See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/259496618

Notes on some endemic Cuban species of Ruelliinae (Acanthaceae), on their

seeds, pollen morphology and hygroscopic features

Article  in  Willdenowia · December 2010

DOI: 10.3372/wi.40.40210

CITATIONS READS

5 96

2 authors:

Werner Greuter Rosa Rankin Rodríguez

Università degli Studi di Palermo University of Havana
387 PUBLICATIONS   5,407 CITATIONS    57 PUBLICATIONS   79 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Phytotaxonomic investigation of the Mediterranean Area View project

Completion of the Euro+Med Plantbase View project

All content following this page was uploaded by Werner Greuter on 14 February 2015.

The user has requested enhancement of the downloaded file.

Willdenowia 40 – 2010 285

WERNER GREUTER1 & ROSA RANKIN RODRíguez2

Notes on some endemic Cuban species of Ruelliinae (Acanthaceae), on their seeds,

pollen morphology and hygroscopic features

Abstract
Greuter W. & Rankin Rodríguez R.: Notes on some endemic Cuban species of Ruelliinae (Acanthaceae), on their
seeds, pollen morphology and hygroscopic features. – Willdenowia 40: 285 – 304. – Online ISSN 1868-6397; © 2010
Berlin-Dahlem.
doi:10.3372/wi.40.40210 (available via http://dx.doi.org/)
Ruellia shaferiana, a seldom collected and incompletely known species, is described in detail and compared with
two vegetatively similar Cuban endemics, Dyschoriste bayatensis and the recently described Hygrophila urquiolae,
and with the two Cuban species of Apassalus (A. cubensis and A. parvulus). All are placed by Scotland and Vollesen,
in their classification of Acanthaceae, in the newly circumscribed subtribe Ruelliinae. Contrary to Dyschoriste, the
other taxa studied have anther thecae that lack basal awns. Absence of these awns is the single character distinguish-
ing Apassalus from Dyschoriste, and it has been suggested that both be combined. Our results lend support to that
suggestion and prompt the transfer of A. parvulus to Dyschoriste. Study of pollen by SEM confirms the generic
placement of the species studied, again associating Apassalus with Dyschoriste. Seed indumentum was studied by
LM and REM, complementing earlier studies of the remarkable hygroscopic mucilaginous hairs, which apparently
characterise the subtribe. Hygrochastic capsule dehiscence was detected in all four species. Pre-dawn anthesis and
flowers withering by noon were observed in R. shaferiana. H. urquiolae, only known from a small area in W Cuba, is
newly reported from E Cuba, and conversely, W Cuban occurrences of R. shaferiana are put on record. For three of
the names lectotypes are designated, as Urban’s original material was destroyed in Berlin in World War II.
Additional key words: Apassalus, Dyschoriste bayatensis, Dyschoriste parvula, Hygrophila urquiolae, Ruellia shafe­
riana, mucilage, bionics, hygrochasy, pre-dawn anthesis

Introduction
When describing a new Cuban species of Hygrophila R. Br., Our paper will focus on the study of some scarcely
H. urquiolae Greuter & al. (2010), we pointed out that collected, presumably rare, Cuban endemic species: the
tropical and subtropical wetland habitats, those of Cuba three already mentioned plus two currently placed in
in particular, had been relatively neglected by botanists. Apassalus Kobuski: A. cubensis (Urb.) Kobuski and A.
We might have added that this is particularly true of parvulus Alain & Leonard. They are currently placed in
Acanthaceae Juss. Indeed, we were then already work- four different genera, but it is debatable whether Apas­
ing on another member of this family that appeared to salus is distinct or should rather, as has been suggested,
present special interest. We had first collected that plant be merged in Dyschoriste Nees. By our observations we
in 2000, and earlier this year (February 2010) we were hope, among other things, to help answering that ques-
able to study it again in the living state. It turned out to tion. Among the aspects studied are seed indumentum,
be Ruellia shaferiana Urb., an incompletely known and pollen morphology and capsule dehiscence.
seldom collected Cuban endemic, superficially similar to
Dyschoriste bayatensis (Urb.) Urb. and of which we are Seeds. — Seed hairs of Acanthaceae were first described
offering a new, detailed description here. in considerable detail by Kippist (1842) and again stud-

1 Botanischer Garten und Botanisches Museum Berlin-Dahlem, Freie Universität Berlin, Königin-Luise-Str. 6–8, 14195 Berlin,
Germany; e-mail: [email protected]
2 Jardín Botánico Nacional, Universidad de la Habana, Carretera del Rocío km 3½, Calabazar, C.P. 19230, La Habana, Cuba;
­e-mail: [email protected]
286 Greuter & Rankin: Notes on endemic species of Ruelliinae

ied by Schaffnit (1906) and Oehm (1932), who, taken and Acantheae Dumort. (Blepharis Juss., Crossandra
together, provide a pretty good coverage of the family. In Salisb.), it is the apical cap that has pectinised cell walls
recent studies these structures were by and large neglect- that lose cohesion when wetted. Subsequent studies in
ed, even though they appear to be of taxonomic value, as Acanthaceae have paid scant attention to the conditions,
we hope to demonstrate. The seed hairs found in many wet or dry, triggering capsule dehiscence.
representatives of the Ruelliinae Nees can be character-
ised as follows. They are slender, very numerous, cov- Systematics. — Due to recent, including molecular, stud-
ering the whole seed or only its marginal portion, and ies, the classi­fication of Acanthaceae is undergoing con-
almost invariably present spiral or annular wall thicken- siderable change. The traditional subdivision of the fami-
ings, or both. In the dry seed they cannot be discerned as ly goes back to the pioneer work of Nees (1847), modified
discrete structures even under the microscope, appear- by Bentham (1876) then Lindau (1895), and is artificial to
ing to be glued down as by a varnish to form a compact a considerable degree. Bremekamp (1965), whose main
layer on the testa. Instantly upon contact with water the experience was with the palaeotropical groups, proposed
hairs are set free from their former confinement, to form a new classification, which, however, failed to gain gen-
a dense spreading pelt. At the same time a mucilaginous eral acceptance. For our purpose we adopt the recent re-
substance is released, which upon drying will attach the arrangement by Scotland & Vollesen (2000) as the best
seed firmly to the substratum. All species studied here currently available system of classification. According
present such mucilage hairs (Fig. 2 – 3). to it, the 23 Acanthaceae genera recognised for Cuba by
Alain (1957), reduced to 21 under Scotland & Vollesen’s
Pollen morphology. — Acanthaceae pollen is extremely (2000) criteria (they include Phidiasia Urb. in Odontone­
diverse and the morphology of pollen grains has been ma Nees and Drejerella Lindau in Justicia L.), represent
used for the purpose of classification ever since Radlko­ all recognised subdivisions of Acanthaceae: the three
fer’s (1883) and Lindau’s (1893) seminal work. Raj’s subfamilies, but very unevenly (Thunbergioideae Kostel.
(1961) extensive study of Acanthaceae pollen makes by Thunbergia Retz., widely naturalised; Nelsonioideae
reference to two of the species here treated. More re- Lindl. by Elytraria Michx.; Acanthoideae Kostel. by the
cently, scanning electron microscopy (SEM) has much remaining 19 genera); both tribes of the latter (Acantheae
improved our knowledge of the subject, as evidenced by with Stenandrium Nees; Ruellieae Dumort., 18 genera);
excellent photographs in many recent papers (e.g. Da­ and the four Ruellieae subtribes: Andrographinae Nees
niel 1998; Scotland & Vollesen 2000). with Andrographis Nees; Barleriinae with Teliostachya
Nees (now Lepidagathis Willd.) and Barleriola Oerst.;
Hygroscopic features. — The most visible function of hy- Justiciinae Nees and Ruelliinae with eight and seven gen-
groscopy in Acanthaceae (only the Acanthoideae, not the era, respectively.
two other subfamilies) is expressed in their explosive cap- The subtribe Ruelliinae, with which this paper is con-
sules, which, assisted by the peculiar structures known as cerned, is characterised by contorted aestivation of the
retinacula (or “jaculators”, as proposed by Lindau (1895) corolla lobes and by the four didynamous stamens be-
for semantic reasons; Sell (1969) prefers the neutral term ing connate pairwise. It is by and large congruent with
“hooks”), disseminate the seeds in the surroundings of the Lindau’s (1895) unranked subdivision “Retortae”, com-
mother plant. One tends to assume that the capsules tear prising six of its seven tribes (Barlerieae Nees being ex-
open when drying (xerochasy), but Bremekamp (1926) cluded). The seven Ruelliinae genera present on Cuba
came to the contrary conclusion, that acanthoid capsules are: Bravaisia DC., Blechum P. Browne, Apassalus, Dys­
open hygrochastically. He discovered that hygrochasy choriste, Ruellia, Hygrophila and the naturalised Eran­
is so to say a phenomenon superimposed on xerochasy, themum L. We will focus on the four to which our target
insofar as first a tension must build up in the fruit as a species belong, which we shall briefly introduce.
result of desiccation, only then can wetting of the capsule
beak, which holds the capsule halves together, trigger the Dyschoriste is a pantropical genus of perhaps 50 or 60
release. Bremekamp was wrong, however, when he gen- species (Wasshausen & Wood 2004; Furness 1995).
eralised his findings to the whole family (according to The New World representatives were monographed by
Sell (1969), he had been studying a single species, Ru­ Kobuski (1928b), who recognised 40 species, to which
ellia tuberosa L.). Sell proved that xerochastic capsules those from Asia and Africa must be added. The generic
are the usual condition in Acanthaceae, from which hy- name is typified by an Old World element. If the Ameri-
grochasy has evolved at least twice and probably repeat- can species were treated as generically distinct (which
edly and/or reversibly. In hygrochastic Ruelliinae species no one presently does), they would have to be placed in
(some of Ruellia L., Hygrophila and Phaulopsis Willd.), Calophanes D. Don. The genus is characterised by the
an apical cap of lignified cells clamps the capsule halves presence of a mucro or awn-like process at the base of
together and imbibition of two subapical tissue pads with the anther thecae (the other genera lack such a process,
pectinised cell walls triggers the release. In hygrochastic except sometimes Hygrophila, in which the thecae may
species of Barleriinae Nees (Barleria L., Crabbea Harv.) be “mucronulate”, Wasshausen 2005). The corolla limb
Willdenowia 40 – 2010 287

is (almost) regular to (slightly) bilabiate. The capsule has tion of a new genus in the Acanthaceae”. Alain (1995)
a short sterile stipe and is spindle-shaped. Seed number accepts Long’s opinion. On the other hand, Scotland &
is (2 – )4. Chromosome counts are available for five species Vollesen (2000) maintain Apassalus. Current usage, as
(Anonymous 2010) and indicate a consistent basic chro- documented by Web-based treatments, is split, with e.g.
mosome number of x  = 15 at the diploid or tetraploid level. Wasshausen (2010) maintaining Apassalus but Wunder-
Mucilaginous seed hairs have been described by Kip- lin & Hansen (2010) merging it under Dyschoriste.
pist (1842) for Dyschoriste erecta (Burm. f.) Kuntze
(as D. cernua Nees) and D. litoralis Nees, by Schaffnit Hygrophila is pantropical, comprising 20 – 30 species
(1906) for D. linearis (Torrey & A. Gray) Kuntze (as Ca­ (Wasshausen 2005; Mabberley 2008) rather than the pre-
lophanes linearis (Torrey & A. Gray) A. Gray) and by viously estimated 100 (Mabberley 1987). The nomenclat-
Oehm (1932) for D. microphylla (Cav.) Kuntze (as C. jas­ ural type, provided by Ruellia ringens L., is an Old World
minum-mexicanum Nees) and D. nagchana (Nees) Ben- element. Hygrophila species consistently have a bilabiate
net (as C. fasciculiflora (Fenzl ex Sond.) Martelli). The corolla limb, often strongly so. The anther thecae have a
hairs cover the whole seed surface, which is described as blunt base or may be mucronulate (Wasshausen 2005).
dull when dry, and have annular and/or spiral thickenings. Seed number (4 – 32 per capsule) varies but is usually
Hairs limited to the thickened seed margin were reported high. The capsule shape is ± cylindrical, with a truncate
by Stopp (1958) for the S African D. depressa (L. f.) Nees base not narrowed into a sterile portion. Chromosome
and other (unspecified) species from that area. counts are available for seven species (Anonymous 2010)
The pollen of 34 species (including the three previ- and indicate a prevalent basic number of x  = 16 at the di­
ously placed in Chaetacanthus Nees) was studied by ploid or tetraploid level, but x  = 6, 11 (or 22) and 12 have
Furness (1995): grains prolate spheroidal to prolate, also been reported.
3-colporate, with a variable number of prominent lon- Mucilaginous seed hairs have been described by Kip-
gitudinal, microechinate exine ribs separated by pseu- pist (1842) for five species and by Schaffnit (1906) for 14
docolpi; the colpi are shorter than most pseudocolpi and species (including one he placed in Cardanthera Buch.-
the rib sides bordering the aperture bulge up on either Ham. ex Benth. and another one in Nomaphila Blume).
side to form an atrium around the ± circular endopore The hairs cover the whole seed surface, which when
(Scotland 1993: fig. 42; Furness 1995: pl. I – V; Daniel dry is said to have a silky sheen, and have annular and/
1998: fig. 4A – C). This pollen type is remarkably uni- or spiral thickenings. Kippist mentions a single species,
form (Furness distinguishes seven subtypes based on H. polysperma (Roxb.) T. Anderson (as Hemiadelphis
pseudocolpi number) and, disregarding one anomalous polysperma (Roxb.) Nees), among those with seed hairs
sample, diagnostic for the genus. lacking mucilage and devoid of thickenings, but verifica-
In Cuba Dyschoriste (when kept separate from Apas­ tion of this assessment is desirable.
salus) is represented by the single D. bayatensis, which The pollen (Scotland 1993: fig. 9 – 11; Scotland &
was not included by Kobuski (1928b) in his revision, be- Vollesen 2000: fig. 9A – B [same images with different
cause it had been described in Hygrophila and not yet species identities]; Furness 1994: pl. I – III) has been de-
transferred to Dyschoriste at the time. scribed in some detail, for 34 species, by Furness (1994).
Hygrophila and the closely related Brillantaisia P. Beauv.
Apassalus was established by Kobuski (1928a) as a ge- share a unique general pollen morphology: suboblate to
nus of three species segregated from Dyschoriste: A. prolate spheroidal, 4-colporate grains with a distinctive
diffusus (Nees) Kobuski endemic to Hispaniola (provid- pattern of raised longitudinal exine ribs separated by the
ing the type), Cuban A. cubensis (Urb.) Kobuski and A. colpi and interjacent pseudocolpi. The apertures are usu-
humistratus (Michx.) Kobuski of the SE United States. ally equatorial, but opposite pairs are sometimes ± shift-
Subsequently, a second Cuban species was described, ed toward opposed poles. The general pattern is reminis-
A. parvulus (Alain 1956). The genus is characterised by cent of that found in Dyschoriste (see above) but differs
creeping habit, the small size of all parts of the plants in aperture number, in colpi which are not shorter than
and principally the fact that the base of the anther the- average pseudocolpi, and in the absence of exine bulges
cae is not mucronate but rounded. The corolla limb is collateral to the pores. Exine sculpture also differs, but is
almost regular. The capsule has a short sterile stipe and is variable in Hygrophila and was used by Furness (1994)
spindle-shaped. The seeds are 2 or 4 per capsule. Chro- to define several different pollen types.
mosome number is 2n  = 30 in A. humistratus (Hilsebeck According to Sell (1969), Hygrophila auriculata
1976). The seed hairs had not so far been described and (Schumach.) Heine (as spinosa T. Anderson; a wide-
the pollen is only known through the brief characterisa- spread palaeotropical species) has hygrochastic but a ma-
tion of D. cubensis by Raj (1961). jority of its congeners has xerochastic fruits.
The taxonomic status of Apassalus is controversial. In Cuba the genus is represented by Hygrophila cos­
Long (1970: 295) writes: “The absence of anther ap- tata Nees & T. Nees, a widespread and variable Ameri-
pendages appears to be the only consistent character for can species also known under numerous synonyms, and
the genus [Apassalus], and this hardly warrants recogni- H. urquiolae.
288 Greuter & Rankin: Notes on endemic species of Ruelliinae

Ruellia, with its 250 or perhaps 300 species (Wasshausen Alain (1957) reports 10 Ruellia species from Cuba,
& Wood 2004; McDade & Tripp 2007), is the second larg- which number, if one follows Ezcurra & Daniel (2007)
est genus in the family after Justicia. It is also pantropical in synonymising R. tweedieana Griseb. with R. simplex
in distribution but is based on a New World nomenclat- C. Wright, is reduced to 9 (but see Ward 2007). Under
ural type, provided by R. tuberosa L. As currently de- Bremekamp’s generic concept, as outlined by Long
fined, Ruellia encompasses at least 21 generic synonyms, (1973), only two of them would remain in Ruellia: R.
which is a good indication of its polymorphism. Ezcurra’s tuberosa and R. nudiflora (Engelm. & A. Gray) Urb.; a
(1993) revision of the southern South American species, third, R. geminiflora Kunth, provides the type of Ulleria
even though limited in geographical coverage, conveys a Bremek.; whereas the “microgeneric” placement of the
fair impression of that variability. Typically the corolla is remaining 6, including R. shaferiana, is unsettled. If, on
large and showy, often vividly coloured, and its limb is the contrary, one accepts the recent synthetic, molecule-
usually ± regular but sometimes distinctly bilabiate. The based approach by Tripp (2007) and Tripp & al. (2008),
anther thecae are rounded at both ends, lacking a basal one must add Ruellia blechum L. (formerly Blechum py­
mucro. Seed number (2 – 26) varies but is typically high. ramidatum (Lam.) Urb. and B. blechioides (Sw.) Hitchc.)
The capsule is ± spindle-shaped, being narrowed at the to the count.
base into a sterile, solid stipe of variable length, but R.
tuberosa and its relatives (i.e. ­Ruellia in the strict sense) In the present paper, we propose to (1) present and
have cylindrical, many-seeded, non-stipitate capsules compare the five target species, their known distribu-
(Long 1973). Chromosome counts are available for c. 40 tion and habitat, based on available herbarium material,
species (Anonymous 2010) and indicate a prevalent basic adding a detailed morphological description of Ruellia
number of x = 17, rarely 16 or (11 –)12, almost invariably shaferiana, a seldom collected and incompletely known
at the diploid level. species; (2) study the seeds, with emphasis on muci-
Mucilaginous seed hairs have been described by Kip- laginous hairs; (3) perform an REM-based study of the
pist (1842) for many, mostly unspecified species, and pollen; and (4) discuss hygroscopic features, capsule
three named ones he placed in Dipteracanthus Nees; and dehiscence in particular, and other traits potentially rel-
by Schaffnit (1906) for 13 species. Variability is great- evant to reproductive biology, such as daily flowering
er than in the other genera. The hairs usually cover the rhythm.
whole seed surface, but in several species (including the
three that Kippist placed in Dipteracanthus, said to have
Material and methods
narrowly winged seeds, plus Ruellia ochroleuca Mart. ex
Nees, R. rubicaulis Cav. and R. solitaria Vell. mentioned Plant material collected by us has been deposited in the
by Schaffnit), they are confined to the margin. The an- herbaria B, HAJB, HIPC, JE and PAL-Gr (in due course,
nular and spiral wall thickenings are usually well devel- available duplicates will be distributed to further her-
oped but may be weak or absent. Finally, in two of the baria). Specimens kept in the herbaria B, HAC, HAJB,
species studied by Schaffnit (1906: 460 – 461), R. elegans HIPC, JE and S, as well as digital photographs from
Poir. (R. formosa “Andrews”) and R. squarrosa (Fenzl) FLAS, FSU, NY, S and USF, have been studied for com-
Schaffnit, the hairs are replaced by mucilage cells closely parison.
resembling those found in Blechum. For Ruellia shaferiana and Hygrophila urquiolae,
The pollen of Ruellia and its close relatives differs morphological details of the flowers were observed on
conspicuously from that of the preceding three genera. FAA-fixed material. Details of flowers of live plants were
According to Lindau’s (1895) classification, it belongs documented with the aid of a Canon digital camera IXUS
to the category of honeycomb pollen (“Wabenpollen”), 9515. In R. shaferiana, diurnal flowering rhythm was ob-
whereas Dyschoriste, Hygrophila and, by implication, served, on three consecutive days, on cut branches kept
Apassalus have ribbed pollen (“Rippenpollen”). The pol- in a vase.
len grains of Ruellia are spheroidal, lacking obvious polar- Seeds were obtained from herbarium specimens, wet-
ity, 3(– 5)-porate or 3-colporate to poly-pantoporate. With ted in plain water and observed in incident light under a
minor exceptions (neglected here) they are surrounded stereomicroscope WILD M5A. Seed and other close-up
by a coarse net of high exine structures best described photographs were taken with a digital camera Leica DFC
as “reticulate handrails” supported by robust, simple or 290.
branched columellae. Good examples can be seen in Long Pollen was extracted from FAA-fixed buds (Ruellia
(1973: fig. 6 – 11), Ezcurra (1993: fig. 4), Scotland (1993: shaferiana and Hygrophila urquiolae) or sampled on
fig. 21 – 24) and Furness & Grant (1996: fig. 13 – 36). herbarium material (Dyschoriste bayatensis, Apassalus
Sell (1969) mentions hygrochastic capsules for Ruel­ parvulus), then acetolysed following the standard proce-
lia tuberosa (Neotropics, including Cuba), R. ciliatiflora dure described by Erdtman (1960). When the pollen sam-
Hook. (as R. lorentziana Griseb.; South America) and R. ple was scant, it was not centrifuged after acetolysis but
prostrata T. Anderson non Poir. (Africa), and xerochastic concentrated manually in a watch glass, and the solvent
fruits for an equal number of other species. pipetted off, as described by Leuenberger (1976).
Willdenowia 40 – 2010 289

For scanning electron microscopy (SEM), pollen and Distribution. — According to Alain (1957), Dyschoriste
seeds were coated with a 20 nm layer of gold-palladium bayatensis occurs in E Cuba, the Camagüey Province in
in a Low Voltage Cool Sputter Coater EMITECH K 550, Central Cuba and the neighbouring island of Hispaniola,
at 20 mA under a vacuum of c. 0.2 mbar. They were stud- but the latter indication is definitely erroneous, as Alain
ied, partly under a FEI Quanta 200 SEM at the Depart- (1995) himself does not mention the species in his Flo-
ment of Zoology and partly under a Zeiss LEO 430 SEM ra of Hispaniola. The main area of the species is in the
at the Department of Systematic Botany, Free University Cauto river basin in E Cuba, from where six gatherings
of Berlin. exist, all made by the Swedish collector Ekman in three
localities between 1915 and 1923: near Bayate (the lo-
cus classicus), Bayamo and Guisa. Alain’s record from
Results
Camagüey is based on a single specimen, collected by
Dyschoriste bayatensis (Urb.) Urb. in Ark. Bot. 24A (4): Acuña, which we have studied and found not to deviate
35. 1932 ≡ Hygrophila bayatensis Urb., Symb. Antill. 9: in any significant detail from the E Cuban plants. – Fig. 1.
127. 1923. – Holotype: Cuba, “Prov. Oriente, prope Ba­
yate in fruticetis densis ad marginem Savanna Miranda, Habitat. — Thickets and fences, scrub margins, occa-
solum hic observata!, 22.3.1917, Ekman 8566 (?B†); sionally a straggling climber; lowland below 200  m of
lectotype (designated here): S [photo], isolectotypes: F altitude.
47389F [photo], NY 1185563 [photo], US 955934 [pho-
to]). Seeds. — There are normally 4 seeds per capsule. They
are lenticular-compressed, broadly elliptic in outline, with
Morphology. — We refer to the full original description blunt wingless margins, measuring c. 2 × 1.4 mm. The dry
of Urban (1923) for details. Important diagnostic fea- seeds (Fig. 2A) are covered by a light-grey, silky layer of
tures, as verified by us, are summarised in Table 1. tightly packed, parallel, antrorsely appressed hairs that are
discernible only towards the distal end, where the densely
Additional specimens seen. — Cuba: [Prov. & Munic. set annular thickenings of the cell wall become evident.
Camagüey]: Caminos al sur de C. Camagüey, 20.12.1925, When wetted, the hairs immediately spread and surround
Acuña 13797 (HAC). – [Prov. Granma, Munic. Bayamo]: the seed as a dense down, releasing mucilaginous mat-
Bayamo, in thickets at Rio Bayamo, 23.1.1923, Ekman ter which, in water, can be seen as a milky, unstructured
16196 (S [n.v.]); [Munic. Guisa]: Guisa, in living fences cloud (Fig. 2B). The mucilage hairs are 0.4 – 0.5 mm long,
along the road, a Spreizklimmer up to 2 m, fl. pale lila, erecto-patent except at the seed base.
3.1.1923, Ekman 16061 (HAC, NY 1185564 [photo],
S [photo]); Guisa, 3.1.1923, Ekman [16061] ex Herb. Pollen morphology. — The pollen grains (Fig. 4) are typical
Cañizares 15a (HAJB). – [Prov. Santiago de Cuba, Mu- for the genus: subprolate 3-colporate “Rippenpollen” with
nic. Mella]: Prope Bayate in fruticetis densis ad mar- equatorial pori. Average dimensions (polar axis, equato-
ginem Savanna Miranda rara, 5.2.1915, Ekman 4549 (S rial diameter and their ratio) are: P = 42.5 (40.6 – 44.7) μm,
[photo], NY 1185565 [photo]); Bayate, Sabana Miranda, E = 35.5 (33.6 – 37.6) μm, P/E = 1.19 (1.11 – 1.29). The pseu­
in thickets, 30.10.1922, Ekman 15595 (S [n.v.]); same docolpi are fairly variable in num­ber (6 – 8 per intercol-
data, 8.12.1922, Ekman 15849 (S [photo], NY 1185566 pium) and often somewhat irregular in size (both length
[photo]); en Bayate, 3.12.1922, Ekman [15849] ex herb. and width), which places our species in “subtype e” as
Cañizares 15 (HAJB); en Sabana Miranda, s.d., Ekman defined by Furness (1995), together with a number of
ex herb. Cañizares 15b (HAJB). Old World (predominantly African) species. However,

Fig. 1. Known distribution of Dyschoriste bayatensis (l), Apassalus cubensis (p) and A. parvulus (q).
290 Greuter & Rankin: Notes on endemic species of Ruelliinae

Furness’ subtypes are but an

artificial, arbitrary classifi-
cation device. The colpi are
small and narrow, so that the
pore appears to open in the
centre of one medially split
twin rib (Fig. 4A, C). The exi­
ne lips on either side of each
circular endopore are well de-
veloped, regular, protruding
above the surface by c. 2 μm,
and occasional fractured pol-
len (Fig. 4D) shows that they
enclose a distinct atrium. The
tectum is densely and regu-
larly microgranulate (granule
size: 0.2 – 0.8  μm), otherwise
compact, with no apparent
reticulum and hardly any mi-
croperforations (Fig. 4B). Raj
(1961: 30, under Hygrophila)
gives a detailed description
of the exine structure, based
on light microscopic study of
sectioned pollen grains, but
his measurements are slightly
at variance with ours.

Hygroscopic features. — Ex-

posing apparently mature
capsules of herbarium speci-
mens to the vapour of boiling
water did not suffice to trig-
ger dehiscence. The capsule
surface, including the tip, is
water repellent. Neverthe-
less, when a water droplet
was placed atop a capsule tip,
the capsule after c. 5 minutes
opened explosively and eject-
ed two of its seeds to some
distance (5 – 10 cm). The open
capsules are eventually shed
from an abscission layer at
their base, and when this hap-
pens the void calyx segments,
in their distal half, curve
slightly but perceptibly in-
ward. This is a reversible hy-
groscopic movement: when
wetted, the calyx segments
straighten within a matter of
minutes to diverge stellately
from above the base. (All
Fig. 2. SEM (A, C) or incident light photographs (B, D – H) of dry (left) and wetted (right)
observations were made on seeds – A – B: Dyschoriste bayatensis (Ekman [16061]); C – D: Apassalus parvulus (HFC
the specimen Ekman ex herb. 46906); E – F: Hygrophila urquiolae (Greuter & al. 26662); G – H: Ruellia shaferiana (Greuter
Cañizares 15a, HAJB). & al. 25199). – Scale bar = 1 mm.
Willdenowia 40 – 2010 291

Fig. 3. SEM photographs of dry (A – B) and once-wet (C – F) seeds – A – D: Hygrophila urquiolae (Greuter & al. 26662); E: Apassa­
lus parvulus (HFC 46906); F: Ruellia shaferiana (Greuter & al. 25199). – Scale bars: A, C, E – F = 100 µm, B = 50 µm; D = 10 µm.

Apassalus cubensis (Urb.) Kobuski in Ann. Missouri 11.1.1956, Alain A-4991 (HAC); el Puntón del Cuero,
Bot. Gard. 15: 2. 1928 ≡ Dyschoriste cubensis Urb., Symb. Via Azul, 11.1.1956, Alain, Morton & López Figueras
Antill. 7: 381. 1912. – Holotype: [Cuba, Prov. Ciudad de la A-4991 (HAC).
Habana, Munic. Habana del Este], Cojimar, 14.3.1906,
Baker 2894 (?B†); lectotype (designated here): NY Morphology. — We refer to the full descriptions of Ur-
114981 [photo]; isolectotype: HAJB ex herb. Canizares). ban (1912), Kobuski (1928a) and Alain (1956) for de-
tails. Important diagnostic features, as verified by us, are
Apassalus parvulus Alain & Leonard in Contr. Ocas. summarised in Table 1. Pubescence is very variable in
Mus. Hist. Nat. Colegio “De La Salle” 15: 20. 1956. – density and length. The square stems are often shortly
Holotype: [Cuba, Prov. Guantánamo, Munic. Imías], pubescent on two opposite faces (those corresponding
subida al Puntón del Cuero, Imías, 11.1.1956, Alain & to the leaf axils of the proximal node) and glabrous on
Morton 4991 (HAC); isotypes: Puntón del Cuero, Imías, the two others.
292 Greuter & Rankin: Notes on endemic species of Ruelliinae

Fig. 4. Pollen of Dyschoriste bayatensis (Ekman [16061]), SEM micrographs – A: equatorial view; B: id., detail of exine ribs and
pseudocolpi; C: subpolar view; D: fractured grain, showing atrium formation and, from within, circular endopore. – Scale bars: A,
C – D = 20 µm; B = 5 µm.

Additional specimens seen. — Apassalus cubensis: 16886 (K 534044 [photo], NY 1185568 [photo]); be-
Cuba: Prov. Pinar del Río, Munic. Viñales: En el agua tween Rio Cojímar and Playa de Bacuranao, vine; fls.
del río [?], finca El Ancón, Viñales, 19.5.1955, Alain white; shady places in coastal sands, 26.12.1910, Wilson
4258 (HAC). – Prov. Ciudad de la Habana, Munic. 9533 (NY 1185571 [photo]); cerca de Playa entre / sea
Habana del Este: lectotype, Cojimar, 10.1.1905, Baker coast between Cojímar y / and Bacuranao, 29.12.1910,
& O’Donovan 4421 (NY 1185575 [photo]); playa de Wilson & León 1912 (HAC, NY 1185573 [photo]). –
Cojímar, 7.5.1951, Daudinet 16833 (HAC); Cojimar, Prov. Habana, Munic. Santa Cruz del Norte: Boca de
coastal thicket, 24.8. 1910, Britton & al. 6300 (NY Canasí (Matanzas), on rocks, 7.12.1927, León 13192
1185574 [photo]); manigua costera / coastal thickets, (HAC). – Prov. Matanzas: ad rupes supra Matanzas,
Cojímar, 14.2.1953, Alain 2756 (HAC, NY 1185569 [1849], Rugel 109 (NY 1185577 [photo]); ad rupes mar-
[photo]); manigua costera, Cojímar, 2.3.1931, León itimas prope Matanzas, [1849], Rugel 741 (NY 1185576
& Alain (HAC); id., 6.7.1957, Alain 6304 (HAC); in [photo]). – Prov. Ciego de Ávila, Munic. Morón: Cami-
coastal ­thickets between Morro and Cojímar, 5.4.1922, no a la cueva del Jabalí, Cayo Coco, Ciego de Ávila,
Ekman 13693 (MO 807432 [photo], NY 1185567 [pho- 19.2.1995, Herrera & al. (HAC 40681).
to]); El Morro to Cojimar, prostrate in shady places, Apassalus parvulus: [Prov. Guantánamo, Mu-
under shrubs; fls. white, 8.12.1910, Wilson 9143 (NY nic. Imías]: Veguita del Sur, alrededores del río Jojo
1185570, 1185572 [photo]); Casa Blanca, in coastal [= Jobo], monte galería semicaducifolio [20°06'50''N,
thickets between Morro and Cojímar, 6.7.1923, Ekman 74°28'52''W], 24.5.1982, Bässler & al. HFC 46906 (B,
Willdenowia 40 – 2010 293

Fig. 5. Pollen of Apassalus parvulus (HFC 46906), SEM micrographs – A: equatorial view; B: polar view; C: some exine ribs and
aperture, showing part of circular endopore and atrium; D: polar cap, detail of tectum with perforations and microgranules. – Scale
bars: A – B = 2 µm; C – D = 1 µm.

JE); id., esquistos cristalinos, 24.5.1982, Bisse & al. and occasional seed abortion is frequent in Acanthace­
HFC 46906 (HAJB). ae. One year later, Alain (1957) keyed out A. parvulus
against A. cubensis, based on indumentum, leaf crena-
Distribution. — According to Alain (1957), Apassalus tion and dimensions, and calyx and corolla length. The
cubensis has a fairly wide distribution along the coasts second, copious gathering of A. parvulus, HFC 46906,
of Cuba, being reported from the provinces of Pinar del from near the locus classicus, shows that the vegetative
Río, [Ciudad de] la Habana and Matanzas, and also from characters of that species are fully congruent with those
“Oriente”. As we did not find E Cuban material identi- of A. cubensis, whereas for calyx, corolla and capsule
fied as A. cubensis in any of the herbaria we consulted length, the values for the E Cuban plants do not fall
(which include those with which Alain was working), we within the variation range of A. cubensis. The differenc-
consider the “Oriente” record as due to a slip of memory. es are slight, but added to the geographical disjunction
A. parvulus is only known from the municipality of Imías they just suffice to recognise distinct species, as we shall
(Fig. 1). argue in the concluding part of our paper. In the follow-
ing discussion, unless otherwise stated, we refer to both
Habitat. — Both species are apparently restricted to species together.
shady places, either on coastal sands or in gallery woods,
below 100 (or occasionally 300) m of altitude. Seeds. — There are normally 4 seeds per capsule. They
are similar in shape to those of Dyschoriste bayatensis,
Variation. — Apassalus parvulus was based on a single with blunt wingless margins, but with a markedly ob-
gathering made in easternmost Cuba at the beginning of lique base, and somewhat smaller (c. 1.5 × 1 mm). The
1956. In the protologue, Alain (1956) failed to compare dry seeds (Fig. 2C) are covered by a brownish shell that
it with A. cubensis, considering it closest to A. diffusus consists of tightly packed, parallel, antrorse hairs glued
of Hispaniola on account of seed number (2 rather than together by a kind of lacquer and discernible only to-
4 per capsule). This is a questionable criterion, as he wards the distal end and sometimes at the bend beside
described A. parvulus as “4-seeded, one pair abortive”, the hilum. The annular thickenings of the cell wall are
294 Greuter & Rankin: Notes on endemic species of Ruelliinae

less pronounced than in the

previous species, because the
wall between the thickenings
is more robust and does not
cave in. When wetted, the
hairs immediately spread at a
right angle and surround the
seed, releasing a conspicuous
mucilage cloud (Fig. 2D). Be-
ing stouter than in the other
species studied, they remain
± upright and do not collapse
upon drying (Fig. 3E). Their
length is c. 0.25 mm.

Pollen morphology. — Same

as Dyschoriste bayatensis,
Apassalus parvulus has sub­
prolate, 3-colporate Rippen-
pollen with equatorial pori
(Fig. 5). Average dimensions
are: P = 31.4 (27.7 – 36.5) μm,
E = 25.4  (22.6 – 29.4)  μm,
P/E = 1.24 (1.11 – 1.30). There
are 6 – 8 pseudocolpi per inter-
colpium (Fig. 5B), which are
somewhat irregular in size
(both length and width), so
that under Furness’ (1994)
classification the Apassalus
pollen falls within the same
subtype as that of D. bayaten­
sis. The shape and dimensions
of the colpi and of the char-
acteristic exine lips protrud-
ing on either side of the pores
(Fig. 5A,  C) do not deviate
from those seen in the latter
species. However, there are Fig. 6. Apassalus parvulus (HFC 46906), incident light photographs – A: capsule with tip burnt
minor differences. The grains off, exploded to eject one seed; B: dry post-fruiting calyx with hygroscopically incurved seg-
are spindle-shaped rather ments; C: same calyx, expanded after wetting, displaying the axial stump on which the now
shed capsule was articulated. – Scale bars: A – B = 1 cm; C = 2 cm.
than broadly ellipsoidal, i.e.
they taper ± conically toward
the poles (Fig. 5A). The tectum is rugulose-scrobiculate obtained by placing a droplet of water, for about 10 min,
underneath the microgranular pattern and is micro- on the tip of a mature capsule of A. cubensis (specimen
perforate, especially on the lips and at the poles (Fig. Herrera & al., HAC 40681). Reversible hygroscopic
5C – D). Raj (1961) characterised the pollen of Dyschor­ bending of post-fruiting calyx segments was observed
iste cubensis somewhat differently, namely as prolate in several specimens. The void capsule, disarticulating
(47 × 28 μm); he did not, however, provide an illustration at the base, is shed; by the time this happens the proxi-
or detailed description. mal portion of the calyx, having become hardened, splits
to the base, from where the formerly upright segments
Hygroscopic features. — The tip of one of the few avail- now diverge, then curve inward in their proximal third,
able ± mature capsules of Apassalus parvulus (from the with the distal part remaining straight. This bending is
specimen HFC 46906, JE) was burnt off by means of a not very obvious in A. cubensis but is extreme in plants
red-hot needle. The capsule readily exploded, ejecting of A. parvulus from both E Cuban localities and results
one of its seeds to a distance of c. 10 cm (Fig. 6A, where in an awkward, “spidery” look of the old calyces (Fig.
the seed has been re-positioned). The same result was 6B), which upon wetting inverts to a regularly stellate
Willdenowia 40 – 2010 295

Fig. 7. Known distribution of Hygrophila urquiolae.

shape (Fig. 6C). This feature, not so far described, lends Prov., is puzzling. The labelling is apparently reliable,
further support to the recognition of both species as dis- but one might suspect a recent introduction as a casual
tinct. roadside weed, because that very same area had been
repeatedly explored by previous botanical expeditions,
Hygrophila urquiolae Greuter & al. in Willdenowia 39: which failed to collect the species. – Fig. 7.
288. 2010.
Habitat. — Marshy flat coastal grassland areas inter-
Morphology. — This rare Cuban endemic was described spersed with copses and patches of shrubbery, near sea
only recently (Greuter & al. 2010). We refer to the full level. An entirely different habitat, montane rainfor-
original description for details and include the character- est above 600 m of altitude, is indicated for the Monte
istic features in Table 1. Cristo specimen. We cannot presently explain this very
unusual ecological discrepancy (see above, under dis-
Specimens seen. — Cuba: As “Dyschoriste bayaten­ tribution).
sis” [lacking label data] (HAC). – Prov. Matanzas, Mu-
nicipio Martí: Ciénaga del Majaguillar al NO de Martí, Seeds. — There are normally 32 seeds per capsule. They
Ciénaga de Gonzalito cerca del Canal de Blanquizal, are similar in shape to those of Dyschoriste bayatensis,
alt. 5 m, 22°59'39''N, 80°58'06''W, herbazales de ciénaga, with blunt wingless margins, but relatively broad and
28.2.2009, Greuter & al. 27013 (holotype: HAJB; iso- somewhat smaller (c. 1.2 × 1  mm). The dry seeds (Fig.
types B, JE, NY, PAL-Gr); same locality, 22°59'45''N, 2E) are covered by a compact brown shell consisting of
80°58'45''W, bosque secundario de ciénaga, 10.2.2007, parallel, antrorse hairs glued together as if lacquered over
Palmarola & al. HFC 84651 (HAJB); same lo­cality, and not discernible individually (Fig. 3A, B). The annu-
flo­res lila pálido con tubo blancuzco, 27.2.2007, Greu­ lar thickenings of the cell wall are very numerous and
ter & al. 26662 (B, HAJB, PAL-Gr). – Prov. Granma pronounced (Fig. 3D). When wetted, the hairs immedi-
[Munic. Río Cauto]: Monte Atejito, Ciénaga de Bira- ately spread in the same way and with the same orienta-
mas, 29.1.1998, Méndez 9692 (HIPC, PAL-Gr). – Prov. tion as in D. bayatensis, amidst a dense mucilage cloud
Guantánamo, [Munic. Yateras]: camino a Felicidad de (Fig. 2F). They are long (0.4 – 0.5 mm) and slender, col-
Yateras (7 – 10 km al N), Montecristo, pluvisilva montana lapsing completely upon drying (Fig. 3C).
sobre caliza, 600 – 800 m/nm, 23.1.1996, Méndez & Ro­
mano 8626 (HIPC). Pollen morphology. — The pollen grains are typical for
the genus: spheroidal to suboblate, 4-colporate “Rippen-
Distribution. — When we described Hygrophila urquio­ pollen” (Fig. 8) with apertures that are either equatorial
lae from a single, restricted locality in the Ciénaga de or displaced by opposite pairs toward opposed poles (Fig.
Majaguillar of Matanzas Province, we already suspected 8A). Average dimensions are: P = 32 (31.3 – 32.8)  μm,
that it might be more widely distributed but was so far E = 34 (30.8 – 39.5)  μm, P/E = 0.94 (0.79 – 1.03). There
overlooked. We thought of similar habitats in W Cuba, are usually 4 (rarely 3 or 5) fairly regular, smooth, nar-
but when revising the Acanthaceae of the HIPC herbar- row pseudocolpi per intercolpium, which together with
ium had the surprise of finding two E Cuban specimens the colpi delimit a total of (17 –)20 exine ribs per grain
unmistakeably representing H. urquiolae. The first is (Fig. 8B). The colpi are almost as long as the pseudo-
from the Ciénaga de Biramas, Granma Prov., an area in colpi, much widened and distinctly granulate in their
which similar, suitable habitats are to be expected. The medial part; there are no marked, lip-like exine protru-
second, from the Monte Cristo region of Guantánamo sions on either side of the circular endopores, which
296 Greuter & Rankin: Notes on endemic species of Ruelliinae

Fig. 8. Pollen of Hygrophila urquiolae (Greuter & al. 26662), SEM micrographs – A: equatorial view; B: polar view; C: aperture,
showing granular colpus, smooth adjacent pseudocolpi and circular endopore; D: subpolar exine portion. – Scale bars: A – B = 2 µm;
C – D = 1 µm.

are, however, bordered by a granular rim (Fig. 8C). The Ruellia shaferiana Urb., Symb. Antill. 9: 128. 1923. –
tectum, on ribs and pole areas, shows a pronounced re- Holotype: Cuba, Prov. Camagüey [Munic. Nuevitas],
ticulate pattern with rather few but coarse perforations vicinity of Tiffin [c. 21°33'N, 77°20'W], 1. – 5.11.1909,
inside the primary lumina, which decrease in size from Shafer 2880 (?B†); lectotype (designated here): NY
the equator to the poles (Fig. 8D). According to Fur- 114918 [photo]).
ness’ (1994) classification, our species groups with four
others of Hygrophila and one of Brillantaisia, all Pa­ Morphology. — Urban (1923) described Ruellia shafe­
laeotropical, which share the Hygrophila salicifolia pol- riana rather briefly, based on an incomplete speci-
len type. men bearing flower buds and immature capsules only.
Thanks to our new material, complemented by observa-
Hygroscopic features. — Exposure of mature but still tions in the field, we can here provide a full and detailed
closed capsules of the type specimen to the vapour of description. The main characteristics are summarised in
boiling water almost immediately triggered their explo- Table 1.
sive opening. Wetting the tip by a drop of water, however,
did not function unless some detergent was added. Not Perennial herb, 20 – 40  cm tall, procumbent below and
only the seeds but the capsule halves were ejected in this rooting at the nodes, the fertile portion ± erect. Epider­
species, breaking free at the basal articulation point. The mis covered with linear cystoliths raised in the dry state,
firm, tubular calyx does not show any hygroscopic prop- which are parallel along the veins on the abaxial leaf
erties. surface, stems and calyces, but irregularly arranged on
the adaxial face of the leaves; in fresh material they are
Daily flowering rhythm. — Hygrophila urquiolae appears barely visible, apparently sunk rather than protruding.
to be a normal day-flowering species. We did not attempt Stems initially quadrangular, later almost terete, not sul-
to establish its exact flowering cycle, but it was certainly cate, glabrous, inflated at the nodes. Lower leaves shed
in full bloom in the early afternoon hours when we took by flowering time, the middle and upper ones decussate,
pictures of it. thin, undivided, the middle measuring up to 8 × 3 cm, the
Willdenowia 40 – 2010 297

Fig. 9. Known distribution of Ruellia shaferiana.

upper much smaller; lamina broadly lanceolate to ovate- apically recurved in bud, becoming straight at anthesis,
lanceolate, acute or subacute, tapering basally into a short its tip reaching the opening of the corolla throat; stigmatic
petiole, with entire but often undulate margins that are branches bluish in bud, diverging from the style at right
ciliolate when young, otherwise glabrous on both faces, angles in the sagittal plane, the anterior filiform, c. 1 mm
pinnately veined with 6 – 7 pairs of abaxially prominent long, eventually becoming porrect, the posterior usually
main veins and a loosely reticulate tertiary venation. In­ reduced to a small tooth (Fig. 10Da) but occasionally of
florescences dichasial; flowers sessile, mostly in threes the same length and shape as the anterior one (Fig. 10Db).
(the lateral borne by linear-spatulate bracteoles) in the Capsule loculicidal, straight, narrowly spindle-shaped,
axils of foliaceous to linear-spatulate bracts, ± spaced or acute and mucronate, equalling or slightly exceeding the
densely crowded on axillary side branches of which only calyx, measuring 9 – 11 × c. 1.5 mm, with a sterile basal
the lowermost are ± elongate but the upper very short. Ca­ and apical portion; valves brown; placental suture a nar-
lyx 7 – 8  mm long at anthesis, accrescent to 10 – 11  mm, row groove; retinacula robust, biseriate.
split to the base into 5 subequal, green, linear segments
narrowing gradually to a setaceous tip; segments devoid Additional specimens seen. – Cuba: Prov. Habana [Mu-
of a membranous margin, glabrous inside, outside on the nic. Melena del Sur or Güines]: Batabanó, La Mora,
midvein and especially along the margins with a mixture in palm savannas behind the manglares [ca. 22°43'N,
of long (up to twice as long as the segment width) gland- 82°06'W], 11.12.1920, Ekman 12631 (S). – [Prov. Ma-
tipped and scattered slender, short, eglandular patent hairs tanzas, Municipio Martí]: Circuito Norte, Man. Gómez y
(Fig. 10A). Corolla c. 14 mm long, exceeding the calyx Martí, 19.8.1956, Acuña, Roig & Acosta 20360 (HAC).
by half, with left-contort aestivation, divided into tube and – Prov. Camagüey, [Munic. Nuevitas]: Tiffin, in pastures
limb (including throat) by the insertion point of stamens; [c.  21°33'N, 77°20'W], 7.10.1922, Ekman 15400 (S,
tube cylindrical, straight except at the ± outward bent top, 2 ×); costa antes del puente de Cayo Sabinal, alt. 0 – 1 m,
as long as the limb (7 mm), glabrous, pale; limb obliquely 21°36'10''N, 77°22'45''W, 3.3.2000, Greuter & al. 25199
inserted on the tube then imperceptibly upward curved (B, HAJB, HIPC, PAL-Gr); costa antes del puente de
(Fig. 10A) but otherwise subregular, glabrous inside but Cayo Sabinal, alt. 1 – 2 m, 21°35'46''N, 77°22'38''W,
outside glandular- and eglandular-puberulent (Fig. 10C), 22.2.2010, Greuter & al. 27118 (B, HAJB, HIPC, PAL-
especially apically and along the veins, bright blue at an- Gr, etc.). – Prov. Las Tunas, [Munic. Manatí]: Sabanas de
thesis turning pale burgundy on withering, marked with Manatí, 24.11.1951, Acuña 17161 (HAC).
5 darker veins that do not attain the margin; throat nar-
rowly campanulate, the 5 ± equal, truncate-obovate or Distribution (Fig. 9). — The area of distribution, as in-
rounded-squarish lobes (3.5 × 2.5  mm) curved outward dicated by Alain (1957), comprises E Cuba (the former
and patent distally. Stamens 4; filaments glabrous, fused Prov. Oriente) and Camagüey Province. A large major-
pairwise at the base, each pair inserted obliquely at the ity of the material seen, including our own gatherings,
tip of the corolla tube, the pairs not united by a transverse comes from the latter. In fact, no specimen from any of
membrane or fold; staminode absent; anthers bithecic, the four traditional “Oriental” provinces (Granma, San-
dorsifix, elliptic, c. 1.6  mm long, parallel, appressed to tiago de Cuba, Holguín, Guantánamo) has been traced,
the dorsal side of the throat, the anterior (outer) ones and it is doubtful whether the species was ever collected
overtopping the posterior by half their length and reach- there. However, the eastern half of Prov. Las Tunas (in-
ing the mouth of the corolla throat (Fig. 10A – C); thecae cluding Manatí) also belonged to the old Oriente Prov-
parallel, free in the proximal third, rounded at both ends ince, a fact that is sometimes overlooked, and it is rea-
and lacking basal awns; pollen and anthers white. Ovary sonable to assume that Alain’s mention of “Or.” is based
cylindrical, glabrous, bilocular; style filiform, glabrous, solely on Acuña’s Manatí specimen. The westward ex-
298 Greuter & Rankin: Notes on endemic species of Ruelliinae

Fig. 10. Ruellia shaferiana – A: flower in lateral view; B: corolla in front view, showing white pollen packages in the dehisced the-
cae and all around the mouth; C: flower in dorsal view; D: variation of stigmatic branches: a, normal, unequal branches, style pro-
truding from corolla bud; b, exceptional, equal branches, style protruding from calyx (corolla removed). – Photographs R. Rankin.

tension of the area of Ruellia shaferiana to the provinces but they might conceivably represent a new taxon and
of Matanzas and Habana, so far unrecorded, requires their identification as R. shaferiana is therefore tentative.
confirmation because the material on which it is based
is exceedingly poor. Neither of the two specimens cited Habitat. — Moist grassland used for grazing, near sea
bears a corolla. That of Acuña & al. is badly damaged level. At our collecting locality, the species occurs locally
by insects; it only has scant remains of two calyces and but plentiful in the shade of Bucida trees, at the margin
one half of a capsule (which is too short and was prob- of wet coastal grassland, on rich black crumbling soil,
ably less than 8-seeded, but may have been malformed). and appears to be an element of the gallery forest under-
Ekman’s gathering from Batabanó is not in much better growth.
shape; Urban (1923) did not cite it in the protologue of R.
shaferiana, and it is not certain that he even saw it (nor Seeds. — There are up to 8 seeds per capsule. They are
is there any statement of who identified it). That speci- similar in shape to those of the three other species, meas-
men deviates in having densely pubescent (not glabrous) uring 1.5 – 2 × 0.9 – 1.2 mm, but have a narrowly winged
floriferous side branches; the calyx segments are shortly margin. In the dry state (Fig. 2G) the surface is light
antrorse-strigulose ventrally and on the back and along brown, smooth and apparently glabrous throughout, as
the margins, densely covered with short eglandular and the lateral faces indeed are. When wetted, the previous-
scarcely longer, robust glandular hairs; finally the cap- ly obsolete hairs along the winged margin immediately
sule, rather than being completely glabrous and smooth, spread to form a of dense, white aureole (Fig. 2H, 3F).
is shortly patent-strigulose along the suture distally. Nei- These hairs are obliquely forward directed, 0.5 mm long
ther specimen fits any other species known from Cuba, near the base but decreasing toward the tip of the seed.
Willdenowia 40 – 2010 299

Fig. 11. Pollen of Ruellia shaferiana (Greuter & al. 27118), SEM micrographs – A: grain viewed from a presumed pole; B & D:
view upon an aperture: note the triplets of “handrails” plunging to the nexine surface at either end, the small, free, knob-headed
columellae bordering the short colpus, and the circular endoaperture; C: grain with a similar structure not linked to an aperture. –
Scale bars: A – C = 50 µm; D = 20 µm.

They are slender, collapse upon drying (Fig. 3F), and muri are tall, with a massive base distally divided like a
show weak annular to spiral wall thickenings. candelabrum into a varying number (2 – 9) of branches
(Fig. 11A); when two neighbouring candelabra fuse, a
Pollen morphology. — The pollen (Fig. 11) is spheroi- two-tiered columellar structure may result. Apertures are
dal, not distinctly polar, 3-colporate “Wabenpollen” peculiar, with very short, ill differentiated colpi almost
with (not easily discerned) apertures that may be taken coextensive with the pore. At either end of the colpus,
to define the equatorial plane. The grains (see Furness three muri (“handrails”) dip down to the nexine surface,
& Grant 1996 for standardised character definitions) are their bases becoming irregularly intertwined (Fig. 11B,
small, with an average diameter (including the c. 10 μm D; this may also, on occasions, occur where no aperture
high muri) of D = 59.6 (51.1 – 64.8) μm. No granules are is found: Fig. 11C). From there, two dense rows of free,
present. The lumina of the reticulum are medium-sized apically expanded, minute columellae, which may fuse
(15 μm across on average in the ± isodiametric lumina). into a wavy membranous structure, extend along the col-
The reticulum is subregular above the poles and inter- pus margins (Fig. 11D).
colpi, where each lumen is surrounded by 6 other lumina
to form a hexagonal pattern (Fig. 11A); the lumina ad- Hygroscopic features. — Dehiscence was triggered by
jacent to the apertures are irregularly shaped. The muri placing a drop of water on the tip of a mature capsule
are rounded and smooth. The columellae supporting the (specimen Greuter & al. 27118). As the surface is water
 300

Table 1. A summary of diagnostic features of the four species treated.

Character Dyschoriste bayatensis Apassalus cubensis and Hygrophila urquiolae Ruellia shaferiana
A. parvulus*

Habit erect (or straggling climber), decumbent for most of its length, erect, to > 1 m tall; base decumbent, ascending, to 40 cm tall; base
to > 1 m tall often rooting at nodes rooting at nodes decumbent, rooting at nodes

Calyx segments united in lower 1/3–1/2, linear- united for c. 1/3, linear- united to at least 2/3, triangular, free to the base, narrowly linear, apex
lanceolate, apex setaceous lanceolate, apex setaceous acuminate setaceous
    segment margins membranous proximally narrowly membranous at base narrowly membranous at base herbaceous
    indumentum homomorphic, eglandular, antrorse none or eglandular-scabrid, none, but segments ciliolate patent, of long gland-tipped and short
segments ciliolate eglandular hairs
    inside appressed-setulose appressed-setulose densely appressed-setulose normally glabrous
Corolla c. 14 mm long, [5 – 6]* / 9 – 10 mm long, 16 – 18 mm long, c. 14 mm long,
limb almost regular, expanded limb subregular limb strongly bilabiate limb regular, expanded
    tube 3 – 4 mm long, [c. 2.5]* / c. 4 mm long, slightly c. 6 mm long, shorter than limb c. 7 mm long, equalling limb
slightly shorter than limb shorter than limb
    limb outside densely eglandular-pubescent all minutely eglandular-pubescent densely glandular-pubescent all over glandular- and eglandular-pubescent mainly
over except at base except at base on lobes and along veins
Anther thecae awned at base rounded at base, lacking awn rounded at base, lacking awn rounded at base, lacking awn
Style strigulose-hairy subglabrous, scabrid at base glandular- and eglandular-pubescent glabrous
Stigma lacking dorsal branch dorsal branch obsolete lacking dorsal branch with ± well developed dorsal tooth or branch
Seeds, no. per capsule (3 –)4 (2 –)4 (28 –)32 (6 –)8
    wing absent absent absent membranous
    hairs on whole surface on whole surface on whole surface only along margin
Pollen grain shape subprolate subprolate* spheroidal to suboblate spheroidal
    apertures 3-colporate, 3-colporate, 4-colporate, 3-colporate
with atrium formation with atrium formation* without atrium formation
    exine structures 4 longitudinal ribs 4 longitudinal ribs 4 longitudinal ribs reticulate muri
(“Rippenpollen”) (“Rippenpollen”)* (“Rippenpollen”) (“Wabenpollen”)
Greuter & Rankin: Notes on endemic species of Ruelliinae
Willdenowia 40 – 2010 301

repellent, the imbibing process took considerable time Upon wetting, the hairs tear off at the base and are carried
(c. 15 min), but then the capsule exploded forcefully, pro- upward by the unfolding mucilage string, which is up to
jecting some of the seeds to a distance of over 30 cm. In 8 times as long and twice the diameter of the hair itself.
this species, too, the calyx shows hygroscopic properties, Our own observations do not bear out any such scenario.
even though they are scarcely noticeable. After the shed- We are unconvinced of the hypothesis that the muci-
ding of the capsule, articulated at the base, the formerly lage comes from within the hairs and rather doubt that
erect calyx segments spread from the very base, then the hairs have any content of note. A much more logical
curve upward to an upright distal portion. In water the assumption, that would better explain the observed phe-
bend is completely straightened out, but upon renewed nomena, is that the glue or lacquer that initially fastens
drying it is recovered and more pronounced than before. down the hairs and the mucilage released in water are one
The hygroscopic portion is limited to the proximal fifth and the same substance.
of the calyx. The main function of the mucilage, which dries out
as a sticky film well visible in SEM images (e.g. in Fig.
Pollination-relevant aspects. — When, in the field, we 3C, D), is to attach the seed firmly to the substrate so as to
found only a few withered flowers but plentiful buds provide a hold during germination. Only exceptionally,
and fruits, we decided to monitor the flowering rhythm when the substrate happens to be the claws or plumage of
of freshly cut material. We had to wake early to do so: waterfowl, this may be an effective way of long-distance
unfolding of the corolla started before dawn, at 5 a.m., dispersal. The hairs themselves serve as the seed’s anchor
and was complete by sunrise. By noon the corollas, which in the glue, which holds them fast thanks to their annu-
from bright blue had turned pale burgundy, were withering lar or spiral surface pattern, and for this to function they
and being shed. The flowers are scentless and lack nectar, must of course remain firmly attached to the seed coat.
so that colour is the only attractant and pollen the only Indeed, we never observed loose hairs floating about af-
possible reward. Indeed the white pollen, partly sticking ter immersion.
to the thecae and partly scattered in clusters around the Mucilaginous seed hairs of the kind here described
mouth of the corolla throat (Fig. 10B), enhances colour (and presumably homologous mucilage cells mentioned
attraction, but its small quantity makes it improbable that above under Ruellia) are arguably a so far neglected
it could serve as a reward for pollen harvesting pollina- synapomorphy for the subtribe Ruelliinae. True, excep-
tors. Are we, then, facing a case of pollination by deceit? tions apparently exist. Seeds with hairs lacking mucilage
Self-pollination may also play a role, but only at the end, have been reported for a few representatives of Ruelli­
when the corolla is falling off; before that time, we did not inae, including Eranthemum and at least two species of
observe any pollen deposited on the stigma, in spite of its Strobilanthes Blume, S. fimbriata Nees and S. wallichii
being placed in close proximity to the anthers. Nees (Kippist 1842); the seeds of Bravaisia are described
as glabrous, smooth and shiny (Daniel 1988). Converse-
ly, mucilaginous seed hairs or papillae exist in Acan­
Discussion and conclusions
thaceae, outside of Ruelliinae, at least in Blepharis (incl.
Seeds. — Seed number, in the species studied, ranges Acanthodium Delile), where they are of a different type:
from 4 to 32 and is species-specific, allowing for slight fasciculate, compound of many concrescent cells with a
downward variation due to abortion. All seeds are strong- free, filiform distal portion resembling ruelliinid single
ly lenticular-compressed. As already noted by Oehm hairs (Kippist 1842: fig. 1 – 5, Schaffnit 1906: 462 – 463).
(1932: fig. 9–11), outline shape varies in function of the Barleria and Crabbea (Barleriinae) were said by Stopp
space available during maturation, which differs depend- (1958) to have mucilaginous, sticky seed hairs, but Kip-
ing on the position of each seed within the capsule. By pist (1842: 71) stated that no mucilage is secreted in the
consequence, seed shape is not suitable as a taxonomic 9 Barleria species he studied, and Schaffnit (1906: 462)
character. confirmed this for B. cristata L. We suggest that a com-
As soon as their hairs unfold in water, a smeary cloud parative study of seed hair types of Acanthaceae and of
of mucilaginous matter forms around the seeds (Fig. 2B, their distribution in a modern, phylogenetically based
2D, F, H). That cloud appears almost instantaneously, classification of the family would be a rewarding subject
and surrounds the hairs in the first place. Kippist (1842: for future research.
69) reports that the mucilage “may be seen to flow from From a practical point of view, we suggest that the
the extremity of the tube”, or “from a terminal pore”, and substance initially holding together the ruelliinid seed
that the hairs “are frequently detached from the testa by hairs, whether or not it be the same as the sticky muci-
the rapid expansion of the fluid within, and numbers of lage released in water, may present properties interest-
them may be seen floating about loose”. Schaffnit (1906: ing in the context of bionics. Adhesiveness coupled with
459) gives a detailed description based principally on Hy­ instant water solubility might be useful in structures to
grophila salicifolia (Vahl) Nees. According to him, the be set free, and processes to be triggered, upon contact
hairs are initially filled with cellulose mucilage, which at with water. The mucilage may also present useful if less
maturity is seen as a vitreous, crumpled gelatine strand. spectacular qualities as a fast and extremely elastic glue.
302 Greuter & Rankin: Notes on endemic species of Ruelliinae

Pollen morphology. — The pollen of Dyschoriste baya­ ing up in the calyx segments at maturation, aimed at
tensis, Hygrophila urquiolae and Ruellia shaferiana fits keeping together the capsule halves at their base. With-
perfectly within the general pollen morphological pat- out such a firm hold the capsule, clamped together by
tern characterising their respective genera and confirms the beak and with tension building up along its valves,
that the current generic placement of these species, based risks to implode before the explosion can be triggered.
on gross morphology, is correct. Apassalus parvulus We have indeed observed imploded capsules (with the
has pollen of the Dyschoriste type, and its study has not valves still united at the tip but spreading basally) in
produced support for the maintenance of Apassalus as a herbarium specimens of Apassalus cubensis and Ruel­
separate genus. lia shaferiana. In Hygrophila urquiolae, where the ca-
lyx is not hygroscopic, its long tubular portion is firm
Hygroscopic features. — In all four species examined, enough to prevent implosion, but not in the other three
the capsules are hygrochastic. One is initially tempted to species. Whereas hygroscopic movements of bracts are
see a contradiction between this result and the fact that well documented in the family (reported by Sell 1969
during the drying of fruiting specimens for the herbar- for Barleria, Blepharis and Crabbea species) and their
ium several capsules open and disperse their seeds. The function is easily explained (exposure of the capsules to
explanation is easy: in the early stages of drying, espe- rain), we have found no published mention of the calyx
cially when some heat is used, moisture accumulates movements here described.
within the plant press and can trigger the release. In
Cuba, the explosive capsules of Ruellia tuberosa and its Dyschoriste and Apassalus. — As we mentioned initial-
congeners are well known among local people, as docu- ly, opinions are currently split on the question whether
mented by the vernacular names “saltaperico”, “dinami- these two genera should be maintained or better merged.
ta”, “fulminante” and “triquitraque” (Roig y Mesa 1963). The single difference of presence versus absence of a
It is a popular pastime to wet a capsule in the mouth and basal mucro in the anther thecae is unconvincing, even
watch it explode. more so as this character varies within Hygrophila
Sell (1969) reports that only one fifth of the Acan­ (Wasshausen 2005). Throughout our study we have
thaceae species he studied (5 out of 25) have hygrochas- therefore been looking for additional differences sup-
tic capsules. This figure is surprisingly low. In a group porting the segregation of Apassalus, but failed to find
such as the Ruelliinae, where the seeds must be wet for any. On the contrary: we observed that the calyx of A.
their adhesive mechanisms to function, one might have cubensis is very similar to that of D. bayatensis, not only
expected hygrochasy to be favoured (as it indeed is on in morphology but also in its hygroscopic properties. We
Cuba). But then, there may be no cogent reason why the therefore come to the conclusion that the merger of Dys­
seeds should become fixed immediately upon release, as choriste and Apassalus is unavoidable.
long as it is ensured that fixation precedes germination. The distinctness of the plants described as Apassa­
At any rate, the Acanthaceae appear to be playing a dual lus parvulus became questionable when material subse-
game, as exemplified by Crossandra nilotica Oliv., in quently collected close to its locus classicus was studied,
which half of the capsules open xerochastically and the which in general aspect, indumentum, and the dimen-
other half hygrochastically, depending on whether the in- sions of its vegetative parts was indistinguishable from
itial tension that builds up in drying suffices to break the A. cubensis. However, the two E Cuban gatherings share
apical clamp or whether it must be weakened by moisture some features that can be used to distinguish them from
(Sell 1969: 434 – 435). If, as in this case, there is a subtle the Central and W Cuban populations: the short calyx
balance between both strategies, that is tilted by slight (5 – 6 versus 6 – 9 mm), corolla (5 – 6 versus 9 – 10 mm), and
quantitative, non-structural changes, it is safe to assume capsule (c. 5 versus 7 – 8 mm), and moreover, the hygro-
that the evolutionary step from xerochasy to hygrochasy scopic incurving of the old calyx segments, much more
and back is an easy one that may frequently happen. pronounced in A. parvulus than in A. cubensis. In view
The hygroscopic movements of the calyx segments of the discontinuous variation and geographical isolation,
that we have observed in three of our species are less we opt for maintaining A. parvulus as a separate species,
easily interpreted. They appear to be associated with a for which, when placed in Dyschoriste, a new combina-
hardening of certain (proximal or medial) portions of the tion is required. The correct names of the two Cuban spe-
calyx segments at the end of the maturation process, and cies formerly placed in Apassalus as follows:
can be observed once the void capsules have been shed.
Dyschoriste cubensis Urb., Symb. Antill. 7: 381. 1912 ≡
There is an abscission layer at the base of the capsules,
Apassalus cubensis (Urb.) Kobuski in Ann. Missouri Bot.
from where they disarticulate (Fig. 6C), apparently a
Gard. 15: 2. 1928.
general though little noticed feature in these genera.
We can see no conceivable functional advantage in the Dyschoriste parvula (Alain & Leonard) Greuter & R.
hygroscopic movements of the segments of the empty Rankin, comb. nov. ≡ Apassalus parvulus Alain & Leo­
old calyces. The most likely interpretation, to us, is that nard in Contr. Ocas. Mus. Hist. Nat. Colegio “De La
these movements are the by-product of pressure build- Salle” 15: 20. 1956.
Willdenowia 40 – 2010 303

Acknowledgements Furness C. A. & Grant M. C. 1996: Pollen morphology of

We acknowledge the help of the keepers of the herbaria some Ruellia species (Acanthaceae) from Africa and
consulted by giving us access to their collections, send- Madagascar. – Grana 35: 231 – 239. [CrossRef]
ing digital specimen images, or lending actual speci- Greuter W., Rankin Rodríguez R. & Palmarola A. 2010:
mens for study. We are particularly indebted to Monika Hygrophila urquiolae (Acanthaceae), a new wetland
Lüchow who produced the incident-light micrographs species from Cuba. – Willdenowia 39: 285 – 291.
and, partly assisted by Christine Grüber, skilfully pre- Hilsebeck R. A. 1976: Reports [in Löve Á. (ed.), IOPB
pared and photographed the SEM seed and pollen sam- chromosome number reports lii]. – Taxon 25: 345.
ples. Hermann Manitz of Jena kindly, once again, helped Kippist R. 1842: On the existence of spiral cells in the
by providing rare items of literature. Thomas Borsch, seeds of Acanthaceae. – Trans. Linn. Soc. London 19:
Director of the Berlin-Dahlem Museum, we thank for 65 – 76. [CrossRef]
granting us liberal access to the institution’s splendid Kobuski C. E. 1928a: A new genus of the Acanthaceae.
working facilities. – Ann. Missouri Bot. Gard. 15: 1 – 8. [CrossRef]
Kobuski C. E. 1928b: A monograph of the American spe-
cies of the genus Dyschoriste. – Ann. Missouri Bot.
References Gard. 15: 9 – 90. [CrossRef]
Alain bro. 1956: Novedades en la flora cubana (VII). – Leuenberger B. E. 1976: Die Pollenmorphologie der Cac­
Contr. Ocas. Mus. Hist. Nat. Colegio “De La Salle” taceae und ihre Bedeutung für die Systematik: mit
15: 6 – 23. Beiträgen zur Methodik rasterelektronenmikroskopis-
Alain bro. 1957: Flora de Cuba 4. – Contr. Ocas. Mus. cher Pollenuntersuchungen und mit EDV-unterstütz-
Hist. Nat. Colegio “De La Salle” 16. ten Korrelationsanalysen. – Diss. Bot. 31.
Alain bro. [Liogier H. A.] 1995: La flora de la Española Lindau G. 1893: Beiträge zur Systematik der Acantha-
7. – San Pedro de Macorís: Universidad Central del ceen. – Bot. Jahrb. Syst. 18: 36 –  64.
Este. Lindau G. 1895: Acanthaceae – Pp. 274 – 354 in: Engler
Anonymous 2010: Index to plant chromosome numbers. A. & Prantl K. (ed.), Die natürlichen Pflanzenfamil-
– Missouri Botanical Garden, published at http:// ien nebst ihren Gattungen und wichtigeren Arten, ins-
mobot.mobot.org/W3T/Search/ipcn.html [consulted besondere den Nutzpflanzen 4(3b). – Leipzig: Engel-
June 2010]. mann.
Bentham G. 1876: Ordo CXXII. Acanthaceae. – Pp. Long R. W. 1970: The genera of Acanthaceae in the
1060 – 1122 in: Bentham G. & Hooker J. D., Genera southeastern United States. – J. Arnold Arbor. 51:
plantarum ad exemplaria imprimis in herbariis ke- 257 – 309.
wensibus servata definita 2. – London: Reeve & Co. Long R. W. 1973: A biosystematic approach to generic
Bremekamp C. E. B. 1926: On the opening mechanism delimitation in Ruellia (Acanthaceae). – Taxon 22:
of the acanthaceous fruit. – S. African J. Sci. 23: 543 – 555. [CrossRef]
488 – 491. Mabberley D. J. 1987: The plant book: a portable diction-
Bremekamp C. E. B. 1965: Delimitation and subdivi- ary of the higher plants. – Cambridge (U.K.): Cam-
sion of the Acanthaceae. – Bull. Bot. Surv. India. bridge University.
7: 21 – 20. Mabberley D. J. 2008: Mabberley’s plant book: a porta-
Daniel T. F. 1998: Pollen morphology of Mexican Acan­ ble dictionary of the higher plants, ed. 3. – Cambridge
thaceae: diversity and systematic significance. – (U.K.): Cambridge University.
Proc. Calif. Acad. Sci. 50: 217 – 256. McDade L. A. & Tripp E. A. 2007: A synopsis of Costa
Erdtman G. 1960: The acetolysis method – a revised de- Rican Ruellia (Acanthaceae), with descriptions of
scription. – Svensk Bot. Tidskr. 54: 561 – 564. four new species. – Brittonia 59: 199 – 216.
Ezcurra C. 1993: Systematics of Ruellia (Acanthaceae) [CrossRef]
in southern South America. – Ann. Missouri Bot. Nees C. G. 1847: Ordo CXLV. Acanthaceae. – Pp.
Gard. 80: 787 – 845. [CrossRef] 46 – 521, 720 – 732 in: Candolle A. L. P. P. de (ed.),
Ezcurra C. & Daniel T. F. 2007: Ruellia simplex, an old- Prodromus systematis naturalis regni vegetabilis 11.
er and overlooked name for Ruellia tweediana and – Paris: Treuttel & Würtz.
Ruellia coerulea (Acanthaceae). – Darwiniana 45: Oehm G. 1932: Beitrag zur Morphologie und Anatomie
201 – 203. einiger Acanthaceen-Früchte und -Samen. – Beih.
Furness C. A. 1994: The pollen morphology of Hygrophi­ Bot. Centralbl., Abt. 1, 49: 413 – 444.
la and Brillantaisia (Acanthaceae: Ruellieae). – Acta Radlkofer L. 1883: Über den systematischen Werth der
Bot. Gall. 141: 267 – 278. Pollen-Beschaffenheit bei den Acanthaceen. – Sit-
Furness C. A. 1995: A pollen morphological study of Dy­ zungsber. Math.-Phys. Cl. Königl. Bayer. Akad. Wiss.
schoriste Nees and Chaetacanthus Nees (Acanthaceae: München 13: 256 – 314.
Ruellieae). – Rev. Palaeobot. Palynol. 84: 331 – 345. Raj B. 1961: Pollen morphological studies in the Acan­
[CrossRef] thaceae. – Grana Palynol. 3: 3 – 108.
 304 Greuter & Rankin: Notes on endemic species of Ruelliinae

Roig y Mesa J. T. 1963: Diccionario botánico de nombres lanae seu fundamenta florae Indiae occidentalis 7. –
vulgares cubanos, ed. 3. – Santiago de las Vegas: Es- Leipzig: Borntraeger.
tacíon Experimental Agronómica. Urban I. 1923: II. Plantae cubenses novae vel rariores a
Schaffnit E. 1906: Beiträge zur Anatomie der Acan- clo. Er. L. Ekman lectae I. – Pp. 59 – 176 in: Urban
thaceen-Samen. – Beih. Bot. Centralbl., Abt. 1, 19: I. (ed.), Symbolae antillanae seu fundamenta florae
453 – 521. Indiae occidentalis 9. – Leipzig: Borntraeger.
Scotland R. W. 1993: Pollen morphology of Contortae Ward D. B. 2007: Keys to the flora of Florida – 17, Ruel­
(Acanthaceae). – Bot. J. Linn. Soc. 111: 471 – 504. lia (Acanthaceae). – Phytologia 89: 293 – 299.
Scotland R. W. & Vollesen K. 2000: Classification of Wasshausen D. C. 2005: 156. Acanthaceae incl. Thun­
Acanthaceae. – Kew Bull. 55: 513 – 589. [CrossRef] bergiaceae. – Pp. 1 – 141 in: Görts van Rijn A. R. A.
Sell Y. 1969: La dissémination des acanthacées. Varia- (ed.), Flora of the Guianas. Series A. Phanerogams.
tions sur le type xérochasique fondamental. – Rev. 23. – Kew: Royal Botanic Gardens.
Gén. Bot. 76: 417 – 453. Wasshausen D. C. 2010: Acanthaceae [in: Flora of the
Stopp K. 1958: Die verbreitungshemmenden Einrichtun- West Indies. Catalogue of seed plants of the West
gen in der südafrikanischen Flora. – Bot. Stud. 8. Indies. Plants of Cuba]. – Smithsonian Institution,
Tripp E. A. 2007: Evolutionary relationships within the Washington, published at http://botany.si.edu/Anti-
species-rich genus Ruellia (Acanthaceae). – Syst. lles/WestIndies/results.cfm?formid=cu
Bot. 32: 628 – 649. [CrossRef] Wasshausen D. C. & Wood J. R. I. 2004: Acanthaceae of
Tripp E. A., Daniel T. F., Lendemer J. C. & McDade L. Bolivia. – Contr. U. S. Natl. Herb. 49.
A. 2009: New molecular and morphological insights Wunderlin R. P. & Hansen B. F. 2010: Atlas of Florida
prompt transfer of Blechum to Ruellia (Acanthaceae). vascular plants. – Institute for Systematic Botany,
– Taxon 58: 893 – 906. University of South Florida, Tampa, published at
Urban I. 1912: Acanthaceae [in: Nova genera et species http://florida.plantatlas.usf.edu/
V]. – Pp. 380 – 391 in: Urban I. (ed.), Symbolae antil-

View publication stats