Flora of North America species comparison

Passiflora tarminiana is sparingly naturalized in the eastern San Francisco Bay area (F. Hrusa et al. 2002) and southward along the coast to San Luis Obispo County, in areas of minimal summer drought.

This species was recently described, and is commonly confused with Passiflora mollissima (Kunth) L. H. Bailey [now usually recognized as P. tripartita var. mollissima (Knuth) Holm-Nielsen & P. Jørgensen]. Many reports of P. mollissima in agricultural, horticultural, and weed-science literature actually apply to P. tarminiana. An attractive plant with large, edible fruits (T. Ulmer and J. M. MacDougal 2004), it is an extremely aggressive weed in Hawaii (A. M. La Rosa 1984, as P. mollissima) and other areas where it has been introduced in the Old World tropics and subtropics. The species is unlikely to become a widespread weed in the continental United States because it cannot survive frost nor occasional desiccation.

A similar, closely related species, Passiflora mixta Linnaeus f., is a rare escape in San Francisco, California; it can be distinguished from P. tarminiana by its angular young stems, persistent stipules (deciduous in P. tarminiana), and a floral tube 80–110 mm deep, 1.6–2.6 times the sepal length (1.3–1.6 times in P. tarminiana).

(Discussion copyrighted by Flora of North America; reprinted with permission.)

Genera 17, species ca. 750 (1 genus, 18 species in the flora).

Passifloraceae have been historically assumed to be closely related to Achariaceae, Caricaceae, Cucurbitaceae, Flacourtiaceae, Malesherbiaceae, and Turneraceae (A. Cronquist 1981); they are currently considered to be most closely related to the latter two families, seated within Malpighiales (D. E. Soltis et al. 2000; M. W. Chase et al. 2002; Angiosperm Phylogeny Group 2003; N. Korotkova et al. 2009). All three families have cyanogenic compounds, flowers with a hypanthium often not bearing the stamens, a floral corona (although uncommon in Turneraceae), tricarpellate ovaries with three elongate styles, parietal placentation, and five stamens (Cronquist; Angiosperm Phylogeny Group). Because of the close relationship of these three families, they have recently been treated as an expanded Passifloraceae (Chase et al.; Angiosperm Phylogeny Group). Passifloraceae are still considered relatively closely related

to members of the now ex-Flacourtiaceae, particularly Salicaceae in the broad sense (Chase et al.), which includes the apparently corona-bearing Abatia Ruiz & Pavón (coronal filaments in Abatia, unlike those in Passifloraceae, are clearly staminal appendages; see A. Bernhard 1999), although the members of Flacourtiaceae that have cyanogenic compounds are possibly more distantly related to Passifloraceae and are now placed in a greatly expanded Achariaceae (Chase et al.). The family description provided above represents Passifloraceae in the narrow sense; Malesherbiaceae and Turneraceae are not included.

Passifloraceae are divided into two tribes: Passifloreae de Candolle, containing genera with tendril-bearing vines and scandent shrubs (or rarely small trees without tendrils), and Paropsieae de Candolle, containing genera without tendrils that are shrubby or arborescent (W. J. J. O. de Wilde 1971). Paropsieae have been assigned to Flacourtiaceae or are seen as transitional between Flacourtiaceae and Passifloraceae (de Wilde). Morphological evidence supports the inclusion of Paropsieae within Passifloraceae (E. S. Ayensu and W. L. Stern 1964), further supported by recent molecular phylogenetic evidence (D. E. Soltis et al. 2000; M. W. Chase et al. 2002). Beyond this, intergeneric relationships are largely unknown within Passifloraceae. The tribes may not be distinct as currently defined and some genera are probably derived from within Passiflora (that is, Hollrungia K. Schumann and Tetrapathaea (de Candolle) Reichenbach; S. E. Krosnick and J. V. Freudenstein 2005).

The floral corona of Passifloraceae is of uncertain origin, considered derived from the perianth (V. Puri 1948) or consisting of staminodes (P. K. Endress 1996). It has an unusual pattern of development, inward from the outer and inner margins of a ringlike primordium; its development is delayed until well after the fertile stamens begin to form; and the arrangement and number of coronal filaments is not consistent with that of the fertile stamens. There is little support for the corona being derived from perianths or staminodes (A. Bernhard 1999).

(Discussion copyrighted by Flora of North America; reprinted with permission.)