The systematic position of tribe Paropsieae,

in particular the Ancistrothyrsus,

and a key to the genera of

W.J.J.O. de Wilde

Summary

The members of tribe Passifloreae of the Passifloraceae are assumed to be originally all tendril-climbers.

have and the ramification They essentially axillary cymose inflorescences, vegetative occurs always through the bud. In tribe and the accessory Paropsieae (mostly or , no tendril-climbers)

inflorescences are axillary, most likely essentiallyracemose, and the vegetative ramification is mostly through the axillary bud.

The tendril-climbing neotropical genus Ancistrothyrsus appeared to belong to Passifloreae.

the tribe remains and Though Paropsieae to occupy an intermediate position between Passifloraceae

Flacourtiaceae, they can best be classified with the Passifloraceae.

A new key is proposed for the distinction of both families and the genera of Passifloraceae.

the of revision of the , and the treatment of the During preparation my genus

Passifloraceae for the Flora Malesiana, I became of course interested in the systematics of the of the tribe several the family, amongst others in position Paropsieae on which subject that this papers were recently published. Almost all authors agree tribe occupies a more

or less transitional place between Flacourtiaceae and Passifloraceae. These two families

are certainly more closely allied mutually than to others and show consequently also

phytochemical similarities (Hegnauer, 1969). The tribe first also was recognized by De Candolle (1828) in Passifloraceae; e.g. Perrier

de la Bathie (1945) and Fernandes (1958) included it in that family. By most British

authors, from Bentham & Flooker (1867) onwards to Hutchinson (1967), the genera it

contains without the of are accommodated in Passifloraceae but recognition a supra

-generic taxon. and For not very convincing anatomical reasons the different habit, viz. arboreous the tribe transferred versus climbing, was to the Flacourtiaceae by Harms (1893), an modern opinion followed until time by various German authors, e.g. Engler (1921), Gilg (1923), Melchior in Engler's Syllabus (1964), and Sleumer (1954). Recent anatomical and induced Sleumer palynological investigations, however, (1970) to refer the genus

Paropsia to Passifloraceae.

Since the of it has treatment Paropsieae by Gilg (1925) appeared that the genus Soyauxia from the with should be excluded tribe, whereas Hounea proved synonymous Paropsia.

Two viz. liana Peru newParopsiaceous genera have been added, Ancistrothyrsus, a from East Brazil and West (Harms, 1931, 1932) and Viridivia, a shrubby closely related to

Paropsia, from Rhodesia (Hemsley & Verdcourt, 1956). the At present tribe Paropsieae is composed of the following genera: Paropsia, Andro- small siphonia, Viridivia, Paropsiopsis, Smeathmannia, and Barteria, all shrubs or trees from the Old World, and Ancistrothyrsus, a liana from South America. 100 BLUMEA VOL. XIX, No. i, 1971

When included the tribe the in Flacourtiaceae, is distinguished mainly by presence of a

corona. Within Passifloraceae it is remarkable that the Paropsieae, with the exception of South the climbing American Ancistrothyrsus, are deviating by their arboreous habit;

the tribe are all tendril-climbers with a few the genera constituting Passifloreae herbaceous, Adenia erect derivatives in and Tryphostemma and a few -like outliers in the large

the related genus and genus Dilkea. Malesian been During myinvestigations in Adeniaand Passifloraceae I have impressed by

the uniform modeoframificationin Passifloraceae, namely through thesecond, serial, acces-

bud. Harms in Nat. Pfl. sory Already (1893, 1925) wrote (family diagnosis Fam., 1925, seriale p. 471): 'oberhalb der Ranke oder des Bliitenstandes eine Beiknospe, die zu einem

Laubspross oder durch Reduktion der Blatter zu einem Bliitenstand auswachsen kann'.

Searching for characteristics for the distinction of the tribe Paropsieae within the Passi-

floraceae better than 'shrubs or trees without tendrils' versus 'climbers, mostly provided

tendrils' I checked the of and with presence the serial accessory buds the corresponding of the mode of (vegetative) ramification in all the genera Passifloreae: Passiflora (inch

Tetrastylus), Dilkea, Mitostemma, Hollrungia, Tetrapathaea, Adenia, Tryphostemma, Cros- and sostemma (inch Schlechterina), Deidamia (inch Efulensia), the Paropsieae, and a large number of Flacourtiaceae. This revealed the following:

all 1) In Passifloraceae (Passifloreae & Paropsieae) an axillary and an accessory bud are Flacourtiaceae bud always present; in an accessory is either present (e.g. in Lindackeria, absent. Ryparosa, Hydnocarpus, many Casearias) or all is into 2) In Passifloreae the axillary bud either abortive, or developed a cymose sterile inflorescence (either tendril-bearing or not) or into a tendril; the second, serial, the accessory bud always provides for vegetative ramification, whereas in certain

A. species (e.g. , Tetrastylis ovalis, Adenia racemosa, globosa, Dilkea sp.)

this raceme-like i.e. shoot appears as an inflorescence, an inflorescences-bearing or The short-shoot. ultimate 'true' inflorescences, which are essentially cymose, never

from the bud. a—d. develop accessory Fig. 1 the the bud flower- 3) In Paropsieae axillary develops into a or inflorescences-bearing also for the ramification. The bud either abor- shoot; it provides vegetative accessory is

tive or rarely developed into a normal branch. In the flower- or inflorescences shoots the flowers inflorescences situated —h -bearing or apparently are axillary. Fig. I e

4) In the Flacourtiaceae the ramification generally resembles that in Paropsieae. The bud for the it axillary usually provides vegetative ramification, or may produce a racemose

inflorescence (raceme-like or thyrsoid e.g. in Homalium, Lindackeria, Casearia, Ryparosa,

Hydnocarpus, and others), or a flower-bearing short-shoot, sometimes reduced to fas- African cicles or even to a single flower (e.g. the Australian Streptothamnus). In the South

genus Kiggelaria the vegetative ramification seems to occur always through the accessory the bud, but axillary inflorescences are racemose, i.e. flower-bearing short-shoots. Fig. I e, k—p.

The difference in mode of ramification between the Passifloreae on the one hand, and

the Paropsieae and Flacourtiaceae on the other, leads me to the assumption that the inflores- in the the uni- cences two latter groups are likely essentially racemose. Whether simple or

few-flowered ultimate inflorescences, as found e.g. in Paropsia or various Flacourtiaceae

(e.g. Casearia, Hydnocarpus, etc.), are essentially reduced racemes or cymes remains, of

course is difficult to decide.

If my observations and considerations on the mode of ramification are correct (which

this rather needs additional checking on several genera of Flacourtiaceae), would mean a fundamental difference betweenPassifloreae and Paropsieae (together with Flacourtiaceae), W. J. J. O. db Wilde: The systematic position of the Paropsieae 101

Fig. 1. Mode of vegetativeramification and various inflorescences in a—d. Passifloreae, e—h. Paropsieae,

Essence in and e, k—p. some Flacourtiaceae; all diagrammatic. Accessory bud solid black, a. of ramification A. Passifloreae; b. Most and Adenias; c. Passiflora racemosa, Tetrastylis ovalis, Adenia racemosa,

P. Essence venenata; d. Passiflora arborea, spicata, Dilkea, Adenia globosa, A. fasciculata; e. of ramification in

and f. Paropsieae Flacourtiaceae; Paropsia, Paropsiopsis, Smeathmannia; g. Paropsia guineensis, Androsiphonia; h. k—n. Various of Bacteria; genera Flacourtiaceae, e.g. Homalium, Lindackeria, Ryparosa, Scolopia p.p., dichasial inflorescences in Laetia Laetia. Hydnocarpus, etc.; o. Phyllobotrieae; p. Axillary sect.

in addition the rather character used till of to only arbitrary up now 'climbing' versus This would 'arboreous'. carry even more weight if a third differential character would be 'inflorescences essentially cymose' against 'inflorescences predominantly racemose'.

To these to the few of test assumptions it appeared interesting investigate cases excep- in tional habit both groups.

In of the P. arborea and Passifloreae some species large genusPassiflora are arboreous, e.g. the consists P. spicata; genus Dilkea mainly of non-climbing shrubs of small trees; and in Adenia and Tryphostemma some erect non-climbing herbaceous species occur.

It appeared that the mode of ramification in all these species is essentially the same as in

of the Dilkea the 'normal' climbing Passifloras, as explained above. Checking genus rather hand. the gave some difficulties because of the poor material at Here situation in the ramificationis usually obscured by the fact that the true axillary bud is suppressed; in a few cases, however, an abortive tendril between leaf scar and inflorescence (i.e. a flower-

is other bearing short-shoot) revealed that thesituation exactly as in Passifloreae. Moreover,

Dilkea is closely related to Mitostemma, in which the situation is quite the same as in Passiflora.

The few scandent Flacourtiaceae, i.e. the Australian Streptothamnus and Berberidopsis No. 102 BLUMEA VOL. XIX, I, 1971

a—b. Fig. 2. Two remarkably different flowers ofAncistrothyrsus tessmannii Harms. Longitudinal section and separate androgynoecium of a mature flower from Ducke 24385, anthers non-apiculate, x 3. c.—d.

Ditto, from Ducke 35681, a small-flowered specimen; note the longly apiculate anthers, x 3. e. Infructes- cence in the axil of a leaf-scar; note the accessory bud; from Ducke 24384, x ½. Wilde: The the W. J. J. O. de systematic position of Paropsieae 103

from Chile, are no tendril-climbers and their ramification is similar to that of other Flacourtiaceae. the but ramified In erect, rarely Phyllobotryeae (Phyllobotryum, Phylloclinium, and Moquerysia), the racemose inflorescence, from the axillary bud, is concaulescent with

the midrib of the leaf, and the accessory bud here occupies the place of the axillary bud. The Flacourtiaceae which the inflorescences and which only examples in in are cymose in is sometimes of Laetia also an accessory bud present, seem to be the few species sect.

Laetia (fig. i p). The seemingly cymose inflorescences of some species of Hydnocarpus

are in fact of a basically racemose construction.

As the all the The to Paropsieae, are non-climbing except genus Ancistrothyrsus. deviating

horseshoe-shaped inflorescences in Barteria can be explained by assuming that the race- the the bottom the mose inflorescence [in which flowers develop from to top, and not

the f. and reverse as erroneously depicted bij Warburg (1893, p. 8, 2) again by in' the and Gilg (1925, p. 384) ] is 'locked between axis the petiole which are concaulescent;

the bracts below each flower inflorescence in which many suggest originally a compound each lateral raceme is reduced to a single flower.

The climbing habit in Ancistrothyrsus, unique in the Paropsieae, induced me to a closer of that described this the but examination genus. Harms (1931) genus in Paropsieae, in his discussion he pointed to the intermediate position of the tribe, between Passifloraceae and Flacourtiaceae. Furthermore he stated on Ancistrothyrsus: 'Oberhalb des Bliitenstandes

steht, wie bei Passiflora, eine winzige Beiknospe, die jedenfalls die Verzweigung besorgt'.

He also pointed to the tendril-bearing inflorescences, resembling those of Passiflora. In a later article den (1932, p. 599) Harms wrote: 'Wie bei Passifloraceae ist die Ranke der

primare Axelspross, und iiber ihr sitzt eine winzige Beiknospe, die die Verzweigung fort-

setzen soil'.

A re-examination of the flowers and the fruit of three collections from the Utrecht

Herbarium (Ducke 24384, 24383, 35681) revealed that Ancistrothyrsus must be referred to the Passifloreae; only the pubescence of corona is reminiscent of a Paropsia flower. It also appeared that in this genus specimens occur with highly differently shaped flowers, of flowers apparently hermaphroditic, but possibly functionally unisexual. Both types The smaller been are depicted in fig. 2. type of flowers has not described by Harms.

The findings on the mode of ramification in the Paropsieae strengthens affinity of this tribe with Flacourtiaceae. other the of For reasons, however, mainly presence a corona,

the — remains and also anatomical and palynological evidence, tribe which to occupy an

intermediate position — seems better be classified in Passifloraceae. Not unlikely, the

of tendril-climbers— which the fundamental divergence into a group in essentially axil- for the lary buds, through the mediation of the cymose axillary inflorescences, provided

— arboreous been necessary tendrils and an group must have an early event in evolution.

A tentative the of the key to genera Passifloraceae, also serving to distinguish Flacourtia-

and 1 ceae Passifloraceae, may run as follows ):

I. (Extra-staminal) coronaabsent. Shrubs or trees, rarely scandent, without tendrils. Inflorescences axillary, racemose or rarely cymose. Vegetative ramification mostly through the axillary bud. 5 to

FLACOURTIACEAE many PASSIFLORACEAE I. (Extra-staminal) corona present Shrubs Inflorescences these 2. or trees, never with tendrils. axillary, racemose or cymose (?); always

produced on inflorescences-bearing shoots which develop from the axillary bud. Vegetative ramifi-

As with Carania examined in *) compared Hutchinsons key and list (1967) the genus Chiov. (type Florence)

is a synonym of Tryphostemma; the ill-known genus Trichostephanus, and also the genera Physena and

be excluded from Soyauxia seem to Passifloraceae; Machadoa and Keramanthus are synonymous with Adenia. 104 BLUMEA VOL. XDC, No. i, 1971

Tribe cation through the axillary bud. Stamens 5 to many Paropsieae

with one broad Corona double. Stamens numerous Barteria 3. Style single, stigma. .... 3. Styles (2—)3—6.

4. Corona double Paropsiopsis

4. Corona single.

5. Stamens 5.

6. at base tube the Filaments connate the into a enveloping ovary . Androsiphonia

6. Filaments free Paropsia

J. Stamens 10 or more.

Viridivia 7. Stamens 10—16. Androgynophore present

Smeathmannia 7- Stamens c. 20 or more. Androgynophore absent

in 2. Mostly tendril-climbers. Inflorescences axillary, cymose; these along normal stems or often special raceme-like short-shoots from the bud. inflorescences-bearingshoots, or rarely in accessory Vegeta-

Stamens tive ramification always through the accessory bud. 4—10 Tribe Passifloreae

8. Androgynophore distinct, if not so than the gynophore distinct (much longer than the ovary).

9. Whole , incl. the tubiform laciniate corona hairy. Flowers 4-merous. Plant (leaves) with

scale-like glands Ancistrothyrsus

Corona 9. various, but not hairy itself. No scale-like glands. 10. Androgynophorelong. Hypanthium well-developed. Stigmas ± globular. Flowers mostly

5-merous Passiflora

10. Androgynophore short. Hypanthium small. Stigmas finely lobed.

11. Flowers 5-merous. Corona composed of two rows of filaments. Styles short Hollrungia

11. Flowers 4-merous. Corona a single row of filaments. Styles long. Flowers unisexual

Tetrapathaea

8. absent shorter than the Androgynophore or very short; gynophore ovary.

12. Leaves simple, though sometimes deeply incised.

13. Corona well-developed. Disk ring-shaped, or absent. Anthers rather broad, dorsifixed.

Flowers bisexual.

14. Style-arms 3 or 4, free or partly connate.

15. Filaments inserted on the hypanthium. 16. Hypanthium short, cup-shaped. Styles free Mitostemma 16. Hypanthium tubiform. Styles partly united Dilkea

15. Filaments inserted on the inner side of the inner corona. . Tryphostemma (incl. Schlechterina) Crossostemma 14. Style single, with a simple stigma Corona Disk of Anthers 13. ill-developed. mostly consisting 5 strap-shaped segments. narrow, basifixed. Flowers mostly unisexual Adenia

Deidamia 12. Leaves compound (incl. Efulensia)

I Dr. Acknowledgement. — feel highly indebted to H. O. Sleumer for discussion and advice on various points.

I herbarium of the Botanisch Museum am gratefulfor the loan of several specimens ofPassifloraceae from the of the University of Utrecht.

REFERENCES

D. HOOKER. Genera BENTHAM, G., & J. 1867. Plantarum 1: 807—816.

CANDOIXE, A. P. DE. 1828. Prod. 3; 321 —338.

A. ENGLER, 1921. In Engler & Drude, Die Vegetation der Erde 9, Pflanzenwelt Afr. 3, 2: 572 —575.

FERNANDAS, A. & R. 1958. Garcia de Orta 6, 2: 241 —262; ibid. 6, 4: 649—671.

GUG, E. 1925. Flacourtiaceae. In Engler & Prantl, Nat. Pfl. Fam. ed. 2, 21: 413—416 (Paropsieae).

FIARMS, H. 1893. Bot. Jahrb. 15: 548—632, tab. 21.

1925. Passifloraceae. In Engler & Prantl, Nat. Pfl. Fam. ed. 2, 21: 470—507.

I93I. Notizbl. Bot. Gart. Berlin-Dahlem 11, 102; 146—149, fig. 4.

Ibid. 1932. 11, 107: 598—600.

HEGNAUER, R. 1969. Chemotaxonomie der Pflanzen 5: 293 —298 (Passifloraceae).

HEMSIEY, J. H., & B. VERDCOURT. 1956. In Hooker, Ic. PL; t. 3555.

HUTCHINSON, J. 1967. The Genera of Flowering 2: 364—374 (Passifloraceae).

MELCHIOR, H. 1964. In Engler, Syllabus der Pflanzenfamilien ed. 12, 2: 322—341.

PERRIER DE LA BATHIE, H. 1945. Flore de Madagascar, Fam. 143, Passifloracees: 1—50.

SLEUMER, H. 1954. Flora Malesiana I, 5: I—106 (Hacourtiaceae), especially 5, 13—14.

1970. Bull. Jard. Bot. Bruxelles 40: 49—75. Nat. Fam. WARBURG, O. 1893. Flacourtiaceae. InEngler &Prantl, Pfl. ed. 1, III, 6a: 8, 25—27 (Paropsieae).