydrocloridric acid-
▪Nuclear DNA content was assessed by flow cytometry by staining with
propidium iodide and using Pisum sativum cv. Ctirad (2C DNA = 9.09 pg) as a
reference standard (Loureiro et al.2007).
Materials & Methods
This study contributes to the knowledge of the genus Calendula L. and provides significant background information
for the taxonomic revision of the genus. In particular, we observed that:
▪morphological characters, both quantitative and qualitative, were found to be useful to distinguish Calendula
▪achene morphology revealed to be particularly important but it is a character that must be used carefully.
Future studies should use molecular methods to corroborate the results obtained and assess evolutionary
relationships among taxa.