1

Plant Systematics & Evolution: June 1999, Vol. 217: 313-333 Intraspecific classification of melons (Cucumis melo L.) in view of their phenotypic and molecular variation

Asya Stepansky, Irina Kovalski and Rafael Perl-Treves+

Department of Life Sciences, Bar-Ilan University, Ramat Gan 52900, Israel

+ To whom correspondence should be addressed

Abstract: Cucumis melo L. (melon) genotypes differ widely in morphological and biochemical traits. Intra-specific classification of such variability has been difficult, and most taxonomists still rely on the work of NAUDIN (1859). A collection of 54 accessions representing diverse genotypes from 23 countries was surveyed. Morphological traits related to the vegetative and flowering stages and mature fruit morphology and quality parameters, e.g., taste, aroma, sugar composition and pH, were scored. These were used to construct a "botanical-morphological" dendrogram that generally reflected the classification of Cucumis melo into several horticultural varieties. DNA polymorphism among the accessions was assessed using the Inter-SSR-PCR and RAPD techniques that detected abundant DNA polymorphism among melon genotypes. Cluster analysis indicated that the largest divergence was between North American and European cantalupensis and inodorus cultivars as one group, and the more "exotic" varieties: conomon, chito, dudaim, agrestis and momordica, as a second group. The molecular phylogeny agreed, broadly, with the classification of melon into two subspecies, and did not contradict the division into "horticultural varieties". It was apparent, however, that the infra-specific division is rather loose, molecular variation being distributed continuously between and within cultivar groups. We suggest that despite the morphological diversity, separation between varietal-groups may be based on a too small number of genes to enable unambiguous infra-specific classification based on DNA diversity.

Introduction

Cucumis melo L. is an important horticultural crop across wide areas of the world. Within the genus

Cucumis, it belongs to the subgenus melo, having 2n=24 chromosomes. Great morphological variation

exists in fruit characteristics such as size, shape, colour and texture, taste and composition, and C. melo is

therefore considered the most diverse species of the genus Cucumis (KIRKBRIDE 1993; WHITAKER &

DAVIS 1962; JEFFREY 1980; BATES & ROBINSON, 1995). The species comprises feral, wild and

cultivated varieties, the latter including sweet "dessert" melons, as well as non-sweet forms that are

consumed raw, pickled or cooked.

The most ancient records on cultivated Cucumis melo (reviewed by PANGALO 1929) appear in

Egyptian mural paintings. Among the vegetables listed in the bible as being eaten by the Hebrews in

Egypt (Numbers 11.5) are the qishu'im, likely identified as non-sweet C. melo varieties, similar to var.

flexuosus or adzhur (M. KISLEV, Bar-Ilan University, personal commun.). Extensive records are also

found in ancient Chinese writings from about 2000 B.C. (WALTERS 1989) and Greek and Roman

documents from the first century BC. PANGALO (1929) maintained that sweet melon forms were not

2

known in the Roman period, and were imported from Persia or Caucasus by travellers, making their

appearance in Europe only around the 13th century.

The extensive variation found in C. melo has led botanists to propose intraspecific classification

schemes. KIRKBRIDE (1993) emphasises that such "horticultural types" should be treated under the

rules of cultivated plant nomenclature, and not as true botanical taxa. He nevertheless regarded the

subdivision of C. melo into two subspecies, ssp. melo and ssp. agrestis, proposed by GREBENSCIKOV

(1953), JEFFREY (1980) and ZOHARY (1983), as botanically meaningful. The morphological character proposed as a key for the two subspecies involves the hairs that cover the female hypanthium: ssp. melo

has pilose or lanate ovaries (i.e., spreading, usually long, hairs), while ssp. agrestis has sericeous ovaries

(appressed, usually very short hairs). According to the above authors, both subspecies include wild,

weedy forms, those of ssp. melo being synonymous with C.trigonus Boiss and C. callosus (Rottl.) Cong.

In 1753 Linn coined the genus name, Cucumis, and described five species of cultivated melons. These

were later united into a single species, Cucumis melo, by NAUDIN (1859), who developed a classification

scheme based on a live collection of 2000 specimens. He performed systematic crosses to ascertain his

species-assignments (rev. PANGALO 1929). Naudin divided melons into 10 varieties, and his work

remained the basis of all subsequent studies. PANGALO (1929) studied a live collection of 3000

specimens at the Vavilov Institute, and proposed a more sophisticated, multi-level taxonomy based on

the idea of homologous series: each of four C.melo varieties was subdivided into two homologous

subspecioids, cultus (cultivated types) and agrestis (wild ones), each being further divided into "types".

HAMMER & al. (1986) inherited a similar 3-level classification system from GREBENSCIKOV (1953) and

tried to simplify it: they grouped under subspecies agrestis two convarieties, the East-Asian conomon,

and wild-growing agrestis. The second subspecies, melo, was divided into 10 convarieties. The latter

classifications only partially overlap with other schemes (e.g. agrestis or inodorus seem to indicate

different types in different publications), and the detailed plant descriptions required to implement them

are not available. This has led MUNGER & ROBINSON (1991) to propose a further-simplified version

of Naudin's taxonomy, dividing C.melo into a single wild variety, C.melo var. agrestis, and six cultivated

ones: cantalupensis, inodorus, conomon, dudaim, flexuosus and momordica. Since the Munger and

Robinson classification scheme served in this study as a reference point, i.e. we compared it to our

cluster analysis results, it is listed briefly in the Materials and Methods section.

In the last years, sensitive DNA fingerprinting techniques have been used to resolve taxonomic

relationships, providing an objective and quantitative measure for genetic diversity between taxa, e.g.

among genera and species (SCHIERWATER 1995; CAMPOS & al. 1994; MILLAN & al. 1996). The sensitivity of the new methodologies also allows genotyping of varieties or cultivars within a species

(e.g., VIRK & al. 1995; FANG & ROOSE 1997; LEE & al. 1996). In this study we combined phenotypic scoring and DNA fingerprinting to address several open questions regarding melon classification. We

asked whether the above mentioned classifications reflect a biologically significant demarcation between

genotypes: is the botanical subdivision into two subspecies substantiated by molecular data? Can we use

3

morphological variation in certain key-characters to sort melons into discrete groups, that would also be

defined by molecular fingerprinting? Alternatively, do melons vary in a more continuous pattern, making

varietal grouping more arbitrary? Which varietal groups within the germplasm contain more genetic

variation?

With these questions in mind, a selection of 54 melon accessions from 23 different countries was

studied. We compared them at the botanical-horticultural level, and applied DNA fingerprinting to

describe the relationships amongst them. In order to evaluate an intraspecific classification system,

several representatives of each group must be included. We therefore sampled a few accessions from

each of Munger and Robinson's varietal groups, to enable a comparison between this popular

classification scheme and our own clustering results. Previous studies have applied molecular markers to

evaluate genetic diversity in the melon germplasm (SILBERSTEIN & al., in press; STAUB & al. 1997 and

references therein), but most of these included a very small number of representative cultivars. Other

studies (e.g., NEUHAUSEN 1992) included numerous accessions, but did not cover the more exotic

germplasm of C. melo, while a study of morphological and physiological diversity by HOSOKI & al.

(1990) was restricted to Far Eastern melons. In the present study, two molecular data sets were collected

using two different PCR-based methods, to exclude possible bias generated by a single technique.

Random amplified polymorphic DNA (RAPD) profiles are obtained using decamer-primers of arbitrary

sequence (WILLIAMS & al. 1993). Inter-simple-sequence-repeat (ISSR) PCR involves longer (16-18

nucleotides) primers encoding microsatellite elements that amplify DNA segments between

microsatellite repeats (GUPTA & al. 1994; ZIETKIEWICZ & al. 1994). The morphological and molecular

data were subjected to cluster analysis, and the implications to C. melo germplasm classification are

discussed.

Materials and Methods Growth of Plant material and its morphological description. The sources of seeds for this study are lister in Table 1. Three plants of each accession were sown in the spring in a field plot near Rehovot, Israel. Plants were drip-irrigated, fertilised and treated against pests and pathogens according to standard agronomic practice. Plants were described morphologically at the flowering stage, fruits were described and photographed when mature. The following traits were scored: stem thickness, density and hardness of stem hairs, leaf colour, male flower diameter, ovary shape, ovary pubescence, sex type, shape, colour and pubescence of immature fruit. Traits examined in the mature fruit included primary and secondary rind colours, rind pattern and textures; fruit size and shape, flesh colour, presence of external and internal aroma, taste (sour, bitter, sweet, insipid), colour change, abscission, splitting and softening at maturity, seed weight. Description followed the guidelines for cucurbit descriptors by ESQUINAS-ALCAZAR & GULIK (1983). At least three fruits from each accession were harvested when mature and taken to the laboratory for biochemical analysis. Classification of melons according to MUNGER & ROBINSON (1991). Following is a description of the seven melon varieties, condensed from MUNGER & ROBINSON (1991), using also descriptions by NAUDIN (1959), PANGALO(1929), GREBENSCIKOV (1953) and HAMMER & al. (1986). The accessions in this study were tentatively assigned to one of these varietal groups (Table 1). 1. C. melo var. agrestis: thin-stemmed, monoecious plants growing as weeds in African and Asian countries. Very small (

4

Andromonoecious flowering in most genotypes, hairy ovary. Includes dessert melon types such as Galia, Ananas, Charentais, "American shippers". 3. C. melo var. inodorus. Large-sized winter melons, with non-aromatic, non-climacteric and long-storing fruits, with thick, smooth or warty rind. Includes sweet dessert melons from Asia and Spain, such as Honeydew and Casaba type-cultivars. Usually andromonoecious, hairy ovary. 4. C. melo var. flexuosus. Fruits are very elongated, non-sweet, eaten immature as cucumbers. Found in the Middle East and Asia, where similar, less elongated types, adzhur and chate, have also been reported as ancient vegetable crops (HAMMER & al. 1986; PANGALO 1929). Usually monoecious. 5. C. melo var. conomon. Far-Eastern cultivars, where the smooth, white-fleshed, thin rinded fruits are eaten as pickles; includes also sweet, crisp fruits eaten with their rind. Andromonoecious vines bear dark, spiny leaves, sericeous ovaries. Corresponds to Naudin's var. acidulus. 6. C. melo var. chito and dudaim were described by Naudin, but grouped together by Munger and Robinson. The former was reportedly of American feral origin, with small plum-size, aromatic fruits used as pickles, monoecious vines and sericeous ovaries. The second is of Persian origin, andromonoecious, sericeous ovaries, bears small, aromatic, red or brown-striped fruits, grown as ornamentals in Oriental gardens. 7. C. melo var. momordica. A group added by MUNGER & ROBINSON (1991) to include Indian accessions with monoecious vines, sericeous ovaries and large, non-sweet fruits with thin rind that splits at maturity. Biochemical analyses of the fruits. Biochemical analyses of mature fruits included pH and total soluble solids (TSS) measurements of the juice, and HPLC determination of sucrose, glucose and fructose in ethanol extracts of the mesocarp, and were performed according to STEPANSKY & al. (in press).

RAPD and ISSR analyses. DNA was extracted from melon leaves according to DELLAPORTA & al. (1983). RAPD reactions were performed using random decamers (Operon Technologies, Alameda - primers OPA7,10,16,18; OPB6; OPC8; OPD7,8,11,13,20; OPL7; OPR2,10), according to WILLIAMS & al. (1993) using a Crocodile II (Appligene) thermocycler. The reaction products were subjected to electrophoresis on 1.5% TBE-agarose gels, stained with ethidium bromide and visualised under UV light. DNA fingerprinting by ISSR were according to the PCR protocol by GUPTA & al. (1994). The primers used, from the University of British Columbia, were: (TC)8C, (AG)8T, (GGGTG)3, (ATG)6, (AC)8YC, (GA)8YG, (TG)8G, (AC)8G, (AC)8T. Annealing temperature was 5 oC lower than the melting temperature, estimated as 2 oC for each

A/T, 4 oC for each G/C (BERGER & KIMMEL 1986). PCR products were separated through 1.8% agarose gels, stained and photographed. Band scoring and cluster analysis. DNA fragments were scored as present (1) or absent (0). In a few cases we differentiated between strong bands (2) and weak ones (1). Morphological and biochemical data were coded as discrete characters for cluster analysis, as specified in the Results. Two different algorithms were used: cluster analysis involving parsimony methods was performed using the PAUP program (SWOFFORD 1993) with different Heuristic search-options, and "majority rule" consensus trees were constructed from sets of shortest trees. The second approach to obtain shortest trees involved distance-based methods, using the program RESTDIST (the PHYLIP package, FELSENSTEIN 1993). The resulting pair-wise distance matrix was fed to the clustering programs NEIGHBOR and KITSCH, and the BOOTSTRAP and CONSENSE programs were used to test the consistency of the data (FELSENSTEIN 1993).

RESULTS

Morphological description and varietal assignment of C.melo accessions. In order to study the

phylogenetic relationships within Cucumis melo, an assembly of 54 accessions (Table 1), including cultivars,

landraces and wild or feral types from 23 countries, was grown and described. We have included

representatives of all the varieties described by MUNGER & ROBINSON (1991) and sampled a substantial

number of genotypes from Africa, Southern and Western Asia, and the Far East, i.e., the primary and

secondary centres of diversity of the species. We tried to assign the accessions to one of the above-

mentioned horticultural groups (Table 1). In the germplasm from I.P.K., Gatersleben, such assignment had

already been done by the Gene Bank curators. The assignment of some accessions to varietal groups was not

5

straight-forwards, because certain traits would not fit the typical descriptions found in the literature. For

example, some accessions that appeared to belong to var. cantalupensis had monoecious flowering (e.g.

AFG), while others would not dehisce at maturity (MEC). Some inodorus accessions did have (weak)

external aroma (CCA, CAS). The African accessions ZA1, ZM1 and ZM3 and the Indian landrace VEL may

be classified as agrestis, except for their fruits which were much larger than typical agrestis. In Tables 2 and

3 we present - out of a more detailed set - data on 2 vegetative and 5 reproductive traits scored at the

flowering stage and seeds, and 14 morphological and biochemical traits scored in mature fruits. Fruit

photographs of 49 of the varieties are presented in Fig. 1, while Fig. 2 exemplifies the different shapes and

pubescence states of the flower ovaries. Since each trait may only provide partial taxonomic information, we

tried to combine all the morphological and biochemical data and subject them to cluster analysis. We have

selected eight vegetative and floral traits and 18 fruit and seed traits (listed in the legend of Fig. 3). When

two traits appeared to be inter-dependent, e.g., TSS and sucrose content, or ovary shape and fruit shape, only

one was included. Since the data contained only seven quantitative traits, they were coded as discrete-

character state (as in HOSOKI & al., 1990) by dividing them into 3-5 discrete classes, and analysed together

with the qualitative traits as non-directional multistate characters by the PAUP program (SWOFFORD 1993).

The heuristic search yielded 100 shortest-trees, and a representative one is shown in Fig. 3A. A consensus

tree computed from all the shortest trees is shown in Fig. 3B.

The main topological features of these trees broadly reflect the horticultural classification into two

subspecies and 5-6 varietal types. The tree may be dissected into two large clusters, one that includes all the

large-fruited, sweet accessions of vars. cantalupensis and inodorus, of Western Asia and European origin,

and the other comprising all the non-sweet, more "exotic" types of African and Asiatic origin. Within the

"sweet" cluster, accessions are parted into sub-clusters. Some sub-clusters comprise cantalupensis

accessions (e.g. AFG, ITA, BNG, DHA), others mostly include inodorus genotypes (CCA, HCR, PDS,

WNT) - but no consistent dichotomic separation between the two varieties is maintained in the majority of

the trees. The "exotic" sub-tree includes a few well-differentiated sub-clusters, that are maintained also in the

consensus tree (Fig. 3B). One includes var. flexuosus (ACK, FLI, FLR etc.), another cluster includes small-

fruited agrestis genotypes (AGA, AGN, SNG etc.), and a third cluster comprises larger, non-sweet fruits

(larger agrestis ZM1, ZM3 and ZA1, var. momordica - USM, MOM). Genotypes belonging to the chito,

dudaim and conomon Oriental varieties form additional, inter-mixed sub-clusters. The tree topology reflects

traits that were used as "taxonomic keys": sweetness, aroma, fruit size and shape, sex type and ovary

pubescence. This implies that the rest of the traits scored do not "mix up" or contradict the classification

suggested by botanists, since the topology was maintained when the additional traits were included.

Cluster analysis of melon varieties based on DNA fingerprints. DNA fingerprints of 54 accessions were

generated using 14 RAPD primers (Fig. 4A is an example). Each primer amplified an average of 6.9 bands.

Following agarose-gel electrophoresis 97 bands were scored, ranging in size between 0.3 and 2 kbp; 70% of

these were polymorphic among the 54 accessions, and the brightest, clearly scorable bands were selected for

cluster analysis. Since we wished to compare the results obtained by two different fingerprinting methods,

we produced a second data set by running ISSR reactions (Fig. 4B) with 9 different microsatellite primers,

6

producing 5-16 bands per primer (average 9 bands/reaction). A total of 116 bands of 0.3-2 kbp were scored

among the 54 genotypes. About 90% of these bands were polymorphic among the genotypes, and the

brightest ones were selected for cluster analysis. Although technically similar to RAPD, ISSR amplify a

different genomic "compartment", i.e. regions that are enriched with specific microsatellite repeats. Each of

the two sets was analysed using two different computational methods, to obtain dendrograms that depict the

relationship among the accessions. The first method was parsimony analysis (using the PAUP program), in

which only phylogenetically informative characters are used to construct the shortest tree, by stepwise

addition of taxa and characters. The second approach involved distance-matrix programs: all characters are

used to calculate pair-wise distances among all taxa; these values are then used by algorithms that construct

the shortest tree that accounts for the observed distances. The programs RESTDIST and NEIGHBOR were

used for this purpose. Cluster analysis of each data set separately produced rather similar trees, so we

combined them into a single set of 86 ISSR and RAPD bands.

Subjecting these data to yielded 50 shortest-trees, 767 steps long; Fig. 4A depicts one of them. In this

dendrogram, Western-world cultivars of the inodorus and cantalupensis types ("Group I" may be separated

(by the 14-step long node) from the non-sweet types of the agrestis, conomon, momordica, dudaim, and

chito varieties ("Group II"). It is also apparent that, despite their morphological distinctiveness (see

dendrogram in Fig. 3), var. flexuosus accessions (FLN, FLI, FLR, FLX, ACK) did not cluster together, but

were dispersed among the sub-clusters of Group I, indicating that they are probably closer to the European

sweet types, than to the non-sweet types of Group II. Within Group I, cantalupensis cultivars tended to

cluster together (e.g., ITA, CHA, LYB, TM; SAL, KUV, IML), as did some of the inodorus genotypes

(CCA, PDS, CAS; BLA and MEA), but sub-clusters were inter-mixed, not allowing a clear-cut separation

between the two types. Only two Indian non-sweet variety (VEL, INB) clustered with Group I and seem to

occupy an intermediate position between the two groups.

Within Group II, a few sub-clusters can be identified. A rather tight cluster (as indicated by the long

internode, 11 steps, separating it from the rest of the tree) comprises the conomon cultivars (CON, COV,

COC, OGO, GIN) and two of the dudaim accessions (DUD and DUG). Chito, momordica and agrestis

accessions are dispersed among a few sub-clusters. The African accessions with medium-size fruits, ZA1,

ZM1 and ZM3 were set aside from the weedy, tiny-fruited agrestis accessions from Africa and Asia (AGR,

AGA, AGN, SEN, SNG, KAK).

The same data were analysed with the RESTDIST and NEIGHBOR programs (applying the Neighbor-Join

algorithm), and a shortest dendrogram based on distance matrix methods was produced (Fig. 4B). The tree is

similar, in its salient features, to the one obtained by PAUP. It similarly allows the division of melon

genotypes into the same two groups, and suggests rather similar sub-clusters within them. Subjecting the

data to Bootstrap resampling-replicates indicated that several nodes in such trees did persist in more than

50% of the replicates (Fig. 4B).

Discussion

7

Melon intraspecific classification based on phenotypic data. Using pre-assignment by gene-bank curators

(mainly from IPK, Gatersleben), and our own morphological description, we classified 54 melon accessions

to the horticultural types described by MUNGER & ROBINSON (1991). Such assignment was based on the

laconic descriptions available in the literature, and relied in fact on a small set of key-traits that are of

horticultural significance. When we performed cluster analysis based on a larger set of 28 morphological and

biochemical traits, including characters not previously used for classification, the subdivision into most of

the varietal groups persisted (Fig. 3). This indicates that the botanical-horticultural traits on which the

traditional classification relies were well-chosen: the other traits did not suggest any consistent, alternative

grouping of the germplasm.

Phylogenetic analysis of melons based on molecular data. It must be recognized that dendrograms

based on morphological data may be subjected to bias: independently scored traits may co-vary or even be

co-inherited (although we have tried to eliminate redundant morphological traits from our cluster analysis);

the coding of qualitative and quantitative traits as discrete characters may not represent biological reality;

selection of traits and division into character-states is rather arbitrary. Dendrograms based on molecular data

- which should represent neutral traits of simple inheritance - may provide a quantitative, more objective

measure of the relationships between taxa. Among the different molecular techniques that may be used,

RAPD fingerprinting (and probably also Inter-SSR PCR) can only provide a first approximation of such

relationships. The problem in applying random PCR-amplicons as phylogenetically informative traits lies in

the assumption that bands of similar size represent the same sequence. Such assumption is not valid for

many PCR bands, and this would introduce noise in such analysis. Homologous sequences from different

molecular lineages may also lead to misinterpretaion of data, as well as the occurrence of PCR bands that

are not genetically inherited (SCHIERWATER 1995; HALLDEN & al. 1996; STAUB & al. 1996). Checking

each polymorphism for Mendelian inheritance and using for cluster analysis only bands exhibiting the

expected segregation would solve the problem, but require considerable additional labour. We have however, proven, by testing a subset of our primers on a sample F2 progeny, that the majority of ISSR bands

that were polymorphic between cultivar Topmark and P.I. 414723, segregated at a 3:1 ratio (STEPANSKY,

KOVALSKI & PERL-TREVES, unpubl.). It will be of interest to implement, in future studies, a more

genetically solid technique such as SSR, RFLP,or direct sequencing of a gene fragment, to test the

conclusions of the present study.

The dendrograms obtained by using two independently derived molecular data-sets, and two clustering

algorithms (Fig. 4), were similar in their general topology. Interestingly, they did not contradict substantially

the topology of the phenotypic-traits tree, and the following conclusions are suggested:

1. A node separating the varieties cantalupensis + inodorus from the other types is consistent with both the

phenotypic and molecular data. The separation between these two sweet-fruited groups - cantalupensis, with

climacteric, aromatic fruits, and inodorus, with long-keeping, non-climacteric ones, is only partially

substantiated by both data sets. Some cantalupensis-like genotypes cluster together, as do some of the

8

inodorus lines, but the two groups are rather close. This was also observed in the smaller-scale surveys by

SILBERSTEIN & al. (in press), and by STAUB & al. (1997), and may indicate that the maturation-related

differences may be controlled by a small number of key-genes and do not represent extensive genetic

divergence.

2. The varieties agrestis, conomon, chito, dudaim and momordica are related to each other, and more distant

from the two "dessert melon" groups and from var. flexuosus. This feature of the tree may be interpreted in

favour of the proposed subdivision into two subspecies, ssp. melo and ssp. agrestis. We note that both the

molecular and phenotypic data would support the grouping, within ssp. agrestis, not only of var. conomon

and var. agrestis, as was proposed by GREBENCIKOV (1953), but also of vars. chito, dudaim and

momordica. We also note that the levels of DNA diversity within Group II are higher, as compared to

"Group I" (cantalupensis and inodorus): the tree-branches within the latter group are shorter (Fig. 5) and

reflect the relatively narrow genetic base of the germplasm commonly used by modern breeders, as

compared to the total diversity found in Cucumis melo.

The trait of ovary pubescence, suggested as a key for the definition of subspecies due to its neutrality, is

only in partial agreement with the molecular phylogeny. Out of 23 accessions in Group II, three agrestis

accessions (AGR, KAK, SNG) and one non-identified Oriental genotype (GUO) had lanate ovaries that

would be typical of ssp. melo; others (SEN, AGA) had ovaries of intermediate hairiness; conomon,

momordica, chito and dudaim accessions indeed had sericeous ovaries. Within Group I, nearly all accessions

had lanate ovaries, with very few exceptions (CCA, JPN). Among flexuosus-like accessions, both sericeous

and lanate ovaries were found. In conclusion, the trait apparently reflects phylogenetic reality, although its

use as a single dichotomic key would be problematic. Variation of this trait among agrestis and flexuosus

genotypes may in fact reflect their diversity of origin, well documented for the former group (PANGALO

1929), less expected in the latter.

3. Due to its peculiar morphology, var. flexuosus - (long-shaped "snake-", or "cucumber-melons"), formed a

tight cluster in the phenotypic-tree (Fig. 3), closer to the non-sweet varieties. However, according to the

molecular data, flexuosus accessions were dispersed among the branches, closer to the inodorus and

cantalupensis types. A dichotomic subdivision would therefore place them in ssp. melo, which fits with the

Near-Eastern origin of this ancient vegetable.

4. Another trait that divides the melon germplasm rather naturally into groups is sex-type, encoded by a

single locus, A (PITRAT 1991). In our survey, all the accessions designated as flexuosus, momordica,

agrestis and chito were monoecious. Accessions of var. dudaim (DUA and DUG) were andromonoecious,

while DUD, that is more similar to chito, was monoecious. Three groups - conomon, inodorus and

cantalupensis - had andromonoecious flowering, although they included a few accessions with monoecious

individuals (AFG, CCA, IRN). Our phylogenetic data may suggest that andromonoecy could have appeared

at least twice, in the conomon-dudaim, and in the "dessert-melon" lineages. In addition, monoecious lines

9

were produced in the last years from cantalupensis and inodorus breeding material to facilitate hybrid seed

production.

5. Varietal groups chito, dudaim and momordica are poorly resolved from each other, and from var.

agrestis both in the phenotypic and molecular dendrograms. On the other hand, the molecular phylogeny

identifies var. conomon as a group that is well distinct from the rest of the germplasm. This probably reflects

an ancient history of independent melon domestication in the Far-East, separate from the domestication

lineage leading to cantalupensis and inodorus melons. The conomon group is thus closer to the larger

agrestis cluster ("Group II"), and the occurrence of sweet fruited genotypes in all three groups - agrestis,

conomon, and "dessert melons" - probably suggests a scenario of multiple domestication events that

occurred in parallel in different places. The somewhat esoteric idea of "homologous series" occurring in

parallel branches of the melon dendrogram is, in a way, re-evoked by the appearance of the sweetness,

pubescence and monoecious traits in different tree-branches.

How distinct are melons varietal groups from each other? We should note that the distinction between

varietal groups as "infraspecific taxa" is not as strong or clear-cut, as compared to the distinction between

different species or genera (e.g., PERL-TREVES & al. 1985). This is reflected in the tree topology (Fig.

5A,B): terminal branches, leading to individual accessions are long, due to the sensitivity of the molecular

techniques, while those internodes that divide taxa into clusters are short. Subjecting the data to the

Bootstrap test (Fig. 5B) proved the same point: The most consistent nodes are the terminal ones, that group

2-3 accessions together, and not the more internal ones. This indicates that the data points (i.e., polymorphic

bands) that served to group together larger clusters of accessions are few, as compared to the numerous

bands that specify terminal branches. As stated above, certain aspects of the PCR technique may have

contributed noise to the cluster analysis, and the use of other techniques may improve intra-specific

resolution in future studies. On the other hand,

a more "continuous" distribution of genetic variation in the germplasm may indicate that the varietal

groups of melon may have formed over a relatively short time-span, which can be probed by only a small

proportion of the traits that are scored. Another, complementary explanation for such pattern of variation

relates to the fact that melon varieties did not evolve any reproductive barriers between them. Wild and feral

genotypes continue to grow, in many countries, in proximity of sweet or vegetable landraces, with which

they may freely hybridise. The occasional occurrence of sweet agrestis fruits may have resulted from such

exchange. Selection by breeders who have combined genetic material from different groups (modern

breeders have frequently hybridised inodorus and cantalupensis lines) may also result in "horizontal

transfer" between tree-branches, causing an apparent poor resolution of the molecular phylogenies.

In conclusion, the molecular data generally seems to support the taxonomic classifications of NAUDIN

(1859), PANGALO (1929) and JEFFREY (1980), albeit not in all details. The distinction between

intraspecific groups in the melon germplasm seems, at present, not a sharp one and should be more "safely"

considered as a horticultural, rather than botanical, classification, as stated by KIRKBRIDE (1993, p. 5):

10

"Those traits desired by man in his cultivated plants, when introduced into botanical taxonomy, produce a

haze of confusion around taxa and lead to a false sense of genetic distinctness... However, those same traits

are key to understanding the variation within cultivated plants and framing their systems of cultivated

nomenclature."

Acknowledgements

The authors express their gratitude to J.H. Kirkbride of USDA, Beltsville, for sending valuable reprints on melon taxonomy. We thank J. Felsenstein of the University of Washington, Seattle, for providing his versatile PHYLIP package and providing continuous, patient instructions for their use. We are grateful to H.M. Munger from Cornell University, K. Reitsma of the P.I. Station, Iowa, K. Hammer of I.P.K., Gatersleben, M. Gomez-Guillamon of Experimental Station La Majora, Y. Cohen of Bar-Ilan University, S. Niego and R. Herman from Zeraim Gedera Ltd., Israel, for providing seed samples for this study. We thank E. Galun and K. Nerson-Kanter for translation from German, and H.S. Paris and V. Gaba for a friendly review of the manuscript. This work was supported by Grant No. IS-2129-92 from BARD, the United States-Israel Binational Agricultural Research and Development Fund., and Israeli Gene Bank Grant No. 5036-1-96 from the Israeli Ministry of Science and Technology.

11

Code Origin Name Accession No. Seed

Donor Melon Variety (tentative)

ACK

Turkey

Acuk

PI 167057

2

flexuosus

AFG

Afghanistan

PI 125951

2

cantalupensis?

AGA

Afghanistan

CuM 146

1

agrestis

AGN

Nigeria

CuM 287

1

agrestis

AGR

Africa

3

agrestis

BAK

South Balkan

CuM 53

1

BLA

Spain

Blanco

C-199

5

inodorus

BNG

USA

Burpee's Netted Gem 4

cantalupensis

CAS

Spain

Rochet Pamal

4

inodorus, Casaba type

CCA

Spain

cc26 (Amarillo orange flesh)

C-446

5

inodorus, Casaba type

CHA

France

Charentais

6

cantalupensis, Charentais type

CHI

unclear

Chito

PI 140471

7

chito/ dudaim

CHT

India

Chito

PI 164320

2

chito

CON

Far East

line 85-893

7

conomon

COV

Vietnam

Kairyo Ogata Kogane Seumari

CuM 246

1

conomon

DHA

Israel

Dvash Haogen2

4

cantalupensis, HaOgen type

DUA

Afghanistan

CuM 254

1

dudaim

DUD

unclear

Dudaim

line 85-895

7

dudaim

DUG

Georgia

CuM 296

1

dudaim

END

Israel

En Dor

4

cantalupensis, Ananas type

FLI

India

CuM 227 (=VIR K2511)

1

flexuosus

FLN

India

CuM 225

1

flexuosus

FLR

Iraq

CuM 349

1

flexuosus

FLX

Lebanon

Faqus

4

flexuosus

GIN

Japan

Ginsen Makuwa (Silver Spring)

PI 420176

2

conomon

GUO

China

Gou Gua

PI 532829

2

HCR

Spain

Hilo Carrete

C-198

5

inodorus?

HON

USA

Honeydew

line 89A-15

7

inodorus

IML

Kazakhstan

Imlyskaja

PI 476342

2

cantalupensis

12

INB

India

PI 124112

2

IRN

Iran

PI 140632

2

cantalupensis

ITA

Italy

CuM 298

1

cantalupensis

JPN

Japan

PI 266947

2

cantalupensis

KAK

India

Kakri

PI 164493

2

agrestis

KRK

Turkey

Kirkagac

PI 169305

2

inodorus

KUV

URSS

Kuvsinka

PI 506460

2

cantalupensis

LYB

Libya

CuM 294

1

cantalupensis

MEA

Spain

Melona Amarilla

C-193

5

inodorus

MEC

China

CuM 255

1

cantalupensis?

MOM

India

PI 414723

7

momordica

OGO

Japan

Ogon No.9

PI 266933

2

conomon

PDS

Spain

Pionet Piel de Sapo C-207

5

inodorus, Casaba type

SAF

Afghanistan

Safed Sard

PI 116915

2

SAL

Ukraine

Salgirskaja

PI 506459

2

cantalupensis

SEN

Senegal

G 22841

PI 436532

2

agrestis

SNG

Senegal

G-22841

PI 436534

2

agrestis

SON

South Korea

Songwhan Charmi

PI 161375

2

conomon

TM

USA

Topmark

7

cantalupensis

USM

USA

PI 371795

2

momordica

VEL

India

Velleri

PI 164323

2

WNT

Turkey

Winter Type

PI 169329

2

inodorus

ZA1

Zambia

ZM/A 5317

PI 505599

2

agrestis?

ZM1

Zimbabwe

TGR 1843

PI 482429

2

agrestis?

ZM3

Zimbabwe

TGR 228

PI 482399

2

agrestis?

Table 1. Melon accessions used in this study. Accessions are ordered alphabetically. Data on country of origin and a tentative assignment to melon varieties are presented. Seed source codes: 1- Institut fr Pflanzengenetik und Kultur-pflanzenforschung, Gatersleben, Germany; 2- Plant Introduction Station, Ames, Iowa; 3- A.P.M. den Nijs, Wageningen, The Netherlands; 4- S. Niego and R. Herman, Zeraim Gedera Ltd., Israel; 5- M. Gomez-Guillamon, C.S.I.C. La Majora, Spain; 6- Y. Cohen, Bar-Ilan University, Israel; 7- H.M. Munger, Cornell University, U.S.A.

13

Code Seed

Weight1 Stem Thickness2

Male Flower Diameter3

Pubesc. Density4

Sex Type5

Ovary Shape6

Ovary Pubesc.7

ACK 36.0 1.6 9.3 0.7 24.7 0.3 2 m 3 2 AFG 32.3 0.7 10.5 1.0 32.5 0.9 2 m 3 2 AGA 6.1 0.2 3.2 0.2 20.5 0.0 1 m 2 3 AGN 6.3 0.3 3.7 0.4 14.5 0.4 2 m 2 1 AGR 4.8 0.1 5.5 0.4 20.2 1.7 2 m 2 2 BAK 46.8 2.4 11.5 0.4 27.5 1.3 3 am 1 2 BLA 40.5 4.0 10.3 0.9 26.0 0.5 1 am 2 3 BNG 29.3 0.9 8.0 0.5 22.7 1.7 2 am 1 2 CAS 47.5 1.6 10.5 1.1 31.7 1.0 3 am 2 2 CCA 42.2 1.5 9.7 0.7 38.3 2.2 2 am, m 1, 2 3 CHA 23.7 0.8 7.5 0.3 27.7 1.8 1 am 1 2 CHI 7.3 0.7 4.7 0.3 14.7 1.2 2 m 1 1 CHT 9.5 0.2 7.0 0.6 18.3 2.0 2 m 2 1 CON 13.3 0.9 5.7 0.2 27.0 1.3 1 am 1 1 COV 17.5 0.2 8.2 0.4 43.5 1.8 1 am 3 1 DHA 37.9 0.5 10.0 0.6 25.7 0.5 2 am 1 2 DUA 29.0 2.1 7.3 0.5 20.7 0.5 2 am 2 1 DUD 3.1 0.1 4.7 0.2 20.5 0.4 2 m 1 1 DUG 19.0 0.4 6.5 0.3 23.7 1.1 2 am 1 1 END 40.1 0.5 8.7 0.2 31.7 2.2 3 am 1 1 FLI 60.3 2.6 9.7 1.2 35.5 1.8 3 m 3 3 FLN 29.2 3.4 7.5 0.8 44.7 1.9 2 m 3 1 FLR 24.7 0.1 7.0 0.8 40.0 0.0 3 m 3 2 FLX 55.1 0.6 11.7 0.3 35.0 0.0 2 m 3 1 GIN 11.5 0.2 7.0 0.0 24.7 0.7 1 am 2 1 GUO 9.0 0.2 7.6 0.5 27.0 0.5 2 am 2 2 HCR 30.7 1.9 10.0 0.6 37.7 1.2 2 am 1 2 HON 39.3 1.7 9.3 0.8 29.0 1.8 3 am 1 2 IML 30.1 2.1 7.8 0.4 33.3 2.0 2 am 1 2 INB 23.7 1.4 7.0 0.5 34.5 1.8 1 m 2 1 IRN 40.9 2.4 8.3 0.4 29.0 1.5 2 am, m 2 2 ITA 33.5 0.7 9.8 1.0 25.3 0.3 1 m 2 2 JPN 24.9 1.3 12.0 0.7 21.3 0.3 2 am 1 1 KAK 5.3 0.3 5.0 0.0 13.5 1.1 2 m 2 2 KRK 41.3 1.2 9.3 0.9 29.0 0.5 3 am 1 2 KUV 44.0 0.6 9.5 0.3 29.8 0.0 2 am 1 2 LYB 28.2 0.7 8.8 0.7 21.8 1.3 1 am 2 2 MEA 45.1 1.8 8.0 0.6 23.7 0.5 2 am 1 2 MEC 40.3 0.7 11.3 0.7 26.7 1.4 1 am 2 2 MOM 20.1 0.8 n.d. 26.8 1.4 1-2 m 3 1 OGO 8.9 0.5 6.3 0.3 23.3 1.4 2 am 2 1 PDS 45.5 1.7 9.7 0.9 35.0 1.5 2 am 2 2 SAF 53.1 1.5 12.0 0.4 34.3 0.7 3 am 2 2 SAL 28.7 1.2 7.0 0.7 30.0 0.0 2 am 2 2 SEN 9.3 0.7 5.3 1.0 17.3 1.1 2 m 1 4 SNG 5.9 0.1 5.0 0.9 9.0 0.0 2 m 2 2 SON 9.1 0.5 6.3 0.3 26.8 1.3 2 am 2 1

14

TM 21.7 0.9 9.0 0.0 24.3 0.5 3 am 1 2 USM 17.1 0.9 7.0 0.4 16.7 0.3 1 m 3 1 VEL 18.3 0.9 9.3 0.7 27.0 2.0 2 m 2 1 WNT 42.1 2.0 9.0 0.9 34.5 2.4 3 am 1 2 ZA1 20.0 0.7 5.7 0.3 22.5 1.2 1 m 2 1 ZM1 15.9 0.9 6.0 0.7 22.7 1.2 1 m 2 1 ZM3 11.7 0.5 6.0 0.0 17.3 1.4 n.d. m 2 1

Table 2. Morphological characters scored in 54 melon accessions. 1 - Seed weight: average of 5 seeds from original GeneBank sample, in mg, and the Standard Error (S.E.). 2 - Stem thickness: measured in mm on fifth node of main stem, averaged from at least 3 plants, and the S.E. 3 - Flower size: diameter of male flowers, in mm, average and S.E. from 3 plants, 3-4 flowers per plant. 4 - Hair density: evaluated on the fifth node of main stem, 1- sparse, 2- medium, 3- dense. 5 - Sex type: m- monoecious, plant bears staminate and pistillate flowers; am - andromonoecious, with staminate and perfect flowers. 6 - Ovary shape: 1- short, (ratio of ovary length:width 1:1 or less); 2- intermediate (ratio approx. 2:1), 3- elongate (ratio higher than 3:1). 7- Ovary pubescence: 1- short and appressed hairs (sericeous ovary), 2- long and spreading hairs (lanate ovary), 3- short and spreading, 4- long and appressed.

15

16

Fig1a

17

Fig1b

18

Fig1c

19

Fig1d

20

Fig2

21

Fig4

22

Fig5

23

24

References BATES, D. M., ROBINSON, R. W., 1995: Cucumbers, melons and watermelons. - In SMARTT, J., SIMMONDS, N. W., (Eds.): Evolution of crop plants, 2nd edn., pp. 89-96 - Essex: Longman Scientific. BERGER, S. L., KIMMEL, A. R.,1986: DNA fingerprinting by oligonucleotide probes specific for simple repeats. - Hum. Genet. 74: 239-243. CAMPOS L. P., RAELSON J. V., GRANT W. F., 1994: Genome relationships among Lotus species based on random amplified polymorphic DNA (RAPD). - Theor. Appl. Genet. 88: 417-422. DELLAPORTA, S. L., WOOD, J., HICKS, J. B., 1983: A plant DNA minipreparation: version II. - Plant Mol. Biol. Reporter 1 :19-21. ESQUINAS-ALCAZAR, J. T., GULIK, P.J ., 1983: Genetic resources of Cucurbitaceae: a global report. Rome: IBPGR Secretariat. FANG, D. Q., ROOSE, M. L.,1997: Identification of closely related citrus cultivars with inter-simple sequence repeat markers. - Theor. Appl. Genet. 95: 408-417. FELSENSTEIN, J., 1993: PHYLIP (Phylogeny Inference Package), Version 3.5c, distributed by the author, Department of Genetics, University of Washington, Seattle. GREBENSCIKOV, I., 1953: Die entwicklung der melonsystematik. Kulturpflanze 1: 121-138. GUPTA, M., CHYI, Y. -S., ROMERO-SEVERSON, J., OWEN, J. L., 1994: Amplification of DNA markers from evolutionarily diverse genomes using single primers of simple-sequence repeats. - Theor. Appl. Genet., 89: 998-1006. HALLDEN, C., HANSEN, M., NILSSON, N. O., HJERDIN, A., SALL, T., 1996: Competition as a source of errors in RAPD analysis. - Theor. Appl. Genet. 93: 1185-1192. HAMMER, K., HANELT, P., PERRINO, P., 1986: Carosello and the taxonomy of Cucumis melo L. especially of its vegetable races. - Kulturpflanzen 34: 249-259. HOEY B. K., CROWE K. R., JONES V. M., POLANS N. O., 1996: A phylogenetic analysis of Pisum based on morphological characters, allozyme and RAPD markers. - Theor. Appl. Genet. 92 :92-100. JEFFREY C., 1980: A review of the Cucurbitaceae. Botanical Journal of the Linnean Society 81: 233-247. KIRKBRIDE, J.H., 1993: Biosystematic monograph of the genus Cucumis (Cucurbitaceae). North Carolina: Parkway Publishers. LEE S. J., SHIN J. S., PARK K. W., HONG Y. P., 1996: Detection of genetic diversity using RAPD-PCR and sugar analysis in watermelon (Citrullus lanatus (Thunb.) Mansf.) germplasm. - Theor. Appl. Genet. 92: 719-725. MALLICK, M. F. R., MASUI, M., 1986: Origin, distribution and taxonomy of melons. - Scientia Hort. 28: 251-261. MILLAN T., OSUNA, F., COBOS S., TORRES A. M., CUBERO J. I., 1996: Using RAPDs to study phylogenetic relationships in Rosa. - Theor. Appl. Genet. 92: 273-277. MUNGER, H. M., ROBINSON, R. W., 1991: Nomenclature of Cucumis melo L. - Cucurbit Genet. Coop. Reports 14: 43-44. NAUDIN, C. V., 1859: Essais d'une monographie des espces et des varits du genre Cucumis. - Ann. Sci. Nat. Bot. sr. 4, 11: 5-87.

25

NEUHAUSEN, S. L., 1992: Evaluation of restriction fragment length polymorphism in Cucumis melo. - Theor. Appl. Genet. 83: 379-384. PANGALO, K. J., 1929: Critical review of the main literature on the taxonomy, geography and origin of cultivated and partially wild melons. - Trudy Prikl. Bot. 23: 397-442 [In Russian, and translated into English for USDA by G. Saad in 1986]. PERL-TREVES, R., ZAMIR, D., NAVOT, N., GALUN, E., 1985: Phylogeny of Cucumis based on isozyme variability and its comparison with plastome Phylogeny. - Theor. Appl. Genet. 71: 430-436. PITRAT, M., 1991: Linkage groups in Cucumis melo L. - J. Hered. 82: 406-411. SCHIERWATER, B., 1995: Arbitrary amplified DNA in systematics and phylogenetics. - Electrophoresis 16: 1643-1647. SILBERSTEIN, L., KOVALSKI, I., HUANG, R., ANAGNOSTOU, K., KYLE, M. M., PERL-TREVES, R.: Molecular variation in Cucumis melo as revealed by RFLP and RAPD markers. Scientia Horticulturae, in press. STAUB, J., JEFFREY, B., POETTER, K., 1996: Sources of potential errors in the application of random amplified polymorphic DNA in cucumber. - HortScience 31: 262-266. - BOX, J., MEGLIC, V., HOREJSI, T. F., MCCREIGHT, J.D., 1997: Comparison of isozyme and random amplified polymorphic DNA data for determining intraspecific variation in Cucumis. - Genet. Res. Crop Evol. 44: 257 - 269.

Stepansky A., Kovalski I., Schaffer A.A., Perl-Treves R. (1999) Variation in sugar levels and invertase activity in mature fruit representing a broad spectrum of Cucumis melo genotypes. Genet Res. Crop Evol. 46: 53-62.

SWOFFORD, D. L., 1993: PAUP: Phylogenetic Analysis Using Parsimony, Version 3.1, Computer program distributed by the Illinois Natural History Survey, Campaign, Illinois. VIRK, P. S., FORD-LLOYD, B. V., JACKSON, M. T., NEWBURY, H. J., 1995: Use of RAPD for the study of diversity within plant germplasm collections. - Heredity 74:170-179. WALTERS, T. W., 1989: Historical overview on domesticated plants in China with special emphasis on the Cucurbitaceae. - Econ Bot 43: 297-313. WHITAKER, T. W., DAVIS, G. N., 1962: Cucurbits - botany, cultivation and utilization. New York: London Leonard Hill Ltd., Interscience Publishers Inc. WILLIAMS, J. G. K., HANAFEY, M. K., RAFALSKY, J. A., TINGEY, S. V., 1993: Genetic analysis using random amplified polymorphic DNA markers. - Methods in Enzymology 218 :704-741. ZIETKIEWICZ, E., RAFALSKI, A., LABUDA, D., 1994: Genome fingerprinting by simple sequence repeat (SSR)-anchore PCR amplification. - Genomics 20:176-183.

26

Figure Legends Figure 1. Fruits of 49 Cucumis melo accessions described in this study. Accessions are designated by codes according to Table 1. Size bars are in cm. Figure 2. Pistillate or hermaphrodite flowers of a few melon accessions. Lanate (pubescent) ovaries (a,b) characterize var. cantalupensis and var. inodorus. Intermediate pubescence was noted in certain flexuosus accessions (e). Sericeous ovaries, typical of var. conomon (c) and var. momordica accessions (d) are rarely found in cantalupensis or inodorus accessions (e.g. JPN, f).Ovary shape varies and is usually correlated with fruit shape. Figure 3. Cluster analysis of morphological and biochemical traits of 54 melon genotypes, performed by the PAUP software. Database included 8 vegetative, floral and seed traits, and 18 fruit traits. Traits were coded as discrete non-directional character states . Traits scored as qualitative included: sex type, ovary pubescence - length and orientation of hairs, density and hardness of stem pubescence, leaf colour. In mature fruits - primary colour of skin, presence of net, wrinkles, ribs or warts; skin design - presence of streaks or stripes, presence of spots or speckles; splitting, abscission, external aroma, internal aroma, flesh colour, fruit shape. The 7 quantitative traits in the data were coded by dividing them to classes or ranges as follows: Stem thickness: 1, 10 mm; 3, 5-9 mm. Staminate flower diameter: 1, 34 mm; 3, 21-33 mm. Fruit length: 0, 30 cm. Flesh pH: 1, 6. Fruit sucrose content: 0, 50 mg/gfw. Fruit glucose and fructose content: 0, 35 mg/gfw. Seed weight: 1, < 13 mg; 2, 13-40 mg; 3, >40 mg. A. A heuristic search was conducted using the random-addition sequence and TBR optimization options, resulting in 100 shortest trees of 290 steps. One of these trees is depicted. Numbers indicate the length (no. of steps) of each branch. B. Majority rule consensus tree computed from the 100 shortest trees obtained from the same data. Numbers indicate the percent proportion of the trees that contained the particular node. Figure 4. Fingerprinting patterns of melon accessions. A. RAPD profiles amplified by primer OPA18. B. Inter-SSR PCR profiles amplified by the primer (AC)8YC. Amplification products were separated in 1.8% agarose gels and stained with ethidium bromide. Size of molecular weight standards is indicated in kbp. Figure 5. Cluster analysis of molecular fingerprint data from 54 melon accessions. Data-base included 44 RAPD bands and 42 ISSR bands. A. A heuristic search was conducted by the PAUP software using TBR optimisation options, resulting in 50 shortest trees of 766 steps. One of these trees is depicted. Numbers indicate the length (no. of steps) of each branch. B. Cluster analysis of the same data performed by the PHYLIP package software. Kimuras genetic distances were computed between all pairs of accessions using the RESTDIST program. These served as input to the NEIGHBOR program, that constructed a shortest-tree using the Neighbor-Join option. The program SEQBOOT were then used to test individual nodes for consistency by random resampling of the data; multiple trees were constructed from 25 resampled data sets by the program NEIGHBOR, and a consensus tree was computed from these by CONSENSE. Nodes that withstood such test, i.e. that appeared in the majority of the tree replicates, are indicated, the figures showing the percent-proportion of replicates containing a specific node.