Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Conspectus of Families

  1. 1.

    Cambium initials not storied; flowers with lobed or crenate, intra- or interstaminal nectary disk on top of ovary (but see Octomeles/Datiscaceae); anthers dorsifixed; [flowers obdiplostemonous; fruit a drupe, samara, or capsule; endosperm 0; cotyledons reduced or 0]. 4/34. Pantropical Anisophylleaceae

  • Cambium initials and secondary xylem and phloem storied; floral nectaries 0 (Octomeles excepted); anthers usually basifixed 2

  1. 2.

    Flowers hypogynous; ovule 1 per carpel, pendent; placentation apical; stylodia unbranched; stipules present, caducous; ellagitannins present 3

  • Flowers (hemi)epigynous; ovules usually many per carpel; placentation parietal; stylodia sometimes branched; stipules present or not; ellagitannins 0 4

  1. 3.

    Leaves opposite or whorled; stipules lateral, small, caducous; fertile stamens 10; carpels 5 or 10, each with a long, slender stylodium stigmatic over its entire surface; pollen 3-aperturate. 1/15+. Worldwide

    Coriariaceae

  • Leaves alternate; stipules intrapetiolar, caducous; fertile stamens 5; carpel 1(2); stylodium (stylodia) with capitate stigma(s); pollen 2-colporate. 1/6. Southwest Pacific region Corynocarpaceae

  1. 4.

    Tendril-bearing dioecious or less often monoecious climbers or trailers, rarely tendrils 0; young stems nearly always with 2 rings of bicollateral bundles; stamens 3–5, often 4 of them joined or connate in 2 pairs; gynoecium(1)3(–5)-carpellate, (semi)inferior; stylodia free or connate into a single style; fruit usually a soft- or hard-shelled berry; seeds flat; bitter cucurbitacins widespread. About 97/960, tropical, some reaching temperate regions Cucurbitaceae

  • Tendrils 0; bundles never bicollateral; fruit capsular or rarely (Begoniaceae) baccate; seeds not flat; seed coat with operculum; cucurbitacins absent, except for roots of Datisca5

  1. 5.

    Leaves simple with mostly large stipules, usually asymmetrical; monoecious, rarely dioecious perennials or rarely annuals or halfshrubs; placentation axile, sometimes parietal; seeds with collar cells arranged in transverse ring around operculum. 2n=16–156 (no clear base number recognisable). 2/ > 1,500. Tropical and subtropical regions of the World and temperate parts of Asia, but not in Australia Begoniaceae

  • Leaves estipulate, simple, lobed, pinnate or pinnatifid, not asymmetrical; (andro)dioecious trees or perennial herbs; placentation parietal; seeds without collar cells around operculum. n=11, c. 23. 3/4. E Mediterranean to SE Asia and Papuasia, and California, Baja California Datiscaceae

Recognition of the close relationship among the core families of Cucurbitales (Datiscaceae incl. Tetramelaceae, Begoniaceae, and Cucurbitaceae) dates back to the 19th century, although in the more recent pre-molecular era these families usually have been included in more comprehensive groupings named Violales or Parietales (for more details, see Matthews and Endress 2004, and Zhang et al. 2006). The addition of Coriariaceae, Corynocarpaceae, and Anisophylleaceae to the core Cucurbitales is an outcome of molecular studies (Chase et al. 1993; Swensen 1996; Setoguchi et al. 1999; Schwarzbach and Ricklefs 2000, among many others). The inclusion of Apodanthaceae, recently favoured by several authorities (e.g., Stevens 2001), is presently not supported (APG III; S.S. Renner, Oct. 2009).

In early molecular studies of the order, using the rbcL gene, these families were not fully resolved and topologies were often contradictory. Still, in recent multigene analyses covering also other orders, statistical support for the branches within Cucurbitales is generally lower than in other angiosperm clades (e.g., Wang et al. 2009). Nevertheless, the analysis of nine loci from three genomes of all Cucurbitales families by Zhang et al. (2006) has resolved Cucurbitales as monophyletic and served as a basis for an understanding of morphological and sexual system evolution within the order, but did not resolve the relationships among all families (Fig. 2). Fagales are now generally viewed as the closest relatives of Cucurbitales; both orders share the essentially unisexual and epigynous flowers. The strongly supported multigene analysis of Wang et al. (2009) has recovered the relationship Fabales[Rosales[Cucurbitales + Fagales]].

Fig. 2.
figure 1_2

Phylogenetic relationships of Cucurbitales families, based on the multigene sequence analysis of Zhang et al. (2006)

Full size image

Anisophylleaceae, formerly included in Rhizophoraceae, are firmly established as the sister group to all remaining Cucurbitales, from which they differ significantly in reproductive and vegetative morphology (Schwarzbach and Ricklefs 2000). Matthews et al. (2001) and Matthews and Endress (2004) have pointed to similarities in floral structure that exist between Anisophylleaceae and Cunoniaceae, but at the same time have also revealed morphological traits in common between Anisophylleaceae and core Cucurbitales, such as unisexual flowers and inferior ovaries. Anisophylleaceae share with other Cucurbitales some anatomical characters of the wood, such as nonbordered or minimally bordered perforation plates and wide rays not accompanied by uniseriate rays, traits that are conservative and less likely affected by ecology. Anisophylleaceae have retained, however, characters that are more conservative than those in the other families of the order, such as absence of storying, presence of tracheids, and heterogeneous rays (Carlquist and Miller 2001). Thus, it appears that this family is correctly placed in Cucurbitales, and that its similarities with Cunoniaceae are due to convergence.

Among the remaining families, Coriariaceae and Corynocarpaceae stand out with 1-ovulate carpels, apical placentation, and superior ovaries, the latter trait, in view of the topology of Zhang et al. (2006), certainly derived. Cucurbitaceae, Datiscaceae s.l. (i.e., including Tetramelaceae), and Begoniaceae have epigynous flowers (as do Anisophylleaceae), essentially basifixed introrse (or latrorse) anthers, trimerous gynoecia, bifurcate carpels, and a peculiar extended neck over the roof of the ovaries or instead (in Begonia and many Cucurbitaceae) a narrow neck at this site (Matthews and Endress 2004). It is notable that Cucurbitaceae share with Coriaria and Corynocarpus a rare combination of wood anatomical traits (vertical parenchyma scanty vasicentric, banded, and ray adjacent, and rays with upright cells strongly predominant; Carlquist and Miller 2001). In the molecular topology, Cucurbitaceae place as sister to Datiscaceae and Begoniaceae, but the precise relationship between the latter remains unresolved.

In view of the amount of morphological differentiation both in Cucurbitaceae and in Begoniaceae, the difference in the numbers of genera recognised in the two families is surprising, if not paradoxical. By the middle of the nineteenth century, the development of taxonomic concepts in both families had reached a comparable level. Further development in Cucurbitaceae led to a steady consolidation of taxonomic concepts, and until the present, the family has remained a field of dynamic systematics activities. Begoniaceae, in contrast, never recovered from A. de Candolle’s degradation of Klotzsch’s 41 genera to sections, in which he has been followed by all students of the family to the present day. Although in principle Klotzsch’s concept survives in the sectional classification of the family, this never has attracted much interest by botanists (for a notable exception, see Doorenbos et al. 1998); instead, they sometimes resorted to an alphabetic sequence of the 1,400 species of Begonia, and the family became a field mainly of floristic, rather than systematics activity. It is true that the decisive differences among begonias are difficult to observe and put into words, many of them being included in the unpopular area of inflorescence morphology. Nevertheless, I am convinced that Klotzsch’s generic concepts would have been further developed had his genera not disappeared out of the focus of botanists through their degradation to sections.