Plant Systematics and Evolution

, Volume 285, Issue 1, pp 83–101

Pollen evolution and its taxonomic significance in Cuscuta (dodders, Convolvulaceae)

Authors

  • Mark Welsh
    • Department of BiologyWilfrid Laurier University
  • Saša Stefanović
    • Department of BiologyUniversity of Toronto Mississauga
    • Department of BiologyWilfrid Laurier University
Original Article

DOI: 10.1007/s00606-009-0259-4

Cite this article as:
Welsh, M., Stefanović, S. & Costea, M. Plant Syst Evol (2010) 285: 83. doi:10.1007/s00606-009-0259-4

Abstract

The pollen morphology of 148 taxa (135 species and 13 varieties) of the parasitic plant genus Cuscuta (dodders, Convolvulaceae) was examined using scanning electron microscopy. Six quantitative characters were coded using the gap-weighting method and optimized onto a consensus tree constructed from three large-scale molecular phylogenies of the genus based on nuclear internal transcribed spacer (ITS) and plastid trn-LF sequences. The results indicate that 3-zonocolpate pollen is ancestral, while grains with more colpi (up to eight) have evolved only in two major lineages of Cuscuta (subg. Monogynella and clade O of subg. Grammica). Complex morphological intergradations occur between species when their tectum is described using the traditional qualitative types—imperforate, perforate, and microreticulate. This continuous variation is better expressed quantitatively as “percent perforation,” namely the proportion of perforated area (puncta or lumina) from the total tectum surface. Tectum imperforatum is likely the ancestral condition, while pollen grains with increasingly larger perforation areas have evolved multiple times. The reticulated tectum, unknown in other Convolvulaceae, has evolved in Cuscuta only in two lineages (subg. Monogynella, and clade O of subg. Grammica). Overall, the morphology of pollen supports Cuscuta as a sister to either the “bifid-style” Convolvulaceae clade (Dicranostyloideae) or to one of the members of this clade. Pollen characters alone are insufficient to reconstruct phylogenetic relationships; however, palynological information is useful for the species-level taxonomy of Cuscuta.

Keywords

ConvolvulaceaeCuscutaDoddersEvolutionPhylogenyPollen morphologyScanning electron microscopyTaxonomy

Introduction

The taxonomic significance of pollen morphology in Convolvulaceae has long been recognized. For example, Hallier (1893) assigned the genera within this family to two major groups, “Echinoconieae” and “Psiloconiae,” based on their echinate or psilate exine, respectively. Together with other characters, the diversity of pollen morphology in the morning glory family has served for the separation of genera such as Calystegia and Convolvulus (Lewis and Oliver 1965), Stylisma and Bonamia (Lewis 1971), Merremia and Operculina (Ferguson et al. 1977), and Maripa, Dicranostyles, and Lysiostyles (Austin 1973a, 1973b), as well as for the circumscription of species, e.g., Ipomoea spp. (Hsiao and Kuoh 1995) and Convolvulus spp. (Menemen and Jury 2002). Not surprisingly, pollen has been used to assess the evolutionary relationships in Convolvulaceae. For instance, Sengupta (1972) proposed an evolutionary arrangement of the family with four major pollen types based on the number and distribution of apertures. Tellería and Daners (2003) found exine to be more relevant taxonomically than aperture features, and based on the former characters, distinguished three major groups of pollen: tectate, microechinate-perforate; tectate, microechinate-perforate with microspines; and semitectate, microechinate-microreticulate.

Cuscuta (dodders), a genus comprising over 180 species of holoparasitic vines (Stefanović et al. 2007), is nested within Convolvulaceae (Stefanović and Olmstead 2004). It represents the third most economically important group of parasitic plants after Striga and Orobanche, because infestation by ca. 15 of its species can result in significant yield losses in numerous crops worldwide (Parker and Riches 1993; Dawson et al. 1994; Costea and Tardif 2006). Additionally, numerous Cuscuta species are rare and endangered, requiring conservation measures (Costea and Stefanović 2009a).

As a result of their parasitic lifestyle, dodders exhibit extreme reductions of the vegetative structures, limiting the morphological characters available for systematic studies to flowers and fruit (Stefanović et al. 2007). It is therefore imperative to search for and discover new characters useful for the taxonomy of the genus, as well as to create a theoretical basis for character evolution analysis.

Pollen information for Cuscuta is relatively scarce. Das and Banerji (1966) described the “rugulate” pollen surface of C. santapaui and C. reflexa, while Jain and Nanda (1966) compared the pollen morphology of Cuscuta hyalina and Convolvulus pluricaulis Choisy. Sengupta (1972) studied 21 species of Cuscuta, which he divided into two groups according to their tricolpate or penta-hexa-colpate pollen. Liao et al. (2005) analyzed four species from Taiwan, also recognizing two main types of pollen based on exine morphology. Type 1, observed in C. australis, C. campestris, and C. chinensis, is characterized by an ektexine “finely reticulate,” whereas type 2 exhibited a reticulate ektexine, as seen in C. japonica (Liao et al. 2005). Despite the limited sampling, these studies concluded that pollen provides important phylogenetic and taxonomic information. Recently, Costea et al. (2006a, 2006b, 2006c, 2006d, 2008a, 2008b) described the pollen of 24 species as a part of taxonomic revisions of major clades that belong to subg. Grammica. Therefore, to date, pollen morphology of only about one-quarter of Cuscuta species is known. More importantly, pollen diversity in this genus has never been analyzed in a broad-scale evolutionary context, in a firmly established phylogenetic framework.

The precise sister group relationships of Cuscuta with other Convolvulaceae members are not clear (Stefanović et al. 2002, 2003; Stefanović and Olmstead 2004). However, well-supported phylogenies based on both plastid and nuclear datasets are available for the genus itself (García and Martin 2007; Stefanović et al. 2007). This newly established phylogenetic framework enables the examination of pollen characters from an evolutionary perspective. Thus, the main goals of this study are to: (1) survey the diversity of pollen morphology across the genus, (2) place this morphological variation into an evolutionary context, and (3) assess the usefulness of pollen exine morphology for the systematics of Cuscuta.

Materials and methods

Sampling and scanning electron microscopy

A total of 148 taxa (135 species and 13 varieties) were examined using 372 herbarium specimens (Appendix). Efforts were made to sample multiple accessions, particularly for those species spanning large biogeographical ranges and/or having a diverse morphology. As a result, with the exception of the species known from only one specimen, all of the examined taxa are represented by two or more collections. Mature anthers were fragmented on the stubs without acetolysis to preserve the exine and intine (Harley and Ferguson 1990). Samples were coated with 20 nm of gold using an Emitech K 550 sputter-coater, and examined with a Hitachi S-570, Hitachi SU-1500 or LEO 1530 FE-SEM at 10-15 kV. Photographs illustrating the details of pollen for all the taxa are provided on the Digital Atlas of Cuscuta website (Costea 2007 onwards). Pollen measurements were performed on digital SEM images using Carnoy 2.0 for Mac OS X (Schols et al. 2002), and ImageJ (Abramoff et al. 2004) was used for determination of areas.

Pollen characters

We used the terminology of Punt et al. (2007) to preliminarily evaluate the morphological variation of tectum perforations into discrete types, potentially utilizable as qualitative state characters. The corresponding tectum types encountered in Cuscuta are: imperforate (no perforations present), perforate (tectum with puncta <1 μm), microreticulate (reticulate ornamentation consisting of muri and lumina <1 μm), and reticulate (similar to the previous, but lumina >1 μm) (Punt et al. 2007). However, the tectum variation in Cuscuta could not be consistently separated into these types, because complex morphological intergradations occur, especially among the imperforate, perforate, and microreticulate pollen grains of different species. Therefore, we defined tectum variation quantitatively as “percent perforation,” namely the proportion of the perforation surface (puncta or lumina) from the total surface of the tectum. Comparable quantitative measures, e.g., the “perforation density” (Vezey et al. 1991) and “percent tectum coverage” (Vezey et al. 1992), have previously been used in other groups of plants, yet they have not achieved widespread acceptance despite the fact that they provide a more accurate description of tectum morphology.

Five other quantitative characters—pollen length (polar axis), polar/equatorial (P/E) ratio, average diameter of perforations (puncta or lumina), average surface of perforations, and number of colpi—were also examined. The number of colpi exhibited a discrete variation: 3(–4) or (4–)5–8 colpi. The remaining characters varied continuously and were coded using Thiele’s (1993) gap-weighting method as implemented by MorphoCode (Schols et al. 2004). Gap-weighting was preferred to various gap-coding methods (reviewed by Wiens 2001; Swiderski et al. 1998) because of the better phylogenetic signal recovered (see also García-Cruz and Sosa 2006). The maximum number of resulted character states (n) was ten for all the quantitative characters, except for the percent perforation where n was eight. Eight character states were sufficient to describe tectum perforation patterns (Table 1).
Table 1

Percent perforation quantitative character states and their corresponding tectum “types” resulted from coding using Thiele’s (1993) gap-weighting method

Percent perforation character states determined with MorphoCode (Schols et al. 2004) (%)

Diameter of puncta/lumina (μm)

Corresponding tectum “types”

0–2.1

0.2 (0–0.6)

Tectum imperforatum (TI)

2.6–6.3

0.43 (0.14–1.2)

Tectum perforatum 1 (TP1)

8.3–12.1

0.62 (0.14–1.5)

Tectum perforatum 2 (TP2)

12.5–17.3

0.67 (0.17–1.62)

Microreticulate 1 (MR1)

17.7–21.9

0.7 (0.2–1.77)

Microreticulate 2 (MR2)

30.8–31.3

1.9 (0.85–2.91)

Reticulate 1 (R1)

34.3

2.65 (1.52–3.85)

Reticulate 2 (R2)

43.6–44.6

3.03 (0.8–5.82)

Reticulate 3 (R3)

Using formal outgroup analysis (e.g., Maddison et al. 1984) to determine character polarity in Cuscuta is hindered by two factors. First, despite considerable efforts, outgroup relationships of Cuscuta are unknown (Stefanović and Olmstead 2004). The position of Cuscuta in Convolvulaceae was, however, narrowed down to three possible phylogenetic scenarios (Stefanović et al. 2002; Stefanović and Olmstead 2004): (a) Cuscuta as a sister to the “bifid-style” clade (Dicranostyloideae) which comprises the tribes Hildebrandtieae, Cresseae, Dichondreae, and in part Convolvuleae, Poraneae, and Erycibeae; (b) Cuscuta as a sister to the “bifid clade” together with “clade 1” (Convolvuloideae) which includes the tribes Ipomoeae, Argyreieae, Merremiae, and some Convolvuleae. Together or individually, these major Convolvulaceae clades account for most of the diversity encountered in the family (e.g., “clade 1” alone has over two-thirds of the species in Convolvulaceae); (c) Cuscuta as a sister to one of the members of the “bifid clade,” although this possibility was deemed “unlikely” and could not be formally tested because the relationships within this clade were unresolved (Stefanović et al. 2002; Stefanović and Olmstead 2004). Second, not all the Convolvulaceae genera/species from these groups have been studied in regards to their pollen morphology. For these reasons, our interpretation of character polarity in Cuscuta also takes into account the ingroup distribution of character states at the level of both Cuscuta and Convolvulaceae (reviewed by Stuessy 2008).

Characters were mapped onto a summary consensus tree built in MacClade 4 (Maddison and Maddison 2000), resulting from the combination of two large-scale molecular phylogenies of Cuscuta based on plastid trn-LF and nuclear ITS sequences (subg. Cuscuta, García and Martin 2007; subg. Grammica Stefanović et al. 2007), and an unpublished phylogeny of the entire genus (Stefanović and Costea, personal communication). ACCTRAN and DELTRAN were turned off and instances of bootstrap values below 85% were considered unresolved and are indicated in the tree as polytomies.

Results and discussion

Number of apertures

Pollen of Cuscuta is heteromorphic (sensu Till-Bottraud et al. 1995) (Fig. 1a–d). Over 95% of the species examined can be characterized as 3-zonocolpate, but this prevalent apertural type may be accompanied in the same anther by a small proportion of 4-, 5- or even 6-zonocolpate grains, and extremely rarely by pantocolpate grains. A similar variation of ±2 apertures can be observed in the species with preponderantly 5- and 7-zonocolpate pollen grains, which in addition may also produce pantocolpate morphs. Apertural heteromorphism is common in numerous angiosperms (Erdtman 1966; Van Campo 1976; Mignot et al. 1994) and can be linked ontogenetically to the succession of events that take place during meiotic cytokinesis (Blackmore and Crane 1998; Ressayre et al. 2002, 2005). Experimental results from heteromorphic eudicots have shown that 4-apertured grains germinate faster than 3-aperturate ones, but the latter have faster pollenic tube growth and better survival than the former (Dajoz et al. 1991; Till-Bottraud et al. 1999). Together, these different morphotypes and their corresponding pollen strategies maximize the chances of successful fertilization under different conditions.
https://static-content.springer.com/image/art%3A10.1007%2Fs00606-009-0259-4/MediaObjects/606_2009_259_Fig1_HTML.jpg
Fig. 1

Variation of colpi number (ad). aCuscuta purpusii, bC. argentiniana, cC. parodiana, dC. boliviana. Tectum variation (el, respectively); see Table 1 for abbreviations of tectum types. eCuscuta brachycalyx (TI), fC. odontolepis (TP1), gC. polyanthemos (TP2), hC. cozumeliensis (MR1), iC. chapalana (MR2), jC. cassytoides (R1), kC. japonica (R2), lC. reflexa (R3). Tectum, surface detail (Table 1) (mp) mC. decipiens (TI), nC. exaltata (TP2), oC. mitriformis (MR2), pC. santapaui (R3). Scale bars (A–L) 5 μm, (M–P) 0.5 μm

Although the number of apertures is not perfectly fixed within the species of Cuscuta, this character is phylogenetically informative. Sengupta (1972) suggested that the 5–6-colpate grains of Cuscuta reflexa are derived compared with the 3-colpate pollen encountered in other dodder species. Our results support this hypothesis, because pollen grains with a higher number of apertures (5–8) have evolved in Cuscuta from the ancestral state with three colpi only in two lineages (Fig. 2), in subg. Monogynella (C. reflexa and C. japonica) and in several species of a South American clade that belongs to subg. Grammica (clade O, see Stefanović et al. 2007). Sengupta (1972) proposed that an increased number of apertures in Cuscuta is associated with polyploidy. While C. reflexa is polyploid with several cytotypes (2n = 28, 30, 32, 42, 48; Kaul and Bhan 1977), the very scarce cytological information available for the remaining species does not seem to support this hypothesis. The entire genus is a polyploid complex, and some of the species with the highest numbers of chromosomes such as C. campestris (2n = ca. 56), C. cephalanthi, and C. gronovii (2n = 60) (Pazy and Plitmann 1995) are tricolpate.
https://static-content.springer.com/image/art%3A10.1007%2Fs00606-009-0259-4/MediaObjects/606_2009_259_Fig2_HTML.gif
Fig. 2

Colpi number optimized onto a summary consensus tree resulting from three molecular phylogenies of Cuscuta based on nuclear ITS and plastid trn-LF sequences (García and Martin 2007; Stefanović et al. 2007; Stefanović and Costea, personal communication). Pollen grains with 5–8 colpi have evolved only in subg. Monogynella and in clade O of subg. Grammica

The number of apertures has received considerable attention in Convolvulaceae. Similarly to other eudicots, tricolpate pollen has been regarded as plesiomorphic, while 5-6-zonocolpate, pantocolpate, and pantoporate grains are considered progressively derived in the family (Wodehouse 1936; Vishnu-Mittre 1964; Manitz 1970; Muller 1970; Sengupta 1972; Austin 1973a, 1973b, 1998; Tellería and Daners 2003). This evolutionary sequence, termed “successiformy” by Van Campo (1976), can be encountered in genera from both Convolvuloideae and Dicranostyloideae. For example, in the former clade, zonocolpate pollen grains with five or more apertures are found in Odonellia, a genus with two species (Robertson 1982), and several Meremia spp. [e.g., M. umbellata (L.) Hallier, Sengupta 1972; Tellería and Daners 2003; Leite et al. 2005; M. vitifolia (Burm. f.) Hallier f., and M. sibirica (L.) Hallier f., Ferguson et al. 1977]. In Dicranostyloideae, Maripa and Jacquemontia species exhibit complex heteromorphic variation patterns from tricolpate to pantocolpate (Robertson 1971; Austin 1973b).

Exine

Exine in Cuscuta is tectate imperforate or semitectate, perforate to reticulate (Fig. 1e–l), with a single layer of unbranched columellae. Supratectal ornamentation typically consists of rounded to acute scabrate processes less than 1 μm long, more or less evenly distributed on the pollen surface (Fig. 1m–p). Larger supratectal conical spines (>1 μm) are present only in subg. Monogynella in C. lehmanniana and C. monogyna. Pollen with reticulate tectum is unknown in other Convolvulaceae (see below) and has evolved in Cuscuta only in some species of subg. Monogynella and clade O of subg. Grammica (Fig. 3; Table 2). Sengupta (1972) characterized the pantoporate pollen grains of Ipomea as “complex-reticulate,” namely reticulate with a superimposed hexagonal pattern (the metareticulate pollen of Borsch and Barthlott 1998; Tellería and Daners 2003) and suggested that this type might have originated from the hexacolpate, “simple reticulate” of C. reflexa. However, as reported by Tellería and Daners (2003), the metareticulate pollen of many Ipomoea spp. in fact has a microreticulate tectum, with a different exine structure and ornamentation.
Table 2

Morphology of pollen in Cuscuta. Species are grouped into subgenera/major clades (García and Martin 2007; Stefanović et al. 2007; Stefanović and Costea, personal communication). Species for which no molecular data was available (indicated by an asterisk) are also tentatively placed into major infrageneric groups based on their morphology. P = prolate; SP = subprolate; S = spheroidal; SO = suboblate

Infrageneric group

Species

Species details

Percent perforation (%)

Avg. perforation area (μm2)

Perforation diameter (μm)

Length (μm)

Width (μm)

P/E ratio

Shape

Number of colpi

Subg. Monogynella (eight species)

C. cassytoides

31.3

0.949

1.523–(2.651)–3.885

24.8–(26.4)–28.6

19.5–(22.3)–24.3

1.18

SP

3 (–4)

C. exaltata

8.3

0.313

0.303–(0.744)–1.366

31.6–(34.5)–37.2

23.1–(25.9)–29.5

1.33

P (–SP)

3 (–4)

C. japonica

30.8

0.855

1.167–(2.491)–4.518

29.5–(31.5)–33.1

21.6–(23.7)–27.5

1.40

P (–S)

(4–) 5 (–6)

C. lehmanniana

19

0.63

0.777–(1.285)–1.685

26.0–(29.6)–32.8

24.0–(26.4)–29.3

1.12

(P–) S

3 (–4)

C. lupuliformis

34.3

0.827

0.855–(1.284)–2.198

27.1–(29.4)–31.4

21.3–(23.5)–26.3

1.25

SP (–S)

3 (–4)

C. monogyna

0

0

0

31.8–(32.9)–33.7

24.1–(28.2)–30.1

1.17

(P–) SP (–S)

3 (–4)

C. reflexa

44.6

2.289

3.867–(4.470)–5.827

29.5–(30.6)–31.7

26.2–(27.7)–29.1

1.10

S (–SO)

5–6 (–7)

C. santapaui*

38

1.524

1.618–(3.202)–4.541

25.0–(26.3)–27.1

29.5–(30.8)–32.2

0.85

(SP–) SO

(5–6)

Subg. Cuscuta (five species)

C. approximata

0.2

0.05

0.286–(0.382)–0.453

22.0–(23.7)–25.1

18.5–(20.6)–24.3

1.15

(P–) SP

3 (–4)

C. epilinum

0

0.002

0.184–(0.257)–0.322

24.6–(26.0)–27.4

17.1–(18.3)–19.2

1.42

P

3 (–4)

C. epithymum

0

0

0

18.6–(19.7)–21.8

12.9–(14.2)–15.9

1.39

P (–SP)

3 (–4)

C. europea

0.4

0.052

0.281–(0.390)–0.584

20.7–(23.1)–25.8

12.4–(13.8)–14.9

1.67

P

3 (–4)

C. planiflora

0

0

0

22.4–(24.1)–25.3

12.2–(17.0)–19.0

1.42

P (–SP)

3 (–4)

Pachystigma clade (five species)

C. africana

3.8

0.257

0.468–(0.564)–0.662

28.0–(29.6)–31.2

17.4–(19.4)–20.6

1.53

P (–S)

3 (–4)

C. angulata

0.1

0.063

0.415–(0.430)–0.461

26.8–(28.0)–29.5

15.2–(15.8)–17.1

1.77

P

3 (–4)

C. appendiculata

0

0

0

17.1–(18.9)–20.0

16.0–(17.6)–19.4

1.07

S (–SO)

3 (–4)

C. natalensis

1.6

0.044

0.168–(0.289)–0.518

17.2–(17.6)–18.3

18.0–(19.2)–19.9

0.92

(P–) S

3 (–4)

C. nitida

0.1

0.029

0.211–(0.288)–0.393

15.6–(18.2)–22.6

14.9–(19.0)–22.1

0.96

(P–) S

3 (–4)

Subg. Grammica (130 taxa) clade A

C. brachycalyx

0

0

0

22.9–(24.0)–24.8

15.0–(16.0)–17.0

1.50

P

3 (–4)

C. californica

0

0

0

19.3–(21.5)–23.0

11.0–(17.5)–20.8

1.23

SP

3 (–4)

C. decipiens

0.1

0.049

0.243–(0.298)–0.334

19.9–(21.3)–23.0

13.5–(14.3)–14.7

1.49

P (–SP)

3 (–4)

C. draconella

0

0

0

15.0–(16.2)–17.4

13.2–(14.6)–16.7

1.11

(P–) S

3 (–4)

C. howelliana

6.3

0.107

0.286–(0.382)–0.542

13.5–(15.6)–16.8

9.2–(11.5)–13.5

1.36

P (–SP)

3 (–4)

C. jepsonii

0.2

0.048

0.181–(0.237)–0.371

19.2–(21.3)–22.9

13.3–(14.0)–14.9

1.52

P

3 (–4)

C. occidentalis

0

0

0.161–(0.295)–0.484

19.4–(21.6)–24.6

12.6–(16.2)–18.8

1.33

P (–SP)

3 (–4)

C. salina var. salina

1.2

0.069

0.314–(0.379)–0.547

16.5–(17.2)–21.3

10.6–(11.8)–14.5

1.46

P (–SP)

3 (–4)

C. salina var. major

1

0.022

0.239–(0.367)–0.800

17.1–(17.7)–18.2

14.7–(15.9)–17.2

1.11

S

3 (–4)

C. subinclusa

0.2

0.04

0.254–(0.327)–0.439

14.7–(15.0)–21.1

15.6–(16.4)–16.9

0.91

(SP–) S

3 (–4)

C. suksdorfii

0

0

0.287–(0.359)–0.499

24.9–(26.6)–27.6

12.4–(15.7)–17.8

1.69

P

3 (–4)

Clade B

C. australis

0.4

0.045

0.246–(0.317)–0.439

18.9–(20.9)–23.5

11.4–(13.7)–16.9

1.53

P (–SP)

3 (–4)

C. campestris

0.8

0.056

0.213–(0.282)–0.367

17.6–(23.6)–26.9

18.0–(18.3)–18.7

1.29

(P–) SP

3 (–4)

C. glabrior

0.1

0.069

0.352–(0.416)–0.485

17.3–(18.0)–21.1

10.8–(12.2)–13.2

1.48

P (–SP)

3 (–4)

C. gymnocarpa

0.2

0.042

0.222–(0.312)–0.445

20.3–(21.1)–22.3

15.2–(17.4)–19.2

1.21

(P–) SP

3 (–4)

C. harperi

0.4

0.037

0.197–(0.230)–0.287

16.5–(17.5)–23.4

9.7–(13.3)–16.0

1.32

(P–) SP

3 (–4)

C. obtusiflora var. obtusiflora

4.7

0.03

0.183–(0.260)–0.396

18.9–(20.1)–21.1

12.9–(13.7)–14.9

1.16

SP

3 (–4)

C. obtusiflora var. glandulosa

0.1

0.066

0.222–(0.309)–0.442

17.4–(21.2)–23.6

14.0–(18.3)–20.1

1.47

P (–SP)

3

C. pentagona

0

0

0

18.6–(19.3)–22.8

12.0–(12.1)–13.5

1.60

P

3 (–4)

C. plattensis

0

0

0

16.5–(17.5)–22.4

16.6–(17.1)–17.4

1.02

S (–SO)

3 (–4)

C. polygonorum*

0

0

0

19.6–(21.8)–23.9

14.2–(16.1)–17.8

1.35

P (–SP)

3 (–4)

C. runyonii

0.1

0.017

0.166–(0.231)–0.282

13.0–(14.0)–18.6

16.2–(17.3)–18.3

0.81

(SP–) SO

3 (–4)

C. stenolepis

0

0

0.283–(0.424)–0.643

16.6–(18.2)–20.0

13.7–(14.1)–14.8

1.29

(P–) SP

3 (–4)

Clade C

C. corniculata

0.5

0.01

0.153–(0.210)–0.335

20.8–(21.9)–23.2

17.5–(17.8)–18.3

1.23

SP (–SO)

3 (–4)

C. incurvata

1.3

0.049

0.211–(0.353)–0.500

13.0–(14.0)–14.8

15.1–(16.5)–17.5

0.85

(SP–) SO

3 (–4)

C. micrantha

0

0

0

13.4–(15.8)–17.9

16.9–(17.5)–18.7

0.90

(SP–) S

3 (–4)

C. parviflora var. elongata

0

0

0.187–(0.199)–0.212

17.5–(18.2)–18.8

16.6–(17.4)–18.8

1.05

S

3 (–4)

C. pauciflora*

0.2

0.031

0.207–(0.345)–0.537

14.4–(16.7)–18.5

15.7–(17.5)–19.3

0.95

S

3 (–4)

C. platyloba

1.2

0.02

0.079–(0.211)–0.383

16.5–(16.9)–17.2

10.8–(12.5)–14.3

1.35

P (–SP)

3 (–4)

C. racemosa var. racemosa*

0

0

0

15.2–(16.7)–18.2

12.6–(13.5)–15.1

1.24

SP (–S)

3 (–4)

C. racemosa var. miniata

2.4

0.038

0.145–(0.273)–0.389

14.4–(15.7)–17.8

15.7–(17.4)–19.9

0.90

(SP–) S

3 (–4)

C. suaveolens

1

0.027

0.195–(0.269)–0.380

14.8–(16.9)–17.6

14.6–(16.1)–17.5

1.05

(P–) S

3 (–4)

C. werdermanii

0.4

0.048

0.209–(0.268)–0.314

15.1–(17.9)–20.0

14.8–(17.0)–18.0

1.05

(P–) S

3 (–4)

C. xanthochortos var. xanthochortos*

0.1

0.023

0.490–(0.521)–0.537

18.9–(19.1)–19.6

21.7–(22.3)–23.3

0.94

(SP–) S

3 (–4)

C. xanthochortos var. carinata

0.2

0.036

0.132–(0.235)–0.420

15.8–(16.9)–19.3

16.6–(18.0)–19.1

0.86

(SP–) SO

3 (–4)

C. xanthochortos var. lanceolata*

0

0.009

0.144–(0.177)–0.195

13.6–(14.4)–15.4

16.2–(16.6)–17.2

0.87

(SP–) SO

3 (–4)

Clade D

C. cephalanthi

20.1

0.137

0.205–(0.378)–0.707

28.6–(29.3)–30.8

20.7–(22.4)–23.6

1.31

(P–) SP

3 (–4)

C. compacta

4.3

0.065

0.423–(0.497)–0.561

22.3–(23.4)–24.6

14.3–(17.2)–19.0

1.36

P (–SP)

3

C. cuspidata

0.2

0.051

0.199–(0.251)–0.331

24.2–(25.2)–25.8

12.2–(14.0)–16.5

1.80

P (–SP)

3 (–4)

C. glomerata

2.6

0.084

0.227–(0.470)–0.751

21.5–(22.7)–23.7

18.1–(19.0)–19.5

1.19

(P–) SP

3 (–4)

C. gronovii var. gronovii

3.6

0.121

0.279–(0.411)–0.628

19.8–(20.6)–23.6

15.1–(16.9)–19.3

1.22

SP

3 (–4)

C. gronovii var. latiflora

12.1

0.075

0.141–(0.328)–0.685

3 (–4)

C. rostrata

15.2

0.098

0.314–(0.666)–1.246

22.2–(23.9)–25.2

13.8–(15.3)–17.8

1.56

P (–SP)

3 (–4)

C. squamata

2.1

0.075

0.232–(0.304)–0.431

21.8–(23.1)–23.9

14.7–(15.6)–17.8

1.48

P (–SP)

3 (–4)

C. umbrosa

1.5

0.029

0.327–(0.508)–0.724

21.9–(23.9)–25.4

14.1–(16.0)–18.3

1.49

P (–SP)

3 (–4)

Clade E

C. denticulata

0

0

0

14.4–(15.2)–16.1

11.4–(11.8)–11.9

1.29

SP

3 (–4)

C. nevadensis

4.9

0.058

0.150–(0.335)–0.716

20.0–(20.8)–22.2

12.9–(13.7)–14.0

1.52

P

3 (–4)

C. veatchii

0.4

0.015

0.137–(0.193)–0.311

17.8–(19.4)–20.6

13.9–(14.9)–16.2

1.30

(P–) SP

3 (–4)

Clade F

C. burrelli

4

0.044

0.183–(0.319)–0.455

16.7–(17.5)–18.0

13.7–(16.5)–16.6

1.06

(SP–) S

3 (–4)

C. haughtii

12.5

0.186

0.298–(0.561)–1.053

19.1–(20.8)–21.8

15.4–(18.4)–21.3

1.13

(SP–) S

3 (–4)

C. longiloba

1.5

0.014

0.210–(0.276)–0.351

15.0–(18.7)–21.8

17.2–(22.6)–25.7

0.83

(S–) SO

3 (–4)

C. partita

3.2

0.019

0.163–(0.330)–0.562

19.6–(21.5)–23.4

14.3–(15.5)–16.6

1.39

P (–SP)

3 (–4)

C. serrata*

0

0

0

18.9–(19.8)–20.7

13.2–(16.8)–19.0

1.18

(P–) SP

3 (–4)

Clade G

C. aurea

0.1

0.052

0.302–(0.408)–0.505

17.5–(18.6)–19.6

13.6–(14.9)–15.8

1.25

(P–) SP

3 (–4)

C. cotijana

20.2

0.405

0.213–(0.306)–1.338

14.3–(18.3)–22.0

13.1–(16.2)–17.4

1.13

(SP–) S

3 (–4)

C. floribunda

8.6

0.095

0.420–(0.560)–0.693

21.3–(22.7)–24.8

14.6–(17.0)–19.3

1.34

P (–SP)

3 (–4)

C. jalapensis

14.5

0.28

0.424–(0.670)–0.889

27.8–(28.6)–30.3

18.7–(19.7)–21.2

1.45

P (–SP)

3 (–4)

C. lindsayi

4.9

0.04

0.305–(0.395)–0.489

22.7–(23.7)–25.6

21.2–(22.0)–23.0

1.08

(SP–) S

3 (–4)

C. mitriformis

17.7

0.136

0.232–(0.440)–0.781

23.2–(24.7)–25.9

23.9–(25.8)–27.0

0.96

(SP–) S

3 (–4)

C. purpusii

15

0.103

0.297–(0.573)–1.240

18.7–(20.1)–22.1

20.9–(22.8)–24.7

0.88

S (–SO)

3 (–4)

C. rugosiceps

9.9

0.166

0.406–(0.773)–1.506

17.7–(18.6)–19.1

15.5–(16.8)–18.2

1.11

(SP–) S

3 (–4)

C. tasmanica

0.2

0.036

0.228–(0.284)–0.330

21.4–(23.2)–25.0

13.4–(15.0)–16.2

1.55

P (–SP)

3 (–4)

C. tinctoria

19.7

0.244

0.406–(0.654)–0.905

22.6–(24.3)–26.3

17.3–(18.1)–19.0

1.34

P (–SP)

3 (–4)

C. victoriana

0

0.005

0.054–(0.117)–0.166

21.7–(25.4)–28.6

16.6–(18.6)–21.0

1.37

P (–SP)

3 (–4)

C. woodsonii

0.6

0.031

0.254–(0.361)–0.490

17.3–(19.2)–21.3

14.5–(16.3)–17.7

1.17

SP (–S)

3 (–4)

Clade H

C. applanata

2

0.066

0.185–(0.311)–0.625

22.2–(23.2)–24.9

14.1–(14.9)–15.9

1.56

P (–S)

3 (–4)

C. chinensis

0

0

0

18.4–(19.6)–20.6

20.6–(21.8)–22.9

0.90

(SP–) S

3 (–4)

C. dentatasquamata*

1

0.076

0.201–(0.324)–0.661

19.3–(20.1)–21.8

13.6–(16.1)–17.9

1.25

SP (–S)

3 (–4)

C. potosina var. potosina

0.2

0.03

0.132–(0.204)–0.318

24.7–(26.6)–29.7

17.2–(18.1)–19.5

1.43

P (–SP)

3 (–4)

C. potosina var. globifera

0.1

0.013

0.229–(0.306)–0.430

24.4–(26.3)–28.9

17.9–(18.4)–18.9

1.47

P (–SP)

3 (–4)

C. sandwichiana

0

0

0

23.3–(24.9)–26.8

17.4–(19.8)–23.1

1.26

SP

3 (–4)

C. yucatana

0.1

0.012

0.124–(0.200)–0.287

17.3–(17.9)–18.6

10.5–(12.2)–13.5

1.47

P (–SP)

3 (–4)

Clade I

C. americana

2.9

0.036

0.221–(0.328)–0.523

17.1–(19.2)–21.0

14.4–(16.1)–16.6

1.19

SP (–S)

3 (–4)

C. cozumeliensis

16.4

0.151

0.241–(0.530)–0.924

14.4–(15.3)–16.5

18.1–(20.2)–21.9

0.76

(SP–) SO

3 (–4)

C. globulosa

1.2

0.05

0.356–(0.454)–0.574

24.2–(25.2)–27.5

18.7–(20.4)–23.1

1.24

SP

3 (–4)

C. macrocephala

17.7

0.319

0.519–(0.860)–1.398

23.1–(25.0)–26.6

15.5–(16.8)–17.7

1.49

P (–SP)

3 (–4)

Clade J

C. corymbosa var. corymbosa*

6.4

0.117

0.255–(0.419)–0.579

18.9–(19.5)–20.6

20.5–(21.5)–22.1

0.91

(SP–) S

3 (–4)

C. corymbosa var. grandiflora

0.4

0.043

0.305–(0.491)–0.761

15.5–(16.5)–18.1

19.5–(20.7)–22.1

0.80

(S–) SO

3 (–4)

C. corymbosa var. stylosa

17.7

0.239

0.591–(0.783)–0.983

22.2–(24.5)–27.3

16.1–(17.3)–18.5

1.42

P (–SP)

3 (–4)

C. prismatica

13.2

0.206

0.565–(0.964)–1.443

20.4–(22.3)–26.1

14.9–(17.5)–19.4

1.27

SP

3 (–4)

Clade K

C. boldinghii

0.4

0.006

0.262–(0.571)–0.737

14.9–(15.8)–17.2

17.7–(20.1)–21.3

0.79

(S–) SO

3 (–4)

C. chapalana

21.2

0.208

0.241–(0.530)–0.822

16.9–(17.6)–18.1

19.4–(21.9)–23.3

0.80

(S–) SO

3 (–4)

C. costaricensis

0.2

0.06

0.369–(0.386)–0.400

23.0–(25.4)–27.8

16.9–(17.9)–19.3

1.42

P (–SP)

3 (–4)

C. erosa

0.4

0.055

0.312–(0.357)–0.455

15.6–(16.8)–17.9

16.6–(18.0)–19.6

0.93

(SP–) S

3 (–4)

C. ortegana*

4.9

0.087

0.248–(0.470)–0.693

18.8–(19.7)–20.5

11.6–(12.9)–14.4

1.53

P (–SP)

3 (–4)

C. strobilacea

11.6

0.136

0.237–(0.530)–0.983

21.3–(23.7)–25.5

14.3–(16.4)–18.3

1.44

P (–S)

3 (–4)

Clade L

C. acuta

0.2

0.024

0.166–(0.237)–0.299

20.5–(22.3)–25.6

13.5–(15.9)–16.1

1.49

P (–S)

3 (–4)

C. desmouliniana

0.3

0.055

0.199–(0.357)–0.454

17.9–(19.5)–21.9

10.2–(13.2)–17.2

1.48

P

3 (–4)

C. hyalina

0

0

0

23.2–(24.9)–26.5

17.6–(18.5)–19.7

1.35

P (–SP)

3 (–4)

C. leptantha

5.4

0.143

0.282–(0.514)–0.816

21.7–(22.9)–23.6

14.5–(15.1)–15.5

1.52

P (–SP)

3 (–4)

C. odontolepis

3.5

0.073

0.279–(0.373)–0.488

25.2–(26.7)–29.8

15.0–(16.3)–18.3

1.67

P (–S)

3 (–4)

C. polyanthemos

15.4

0.114

0.116–(0.368)–0.708

24.6–(27.5)–31.4

18.8–(20.6)–22.9

1.33

P (–SP)

3 (–4)

C. tuberculata

0

0.02

0.299–(0.358)–0.418

15.5–(16.0)–16.5

19.7–(11.8)–12.6

1.36

P (–SP)

3 (–4)

C. umbellata var. umbellata

0

0

0

16.6–(18.2)–19.0

14.6–(15.9)–17.7

1.14

(P–) SP

3 (–4)

C. umbellata var. reflexa

0

0

0.149–(0.294)–0.431

16.9–(19.1)–21.5

12.4–(14.3)–16.5

1.34

P (–SP)

3 (–4)

Clade M

C. coryli

0

0

0.128–(0.173)–0.239

27.9–(29.4)–30.5

19.3–(20.9)–23.2

1.41

P (–SP)

3 (–4)

C. indecora var. indecora

0

0.023

0.366–(0.490)–0.586

26.8–(27.7)–29.8

20.7–(22.5)–24.1

1.23

(P–) SP

3 (–4)

C. indecora var. attenuata

0

0

0.225–(0.269)–0.296

25.9–(27.3)–28.9

17.2–(18.4)–19.0

1.48

P (–SP)

3 (–4)

C. indecora var. longisepala

0.2

0.054

0.226–(0.281)–0.376

26.0–(27.6)–30.3

13.8–(17.6)–21.2

1.57

P (–SP)

3 (–4)

C. indecora var. neuropetala

0

0.011

0.248–(0.322)–0.395

23.7–(27.7)–32.1

16.1–(19.0)–22.2

1.46

P (–SP)

3 (–4)

C. warneri

0

0

0.152–(0.203)–0.246

23.4–(24.7)–26.5

14.6–(15.1)–15.7

1.64

P (–SP)

3 (–4)

Clade N

C. choisiana*

0.3

0.073

0.177–(0.346)–0.422

19.7–(20.2)–20.5

16.4–(16.7)–17.0

1.21

SP

3 (–4)

C. deltoidea

0.6

0.041

0.182–(0.283)–0.460

15.2–(17.5)–19.8

13.8–(15.7)–17.2

1.11

(SP–) S

3 (–4)

C. gracillima

0

0

0.228–(0.277)–0.360

18.9–(19.8)–20.7

11.1–(13.7)–14.9

1.44

P (–SP)

3 (–4)

C. mcvaughii

0.1

0.008

0.197–(0.241)–0.298

18.9–(20.6)–22.0

15.8–(18.7)–21.4

1.10

(SP–) S

3 (–4)

C. punana

0

0

0

20.1–(21.9)–23.9

14.9–(16.2)–17.0

1.35

P (–SP)

3 (–4)

C. sidarum

0

0

0

16.8–(18.5)–20.0

13.3–(14.1)–15.1

1.31

(P–) SP

3 (–4)

C. vandevenderi

0

0

0

21.8–(23.9)–25.8

15.3–(16.1)–17.4

1.48

P

3 (–4)

Clade O

C. argentiniana

1.1

0.045

0.307–(0.329)–0.497

17.8–(19.5)–20.7

11.0–(12.4)–13.1

1.57

P (–SP)

4 (–4)

C. acutiloba*

0.8

0.037

0.221–(0.292)–0.395

21.1–(21.8)–22.6

15.5–(16.1)–17.1

1.35

P (–SP)

3 (–4)

C. bella*

0.7

0.112

0.371–(0.534)–0.784

21.2–(22.4)–24.1

22.6–(24.0)–25.1

0.93

S

3 (–4)

C. boliviana*

6.2

0.068

0.204–(0.325)–0.583

18.7–(21.2)–23.5

20.4–(23.2)–26.4

0.91

S

(5–6)–7

C. chilensis

19.8

0.311

0.443–(0.954)–1.776

21.4–(22.9)–25.9

17.9–(23.3)–26.9

0.98

(SP–) S

3 (–4)

C. cockerellii

4.9

0.319

0.930–(1.064)–1.224

20.8–(22.6)–24.0

17.0–(18.5)–21.6

1.22

SP

3 (–4)

C. cristata

6.3

0.086

0.231–(0.421)–0.669

15.9–(16.7)–17.4

14.3–(15.3)–15.9

1.09

S

3–4

C. flossdorfii var. pampagrandensis*

26.8

0.47

1.040–(1.469)–2.214

15.8–(19.9)–24.9

17.5–(20.5)–23.4

0.97

(SP–) S

(5–6)–7

C. foetida var. foetida

10.4

0.146

0.485–(0.707)–1.343

21.0–(23.2)–24.6

17.5–(18.4)–19.7

1.26

SP (–S)

3 (–4)

C. foetida var. pycnantha

12.1

0.217

0.478–(0.979)–1.513

21.4–(22.3)–23.6

22.4–(22.9)–23.2

0.97

S (–SO)

3 (–4)

C. friesii

0.1

0.028

0.158–(0.247)–0.405

17.8–(18.8)–20.1

14.7–(16.0)–17.3

1.17

SP

3 (–4)

C. globiflora

2.1

0.093

0.302–(0.456)–0.664

21.7–(23.1)–25.6

18.8–(22.0)–25.0

1.05

(SP–) S

6–8

C. goyazina*

4.5

0.118

0.294–(0.515)–0.773

15.3–(17.3)–19.0

17.5–(18.2)–20.1

0.95

(SP–) S

3 (–4)

C. grandiflora

0.3

0.05

0.371–(0.468)–0.564

15.1–(17.3)–18.9

15.2–(16.2)–18.4

1.07

(SP–) S

(4–) 5–6

C. killimanjari

21.9

0.212

0.203–(0.859)–1.615

17.6–(19.0)–21.0

18.5–(19.9)–21.6

0.95

(SP–) S

3 (–4)

C. microstyla

21.5

0.189

0.360–(0.702)–1.105

15.3–(17.7)–19.6

10.4–(13.0)–16.8

1.36

P (–SP)

3 (–4)

C. odorata var. odorata

9.6

0.184

0.395–(0.749)–1.190

20.9–(22.8)–25.4

15.3–(15.8)–16.2

1.44

P (–SP)

3 (–4)

C. orbiculata*

8.4

0.082

0.242–(0.433)–0.713

18.6–(19.8)–20.4

17.7–(19.4)–22.0

1.02

(SP–) S

3 (–4)

C. paitana

43.6

0.912

0.749–(1.619)–2.606

16.8–(18.8)–19.9

20.9–(22.7)–23.8

0.83

(S–) SO

3 (–4)

C. parodiana

34.3

0.903

1.076–(1.933)–2.917

19.0–(20.5)–22.3

21.5–(23.0)–24.0

0.89

(SP–) S

(4–) 5–6

C. purpurata

0

0

0

19.4–(21.1)–22.7

17.0–(18.3)–20.7

1.15

SP (–SO)

3 (–4)

C. rubella*

0.3

0.09

0.468–(0.505)–0.550

17.7–(21.0)–22.9

17.4–(17.9)–19.0

1.17

SP

3 (–4)

https://static-content.springer.com/image/art%3A10.1007%2Fs00606-009-0259-4/MediaObjects/606_2009_259_Fig3_HTML.gif
Fig. 3

Evolution of tectum perforation in Cuscuta. As a result of implementing the gap-weighting method (Thiele 1993), percent perforation is represented on a continuous scale, with light branches indicating an imperforate tectum and black branches indicating a reticulate tectum. For more information on clades A–O, see Stefanović et al. (2007). Tectum imperforatum is likely the ancestral condition, while pollen grains with increasingly larger perforation areas have evolved multiple times. The reticulated tectum has evolved in Cuscuta only in subg. Monogynella, and clade O of subg. Grammica

Pollen grains with smaller perforations (usually ≤1 μm) form a continuous transition from imperforate to microreticulate (Fig. 1e–l) with the former condition prevalent in Cuscuta (ca. 60% of species) and encountered in many Convolvulaceae (Sengupta 1972; Austin 1973a, 1973b; Tellería and Daners 2003). Because of the intergradations observed, the types of pollen previously recognized on the basis of perforation size in Cuscuta (e.g., Liao et al. 2005) or those derived from the currently accepted tectum categories (Punt et al. 2007) are arbitrary ranges of variation. If a separation of “types” is desirable for description purposes, the template based on the eight quantitative character states (Table 1; Fig. 1e–l) provides a better resolution. Increasingly larger tectum perforation areas have evolved in subg. Monogynella, and multiple times in subgenus Grammica (Fig. 3), but the advantage of this feature in Cuscuta is unclear.

The evolution of tectum in Cuscuta parallels that of early angiosperms which were inferred to have had an imperforate or microperforate tectum, with the reticulate condition evolving in the common ancestor of Austrobaileyales and “mesangiosperms” (e.g., all angiosperms other than the ANITA lines; Doyle 2005, 2008). Reticulate exine is common in angiosperms, and it was debated whether if it is associated or not with sporophytic self-incompatibility (Zavada 1984, 1990; Gibbs and Ferguson 1987). Unfortunately, very little is known about the breeding systems in Cuscuta (Costea et al. 2006a; Costea and Tardif 2006). Reticulate pollen was functionally linked to entomophily (e.g., Ferguson and Skvarla 1982; Hesse 2000), hydrophily (Cox 1988) or anemophily (e.g., Lisci et al. 1994; Tanaka et al. 2004), suggesting that this microarchitectural feature of pollen is not directly correlated with a certain pollination vector.

Pollen size and shape

Pollen size is polymorphic (Table 2; Online resource 1). However, the species of subg. Monogynella have the largest pollen grains, 25–37.2 μm long, while in the remaining subgenera the average is 21 μm (Table 2). Convolvulaceae pollen is usually at least twice as large, averaging between 50 and 80 μm (Sengupta 1972; Lewis 1971; Ferguson et al. 1977; Tellería and Daners 2003; Leite et al. 2005; Menemen and Jury 2002; Martin 2001). Humbertieae, which forms a sister lineage to the rest of Convolvulaceae, has also large pollen (50–80 μm) (Lienau et al. 1986). In contrast, small pollen grains were reported from Cardiochlamyeae (e.g., Cordisepalum, ca. 12 μm; Dinetus 12–18 μm, Tridynamia, 12–14 μm, Staples et al. 2009; Cardiochlamys and Poranopsis, 18–20 μm, Sengupta 1972) and Erycibe (28–39.6 μm, Rao and Lee 1970; Sengupta 1972), all inferred to have diverged earlier than Cuscuta (Stefanović et al. 2003). Small pollen grains are also known only from the “bifid clade” (Dicranostyloideae), in Dipteropeltis (12–14 μm, Staples et al. 2009), Hildebrandtia (28–32 μm, Staples et al. 2009), Dichondra (22–33 μm), some Cressa species (24–30 μm) (Tellería and Daners 2003), Dicranostyles (18–21.6 μm), Lysiostyles (21.6–25.2 μm) (Austin 1973b), and Metaporana (14–16 μm) (Staples et al. 2009). Based on this information, the polarity of this character is equivocal. If the tribe Humbertieae, currently comprising only one genus and species (Humbertia madagascariensis Lam.), would be considered a distinct family, Humbertiaceae (Pichon 1947), small pollen grains are likely the ancestral condition in both Cuscuta and Convolvulaceae.

Pollen size in Cuscuta may be associated with chromosome size, ploidy level, and nuclear genome size. Species with the largest pollen in the subg. Monogynella have also the largest chromosomes in the genus (6–23.1 μm in C. reflexa; Kaul and Bhan 1977) and among the highest estimates for the nuclear genome (44.93 pg/2C in C. lupuliformis; McNeal et al. 2007). Subgenus Grammica has typically the smallest chromosomes (typically ≤4 μm) and pollen grains, but some of its species with larger chromosomes (8–16 μm in C. indecora, Fogelberg 1938) and higher nuclear genome size (65.54 pg/2C McNeal et al. 2007) have also pollen grains approaching 30 μm in length. In this latter size category can also be included some Grammica species such as C. cephalanthi and C. campestris which are characterized by small but numerous chromosomes (2n = 60 and 2n = ca. 54, respectively; Fogelberg 1938; García and Castroviejo 2003) and higher genome sizes (10.83 pg/2C; McNeal et al. 2007). Similarly, C. epilinum, with 2n = 42 (García and Castroviejo 2003) and 7.74 pg/2C estimated genome size, has larger pollen grains than the other species of subg. Cuscuta with 2n = 14 (genome size is known only in C. europaea—2.15 pg/2C). However, a rigorous corroboration of this apparent correlation is not possible because only a few species have their karyotype known (reviewed by García and Castroviejo 2003) and their genome size estimated (McNeal et al. 2007). The presence of 5–8 zonocolpate pollen grains in Cuscuta is consistent with all three phylogenetic scenarios mentioned in the “Material and methods” section, but together with the small size, the overall morphology of pollen suggests for Cuscuta sisterhood either to the “bifid-style” clade (Dicranostyloideae) or to one of the members of this clade.

Shape of pollen is polymorphic in Cuscuta (Table 2; tree not shown). Although over 50% of species have prolate grains, shape varies greatly among species of the same clade, and to a certain extent within the same species in the same flower/anther (Table 2). Nevertheless, because pollen grains with an increased number of apertures (see above) tend to be associated with spheroidal or subspheroidal shapes, the prolate or perprolate shapes (P/E ratios >1.33; Table 2) are likely to be primitive, as suggested by Austin (1998) for Convolvulaceae in general.

Taxonomic significance of pollen characters in Cuscuta

Cuscuta is one of the most difficult parasitic groups taxonomically. The last comprehensive treatment of the genus was provided by Yuncker more than 75 years ago (Yuncker, 1932). Following Engelmann (1859), Yuncker (1932) proposed a classification with three subgenera (Cuscuta, Grammica, and Monogynella). While this arrangement has been largely confirmed by phylogenetic studies, the numerous sections and subsections created by Yuncker (especially in subg. Grammica) have been shown to be polyphyletic (García and Martin 2007; Stefanović et al. 2007). At the species level, the systematics of Cuscuta is currently undergoing major taxonomic revisions through studies aimed at understanding the evolutionary relationships, speciation, and biogeography by using various molecular, morphological, and micromorphological data (Costea et al. 2006a, b, c, d, 2008a; Costea and Stefanović 2009b).

It is clear that pollen characters alone are insufficient to reconstruct phylogenetic relationships within Cuscuta, but considering the overall morphological minimalism that characterizes the genus, the variation of pollen (Table 1) is important for future taxonomic revisions at the species level. In general, subg. Cuscuta and several of the 15 major clades of subg. Grammica exhibit little pollen variation. (e.g., clades A, B, C, E, H, L, and N, see Stefanović et al. 2007; Table 2). Nevertheless, even in groups such as clade A (C. californica complex, Costea et al. 2006d), clade B (C. pentagona complex, Costea et al. 2006b), and clade L (C. gracillima complex, Costea et al. 2008a), basic pollen characters such as size, shape, and diameter of puncta/perforations have already been used together with other characters to separate species. Pollen will play an increasingly significant role in the subg. Monogynella and many of the Grammica clades (e.g., D, F, G, I, K, and O, see Stefanović et al. 2007) which exhibit significantly more palynological diversity (Table 2). These infrageneric groups are the least known in the genus, and their future species-level taxonomic revisions will benefit enormously from these additional pollen characters. For example, clade O (subg. Grammica), comprising over 20 species distributed mostly in South America (but also 1 in Africa), is perhaps the most diverse and challenging in Cuscuta (Stefanović et al. 2007). Pollen is equally varied in this group, encompassing practically almost entirely the variation documented in the genus (Table 2). While most species are tricolpate, C. boliviana, C. grandiflora, and C. purpurata are 5-8-colpate. Tectum varies from imperforate in C. purpurata to reticulate in C. parodiniana and C. paitana. Additionally, size and shape can also be used to separate closely related species in this clade.

Conclusions

Placed in the context of the evolutionary history of pollen in Convolvulaceae (3-colpate → 5-6-zonocolpate → pantocolpate → pantoporate), the pollen of Cuscuta can be considered one step above the primitive because some species in two major lineages (subg. Monogynella and clade O in subg. Grammica) have evolved 5–8 zonocolpate pollen, and because pantocolpate grains, although rare, are present. Reticulate pollen has evolved two times in Cuscuta: in subg. Monogynella and clade O of subg. Grammica. The traditional, qualitative tectum “types” represent arbitrary ranges of variation, which in Cuscuta are better described quantitatively. Overall, the morphology of pollen supports Cuscuta as a sister either to the “bifid-style” clade (Dicranostyloideae) or to one of the members of this clade. Although the pollen characters are insufficient to reconstruct the phylogeny of the genus, pollen morphology is useful for the taxonomy at species level.

Acknowledgments

The authors warmly thank the curators and directors of AAU, ALTA, ARIZ, ASU, B, BAB, BOL, BRIT, CANB, CAS, CEN, CHR, CHSC, CIIDIR, CIMI, CTES, DAO, ENCB, F, G, GH, H, HUFU, IAC, IBUG, IEB, IND, J, JEPS, LL, LP, LPB, LPS, K, MEL, MERL, MEXU, MICH, MO, NMC, NU, NY, OKLA, OSU, OXF, PACA, PRE, QCNE, QFA, P, PACA, RB, RSA, SAM, S, SD, SGO, SI, SPF, TEX, TRT, UA, UB, UBC, UCR, UCT, UNB, UNM, UPRRP, UPS, US, USAS, VEN, WTU, and XAL for supplying plant material. We also thank Frédérique Guinel and two anonymous reviewers for their comments on an earlier version of the manuscript. Thierry Deroin provided the information about the pollen of Humbertia. This research was supported by Discovery grants from the Natural Sciences and Engineering Research Council (NSERC) of Canada to M. Costea (327013–06) and to S. Stefanović (326439–06).

Supplementary material

606_2009_259_MOESM1_ESM.jpg (3.9 mb)
Online resource 1: Pollen size optimized onto a summary consensus tree resulted from three molecular phylogenies of Cuscuta based on nuclear ITS and plastid trnL-F sequences (García and Martin 2007; Stefanović et al. 2007; Stefanović and Costea, personal communication). Pollen size is polymorphic but the species of subg. Monogynella have the largest pollen grains

Copyright information

© Springer-Verlag 2010