Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Trigoniaceae
Author(s): Eduardo Lleras
Source: Flora Neotropica, Vol. 19, Trigoniaceae (Apr. 28, 1978), pp. 1-73
Published by: New York Botanical Garden Press on behalf of Organization for Flora
Neotropica
Stable URL: http://www.jstor.org/stable/4393715
Accessed: 21/12/2008 12:34
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=nybg.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.
New York Botanical Garden Press and Organization for Flora Neotropica are collaborating with JSTOR to
digitize, preserve and extend access to Flora Neotropica.
http://www.jstor.org
FLORA NEOTROPIC
MONOGRAPHNO. 19
TRIGONIACEAE
by
Eduardo Lleras
-(\' -
?
(~ * " ?~ -- --- - ----- --- - ---------
/ / C f
~TROPIC 0O CANCRt
FLORA
NEOTROPICAI
TrlOPIC OF CAPRICORN
Published for
by
The New York Botanical Garden
New York
1978
Publishedby
The New York BotanicalGarden
Bronx, New York 10458
All material subject to this copyright may be photocopied for the non-commercial purpose
of scientific or educational advancement.
MONOGRAPHOF THE FAMILYTRIGONIACEAEI
EDUARDOLLERAS2
INTRODUCTION
TAXONOMICHISTORY
althoughit has been treatedin severallocal revisions,in all cases includingonly one
genus. The only importantoccurrencebetween 1897 and the presentwas Leandri's
(1949) descriptionof Humbertiodendron(Humbertodendron)from Madagascar,
addinganothercontinent to the known distributionof the family.
Trigoniawas treated by Standley (1924) for North AmericanFlora, by
Stafleu (1951) for the Flora of Suriname,by Reitz (1967) for Flora Iustrada
Catarinense,and most recently by Austin (1968) for the Flora of Panama.
Trigoniastrumwas revisedby Van Steenis (1949) for FloraMalesianaand by
Ng (1972) in TreeFlora of Malaya.Apartfrom the originalpublicationin 1949,
Humbertiodendronwas studied by Perrierand Leandri(1955) for the Flora de
Madagascar,in which they changedthe orthographyto Humbertodendron.This
second name is acceptedhere as the correctorthographicvariation,based on the
recommendationof the InternationalCode of BotanicalNomenclature(Stafleu,
et al 1972).
GEOGRAPHYAND ECOLOGY
_3
. ..
FIG 1. Geographic distribution of Trigoniaceae. , Range of the genus Trigonia. 2, Range of the
genusHumbertodendron(Humbertodendron
saboureaui).3, Rangeof the genus Trigoniastrum
(Trigoniastrum
hypoleucum).
and the fruits were probably capsules with severalvillous seeds. (The seeds of
Trigoniastrumand Humbertodendronare relictuallyvillous, but the trichomes
serveno apparentfunction, for in Trigoniastrumthe seeds are reportedas germinat-
ing in situ by Ng, 1972.) It is probablethat the pubescenceon the seeds of the
ancestralstock was not as functional for wind dispersalas that found in present
Trigonia;it could have been much shorterand still have providedadequatedispersal
for tall trees (acting as a parachutein the same manneras the fruit of the paleo-
tropicalgenera).Witha progressivechangein habit from tall trees to lower shrubs,a
more effective dispersalmechanismwas probablyselected for in the ancestryof
Trigonia.The leaves of the ancestralstock were probablyopposite, with the alter-
nate leavesof Trigoniastrumrepresentingthe derivedcondition. This can be postu-
lated, based on the presenceof opposite leavesin TrigoniaandHumbertodendron,
and suggestsan ancestralstock for these generawith opposite leaves, as the neo-
tropicaland paleotropicalmembersof Trigoniaceaewere probablyseparatedbefore
the ancestorsof Trigoniastrumand Humbertodendron.The arborescentcondition
in Trigoniastrumand Humbertodendronis probably relictual, and the shrubby,
vininghabit of Trigoniaderived.
Although membersof the Trigoniaceaeand especially Trigoniaare fairly
common in some areas,the groupis, with some exceptions, a poorly collected one.
Thereis probablyno one reason for this; Trigoniaas a whole is found close to river
banks in periodically flooded forests, and in drier areas along gallery forests. In
general,species representedin galleryforests are collected more often than those
found along periodicallyflooded riveredges. These collecting practicescould be
GeneralConsiderations 5
00 .00 .....0
/ N
.'
FIG 2. Forest refuges proposed by Prance (1973). 1. Choc6; 2. Nechi; 3. Santa Marta; 4.
Catatumbo;5. Rancho Grande;6 Paria; 7. Imataca; 8. Guiana; 9. Imeri; 10. Napo; 11. Olivenca;
12. 'Tefe; 13. Manaus; 14. East Peru; 15. Rond6nia-Aripuana; 16. Belem-Xingu. The areas
surrounded by a heavy black line correspond to areas flooded during the pleistocene.
0,
80 70 60 50 40
ANATOMICALCONSIDERATIONS
80 0
,
it"..
pubescens; o T...ni.......ci.l..
FIG 7. Known distribution of species of Trigonia. * T. nivea var nivea; (3T. nivea var
pubescens; 0 T. nivea var fasciculata.
Mesophyllisis with
with branchedsclereids
sclereids in Trigoniastrumand Hu
Trigoniastrum
in Humbertodendron;
and
rbertodendron;
palisadeparenchymais usually l-layered but sometimes2-layeredin Trigoniastrum
and Humbertodendron;rhomboidor irregularlychamberedcrystalsof calciumare
presentinand
niastrum Trigoniastrum, Humbertodendron
other Trigonia and several speciesisofsimple
Trigonia.
species. In the
The petiole has a simple stem,
all
the epidermis
three
and often
epidermis in genera;the trichomes,when
present,are simple and unicellular.The phloem completely surroundsthe xylem in
Humbertodendron and some species of Trigonia, or is open at the top as in Trigo-
has tannins.The cork i from subepidermalphellogen,and the cork cells
species
has tannins.The cork is derivedfrom subepidermalphellogen,and the cork cells
14 Flora Neotropica
often have tannins, and the regionjust below the peridermisis often subcollen-
chymatous. Corticalsclereids are often present and are especially common in
Trigoniastrumand Humbertodendron.Calciumoxalate crystalsare frequentin all
three genera (lacking in some species of Trigonia).The secondaryphloem has
sclereids. Intraxylaryphloem is absent, and the vessels are solitary (although
Heimsch1942, reportsporemultiplesin some speciesof Trigonia).The perforations
are simple and elliptic (sometimes double in T. eriosperma,fide Barth).The paren-
chyma is apotrachealin all three genera(1 specimenof Trigoniastrumreportedas
havingparatrachealparenchymaby Metcalfe& Chalk). The rays are 1-5-seriate,
heterogeneous(Kribs' type IIa and IIb). The wood is diffuse-porous.The pith
often has tannins,with or without crystals;in Humbertodendron,rhomboidcrystals
arevery common.
Cross sections and dissections of the flower of Trigoniaspruceanaand T.
prancei were made in order to determinethe nature of the floral glands. The
vascularizationof the glandsis the same as that found in the staminalring,with
2-3 vascularstrandsattachedto (ie entering)each gland.This supportsmy views
based on morphologythat the glandsare actuallymodified stamens.
Anatomically,the familyis fairlyuniform,with no single characterfacilitating
differentiationamongthe genera,althoughthere are trendsthat are more noticeable
in some groupsthan in others.
Humbertodendronis anatomicallyvery close to Trigoniastrum; in fact, the
anatomy of the leaf blade is so similarin the two generathat laminarmaterial of
the two generaeasily can be confused in the laboratory.The petioles are somewhat
different,however,especiallyin the position of the vascularbundles.As indicated
in the anatomicaldescription,the phloem in Humbertodendrontotally surrounds
the xylem, but is open in Trigoniastrum.In Humbertodendronsclereidsare present
only alongthe rays,while they are randomlydistributedin Trigoniastrum.Although
rhomboidcrystalsoccur in the pith in both genera,they are much more abundant
in Humbertodendron.
Barth(1896) consideredthe Trigoniaceaeas relatedto the Dichapetalaceae
on anatomicalgrounds.Heimsch(1942), based on his work on secondaryxylem,
consideredthe Trigoniaceaeclose to the Polygalaceae,the Tremandraceae,the
Zygophyllaceae,the Malpighiaceaeand the Vochysiaceae.Metcalfe and Chalk
(1951) consideredthe family closest to the Vochysiaceae.
The Trigoniaceaediffers from the Vochysiaceaein the absenceof intraxylary
phloem, the absenceof libriformfibers, and in havingapotracheal(vs paratracheal)
parenchyma.
Anatomically, the Trigoniaceaehave undoubtedly close affinity to the
Vochysiaceaeas well as the Polygalaceae.The suggestedrelationshipswith Dicha-
petalaceaemay also be correct. The taxonomic positions and familialrelationships
of the familiesin both Cronquist's(1968) Sapindalesand Polygalales,as well as the
relationshipof these familieswith the ancestralRosales,remainratherunclearat
present.
On the basis of anatomy, which corroboratesthe morphologicalobservations,
I agree with Barth that Euphroniashould be removed from the Trigoniaceae.
Anatomically,thereis no seriousobjectionto returnEuphroniato the Vochysiaceae,
with which it has many charactersin common. The data used to support this
decision are discussedelsewhere(Lleras, 1976).
GeneralConsiderations 15
'4 S7l
t: "'. '
~ ...*,
g,* 4
A B
Wi 35
.
eIf'j : '-t?
'I,r~ *...
' ?
' 9.. ? .
E F
FIG 8. Anatomy of Humbertodendron saboureaui (Perrier & Louvel 14896bisJ. A, young stem,
longitudinal section X 9.2;B, young stem, longitudinal section showing rhomboid crystals X 34;
C, young stem, cross section X 9.2;D, petiole, cross section X 18; E, Lamina, cross section X
66; F, lamina
la :.inacentralnervesX 18.
16 Flora Neotropica
POLLEN
A B
E F
ovariesof the last two species, by the time the fruit reachesmaturityno trace of
the lateralsepta remains.
Evidenceof a gradualreductionof the androeciumcan be seen in livingspecies
of the Trigoniaceae.The most primitiveexisting condition occurs in Trigonia
macranthaand T. nivea, whose flowershave fairlylargenumbersof stamens(ca 12),
some of which are reducedto staminodes.The intermediatecondition is found in
species in which the staminodeshave been lost, and all of the stamensare fertile,
as in T. virens.The condition which I considerto be the most evolved, where there
areno staminodesand a relativelysmallnumberof stamens(ca 6), is presentin both
Trigoniastrumand Humbertodendron.
If we acceptan ancestorin the Rosales(as discussedundersystematicposition
of the Trigoniaceae)as the precursorfor Trigoniaceae,it was probablya groupof
plants with a largenumberof indefinite or definite stamensdisposedin a staminal
ring.The reductionprocessprobablyoccurredas follows: first, a complete reduc-
tion of a whole section of the staminalring,with subsequentevolution of the disc
glands. I interpret these glands as reduced and modified stamens, based on the
laciniaewhich are morphologicallyvestigialfilaments;anatomicalevidence for this
is discussedin the section on anatomy. Next there was a loss of antherson some of
the remaining(outer) stamensof the portion that remained;and finally, a loss of
staminodes,giving rise to the extant taxa displayingfew fertile stamens and no
staminodes.
The obvious differencein the dispersalmechanismbetween the neotropical
and paleotropicalTrigoniaceaerepresentsan importantevolutionarychange. In
both Humbertodendronand Trigoniastrum,the seeds remainin the winged fruit,
and are dispersedwith it. In Trigonia,however,the fruits are dehiscent,and the
seeds are adaptedto wind and water dispersal.This transferenceof function for
seed protection and dispersalfrom the seed to the ovary and vice-versais common
in the Angiosperms(Stebbins, 1974).
As is discussedin the section on geographyand ecology, and based on geo-
graphicaland ecologicalevidence, the plants of the ancestraltrigoniaceousstock
were probablytrees, with many-seededcapsules.The seeds were probablyvillous,
although the trichomes were probably shorter than those found on present-day
species of Trigonia.
Figure 10 is a schematic representationof my opinion of similaritiesand
differencesin Trigonia.The distancesdo not representcloseness of relationshipand
only serveto separatetaxa of dubiousposition. It is not my intention to suggest
phylogeny, as I believedthat presentevidenceis insufficient to do so. It is also
difficult to determinewhich species are primitiveor advanced,as in almost every
case a species that presents a characterthat I consider primitivealso presents
charactersthat I consideradvanced.
SYSTEMATICPOSITIONOF THETRIGONIACEAE
,iiE,~i
and some petals are very similar(almost identical) to those of the Old Worldgenera
of the latter family. Another significantcharacteris the occurrenceof winged
samarasin all three families.
TableI presentsa comparisonof morphologicalcharactersof the Trigoniaceae,
the Vochysiaceae, the Polygalaceae,the Sapindaceaeand the Malpighiaceae.A
comparisonof charactersin the table is indicativeof the difficultiesencountered
when trying to establishrelationshipsbased on morphologicaldata.
The evidencecited above supportsCronquist(1968) who accepted a general
relationshipamongthese families,which he believesare derivedfrom the common
Rosalean stock, but I am not entirely convinced that the degree of relationship
amongthe familiesis clearat present.Thereis no doubt that the familiesincluded
in the Polygalalesare related,and are close to the Sapindales(Cronquist,1968) and
possibly the Celastrales(Prance, 1972). However,the relationshipsamong the
familiesin these three ordersare complex. Indicationsfor this have been givenby
Prance(1972), and Breteler(1973) on morphologicalgrounds,and anatomical
evidencehas been providedby Barth(1896) and Heimsch(1942).
The evidencepresentedabove supportsa common originin the Rosalesfor
the Polygalales,Celastralesand Sapindalesas has been suggestedby Prance(1972).
This would explain a numberof familiesthat seem to have affinitiesin all three
orders.Thus, the Trigoniaceaetogetherwith the Vochysiaceaeand the Dichapetal-
aceae and probablythe Malpighiaceae(fide Breteler,1973) would occupy positions
close to the point of divergenceof the three orders.
The Trigoniaceaehave (as have the other three families)evolvedin a direction
of their own, and probablyarrivedat the papilionaceoustype of flower independ-
ently. This suggeststhat this distinctiveflower type has evolved at least three times
in responseto insect pollinators(the Fabaceae,the Polygalaceaeand the Trigoni-
aceae). At presentit is difficult to establishthe closest relativesof the Trigoniaceae,
but it would be misleadingto give undue emphasisto the papilionaceouscorolla.
It is significantthat the relativessuggestedfor the Trigoniaceaeall belong to
familiesthat have significantlymodified corollas(Breteler,1973). This suggestsa
highly adaptablegroupin relationto the corolla, anotherfactor that would suggest
a common ancestry.
Taxonomically, I maintainTrigoniaceaein the Polygalalesas defined by
Cronquist(1968), as most of the possible relativessuggestedin this paper are
includedin that order.
TAXONOMICPOSITIONOF EUPHRONIA
number of sepals 5 5 5 5
staminodes + (-) + -
glands + + + (ext) +
endosperm - - -
SPECIFICCONCEPTS
TO THESYSTEMATICTREATMENT
INTRODUCTORYCOMMENTS
The following brief comments are given to aid the user of the key and to
clarify the descriptions.
Since sterile materialis not adequatelyrepresentedin collections, and sterile
shoots and vegetative branchesoften have much largerleaves, leaf size measure-
26 Flora Neotropica
' '
?.
.. , :.. - . ,'? .i'.'..
"... ?: .
SYSTEMATICTREATMENT
2. Fruit a three-winged samara; ovary with lateral ridges; ovules one per locule;
Madagascar. 2. Humbertodendron.
1. Leaves alternate; Malaysia. 3. Trigoniastrum.
Frutex, ramulisjuvenilibustomentellis,glabrescentibus,lenticellatis.Stipulae
caducae,ca 5.0 mm longae, ad basinconnatae. Folia opposita,petiolo 8.0-15.0 mm
longo, 0.8-1.5 mm crasso, leviter tomentello; laminae ellipticae vel obovatae,
6.0-11.0 cm longae, 3.0-5.8 cm latae, subcoriaceae,margineintegrae,apice acuto
34 Flora Neotropica
Vs.2
F G
D E
J K L
FIG 12. TrigoniarotundifoliaA, habit X0.35; B, crosssection of ovaryXI 3.2; C, flower X3.3; D,
sepalsX3.3; E, ovaryX3.3; F, stamensX3.3; G, anthersX6.6; H, glandsX6.6; J, keel petal X3.3;
K, wingpetal X3.3; L, standardX3.3.
SystematicTreatment 35
J?
FIG 13. TrigoniapranceiA, habit X0.35; B, flower X3.3; C, ovaryX3.3; D, keel petal X3.3; E,
wingpetal X3.3; F, standardX3.3; G, stamensX3.3; H, anthersX6.6; J, crosssection of ovaryX13.2.
36 Flora Neotropica
6. Trigonialaevis Aublet, Hist. PI. Guian.Fr. 1: 390 pl. 150. 1775; de Candolle,
Prodr.1: 571. 1824; Warming,(Trigoniaceae)Mart.Fl. Bras. 13(2): 131. 1875.
6a. Trigonialaevisvarlaevis
Flowersin groupsof 1-3; petals 2.5-3.5 mm long. Fruit oblong, 1.5-3.0 cm
long, ca 1.0 cm wide; seeds usually3-4 per locule.
Type. Aublet sn, French Guiana,Cayenne, fl, fr (holotype BM, isotypes
F,P).
Distribution.Known only from a few collections in FrenchGuiana,and one
collection (type of Trigoniakaieteurensis)from Guyana.
FRENCH GUIANA. Gabriel 1802 fl, fr (G); Soubirou sn fl, fr (P).
Withoutfruitingmaterial,it is practicallyimpossibleto separatethis variety
from varmicrocarpa.It is quite possible that some materialplaced by me in var
microcarpamight actually be varlaevis.
6b. Trigonialaevisvar microcarpaSagot, Ann. Sci. Nat. 6(2): 176. 1881.
8. TrigoniareticulataLleras,sp nov
Frutex scandens,ramulisjuvenilibustomentellis,glabrescentibus,lenticellatis.
Stipulascaducae,haud visae. Folia opposita;petiolus 9.0-18.0 mm longus, 1.0-3.0
mm crassus,tomentellus;laminaeovataevel ellipticae,interdumobovatae,4.5-11.0
cm longae, 3.0-6.5 cm latae, subcoriaceae,margineleviter revolutae,apice acuto
usque acuminatovariante,basi obliquavel subrotundata,supraglabrae,infralanato-
tomentellae;costa media supraplana,infra prominens,strigulosa;costis secundariis
(8-)9-11(-12)-jugis,costis tertiariisquaternariisquereticulatis,evidentibusinfra.
Inflorescentiaein thyrsisterminalibusvel axillaribusdispositae,4.0-15.0 cm longae.
Flores in dichasiassimpliciausque composita dispositi; axes dichasiorum0.5-2.0
mm longi, 0.5-0.8 mm crassi,tomentelli;pedunculi0.3-1.0 mm longi, 0.5-0.8 mm
crassi,tomentelli, bracteissubulatis, 1.0-2.5 mm longis, margineglanduloso-papil-
latis, tomentellis;pedicelli 1.0-2.5 mm longi, 0.3-0.5 mm crassi,tomentelli, bracte-
olis subulatis, 1.0-1.5 mm longis, margineut in bracteis,tomentellis;sepalaovata
vel oblonga, 2.8-4.0 mm longa, 1.7-2.2 mm lata, nonnumquamglandulosa,
tomentella;vexillum 4.3-4.6 mm longum, 2.5-2.8 mm latus, usque ad medium
longitudinalitersaccatum,apice leviter revolutum,intus barbatum;alae spathulatae,
3.3-3.8 mm longae, 1.3-1.6 mm latae, ad basin barbatae;carinaepetala saccata,
2.8-3.0 mm longa, 1.0-1.8 mm lata; stamina8-10, sterilia24, fertilia6, staminodiis
interdumappendiculatis,filamentisad mediumconnatis, 1.0-1.4 mmlongis, antheris
oblongisvel ellipticis, 0.5-0.7 mm longis, 0.3-0.4 mm latis; glandulae2, unilobae, a
fronte visae labiatae, lanato-villosae;stylus erectus, 1.0-1.2 mm longus, villosus,
stigmatecirculari,ca 0.3 mm diametro,albo; ovariumsubglobosumusque pyrami-
dale varians,ca 0.6 mm latum, 3-loculare,villosum, ovulis in quoque loculo
numerosis.Fructusjuvenilisoblongus,dense villoso-tomentosus.
Type. Tillett & Tillett 45524, Guyana,UpperMazaruniRiverBasin,Kako
River,fl (holotype NY; isotypes F, GH, P, US).
Distribution.At edge of forest in open areasalong roadsor rivers.
VENEZUELA.Bolivar:Steyermark&Aristeguieta77 fl, imm fr (VEN,U). GUYANA.
Forest Department7992, KukuRiver,Fl (NY).
Due to lack of adequate fruiting material, the relationships of this species are
impossible to determine. It has been confused with Trigonia villosa var macrocarpa
(T. macrocarpaBentham)and somewhat resemblesit but differsin the reticulation
of the leaves, the more complex inflorescence,and severalother minor characters.
Trigoniareticulata and T. bracteataare unique in Trigoniain havingglandular
papillaeon the marginsof the bracts,bracteolesand sometimesthe sepals.
?? r
FIG 14. Trigonia candelabra A, Habit, vegetative branch; B, Habit, flowering branch; C, Habit,
fruiting branch, X 0. 3?3.
SystematicTreatment 43
P*~~~ '
D E F
FIG 15. Trigonia candelabra A, flower X3.3; B, ovary X3.3; C, anthers X6.6; D, Keel petal X
3.3; E, Wing Petal X3.3; F, Standard X3.3; G, compound dichasium in bud X3.3.
This very distinct species shows affinities with Trigoniahypoleuca, in the leaf
(shape, pubescence, color, etc) and the fruit. However,the inflorescenceis very
distinct. It is the only speciesin the genusin which the flowers are arrangedin
perfectcompounddichasia.The sterile (vegetative)branchesbear some of the largest
leaves in the genus, another characterthat separatesit from T. hypoleuca. The
pubescencein the youngerleaves cannot be seen with the naked eye, or even with a
dissectingmicroscope(100X).
I have selected floweringmaterialas the holotype for the species. It is quite
possible that the fruitingmaterialdesignatedas paratypeis from the same plant,
gatheredat a later date.
The specific epithet, candelabra,refersto the arrangementof the ultimate
inflorescences.
Croton eriospermum Lamarck, Encycl. 2: 211. 1786. Type. Commerson sn Brazil, Rio
de Janeiro, fr (holotype P -Herb. Lamarck; isotype P).
Mainea racemosa Vellozo, Fl. Flum. 275. 1825; Vellozo, Iconog. 7: pi. 8. 1827; Netto,
Arch. Mus. Rio de Janeiro 5: 260. 1881.
Trigonia crotonoides Cambessedes, (Hippocrateaceae) St. Hilaire Fl. Bras. Merid. 2:
83. pl. 105. 1829; Grisebach, Linnaea 22: 31. 1849; Warming, (Trigoniaceae)
Mart. Fl. Bras. 13(2): 127. 1875. Type. St. Hilaire c102. Brazil, Rio de Janeiro.
fl, fr (holotype MPU; isotype P).
Trigonia crotonoides var incana Cambessedes, (Hippocrateaceae) St. Hilaire Fl. Bras.
Merid. 2: 83. 1829; Warming, (Trigoniaceae) Mart. Fl. Bras. 13(2); 127. 1875.
Type. St. Hilaire sn, Brazil, Rio de Janeiro, fr (holotype MPU).
46 Flora Neotropica
membranacea
Trigonia A. C. Smith,Phytologia
3: 128. 1935.
rasaStandley&Steyermark,
Trigonia Publ.FieldMus.Nat.Hist.23(2):59. 1944.Type.
Standley78584Guatemala, SantaRosa,RioPanal,fr (holotypeF;isotypeF).
wide, the pouch extending to 2 of the length, erect along the upperportion or
nearlyso, the apex revolute,barbateat the throat, the wings spathulate,2.5-3.1
mm long, 0.8-1.0 mm wide, glabrousat base, the keel petals 2.5-2.8 mm long, ca
2.0 mm wide, with the pouch extendingfrom 1/3 of the length up, the apex revolute,
glabrous;stamens 8, fertile ones 5-6, staminodes2-3, the filaments 1.0-1.5 mm
long, free for 1/3 of the length, the anthersovate or oblong, 0.4-0.5 mm long, ca
0.2 wide; glands2, 2-3 lobed; deltoid or trapezoid,ca 0.4 mm per side, glabrous;
style 0.8-1.1 mm long, glabrousor villous, the stigmatrilobateca 0.1 mm in diam-
eter;ovarysubglobose,ca 0.4 mm in diameter,villous-barbate,the ovulesnumerous.
Fruit not known.
Type. Schomburgk56 Guyana, fl (lectotype K - Bentham Herb.;isotypes
CGE,G, NY, W).
Distribution.Known only from Schomburgkcollections made in Guyana.No
localities are reported.
GUYANA. Schomburgk 63 fl (paratypes BR, CGE, G, OXF, U, W); Schomburgk 249 fl
(F, W); Schomburgk 373 fl (GOET, K).
Trigoniasubcymosais easily recognizedby its short, pyramidalpanicles,the
long peduncles, the typically curved bracts and bracteoles, and its fairly small
leaves and flowers.
Benthamcited both Schomburgk56 and 63 in his descriptionof this species.
These specimensarevery similarand either one could be selected to typify the
taxon. I have selected Schomburgk56 as the lectotype. Thereare other elements
labeled as 63 and 56 that are not referableto this speciesbut Trigoniavillosa,and
this mixturehas led to confusion in the past. The specimenfrom Bentham'sherb-
ariumat Kew has Schomburgkcollections 56, 63, and 373 mounted on the same
sheet. This could well mean that they constitute elements of the same gathering
that were numbereddifferently.The specimensare extremely similar.
The lack of fruit makesit extremely difficult to determinethe relationships
of this species.
Trigonia candida Warming,(Trigoniaceae) Mart. Fl. Bras. 13(2): 139. 1875. Type.Glaziou
2505, Brazil, Rio de Janeiro, fl, fr (holotype C; isotypes A, BR, C, F, GH, IPEAN,
NY, US).
Trigonia ovalifolia Glaziou, Bull. Soc. Bot. Fr. 3(52): 34. 1905. Type. Glaziou 14690,
Brazil, Minas Gerais, fl (holotype P; isotypes C, F, G, P).
Trigonia nivea forma paniculata Chodat & Hassler, Bull. Herb. Boiss. 2(8): 801. 1903.
Type. Hassler 8416, Paraguay, Apa River Headwaters, fl, fr (holotype G; isotypes
BM, G, MO, NY).
pubescensCambessedes,
Trigonia (Hippocrateaceae) St. HilaireFl. Bras.Mer.2: 114.
1829;Grisebach,Linnaea22: 28. 1849;Warming, (Trigoniaceae)Mart.Fl. Bras.
4. 1967.
13(2):135. 1875;Reitz,Fl. Ill.Catar.1(Trigoniaceae):
with the upper half revolute, irregularat apex, barbateat the throat, the wings
spathulate,6.0-6.5 mm long, 2.0-3.0 mm wide, barbateat the base, the keel petals
4.0-5.0 mm long, 3.04.0 mm wide, the pouch extending along 2/3 of the length,
barbateat the base;stamens 10-11(-12), 6-7 fertile, staminodes3-4, the filaments
2.5-3.0 mm long, free for 1/3 of the length, the anthersobovate or oblong, 0.5-0.7
mm long, ca 0.4 mm wide; glands2, 2-3 lobed, the lobes deltoid, 0.2-0.3 mm per
side, glabrous;style 2.5-2.8 mm long, glabrousor slightly villous, the stigmatrilo-
bate, ca 0.2 mm in diameter;ovarysubglobose,ca 1.0 mm in diameter,barbate
pubescent,the ovules numerous.Fruit4.5-11.0 cm long, the valves6.0-20.0 mm
per side; exocarp thin (ca 0.7mm) fleshy, with reddish-brownor yellowish-brown
velutinous-tomentosepubescence;mesocarpwoody, separablefrom the endocarp;
endocarpthin, sometimesdensely coveredwith long, brown velutinous-tomentose
pubescence,to glabrousor nearlyso. Seeds ca 20 per locule, ovoid, barbatepubes-
cent, the trichomesto 20.0 mm long, spirallydisposedaroundthe seed.
This speciesis dividedinto two varieties.
22. TrigoniahypoleucaGrisebach,Linnaea22:30.1849;Warming,(Trigoniaceae)
Mart.Fl. Bras. 13(2): 140. 1875; Macbride,Publ. Field Mus.Hist. Bot.
11(2): 69. 1931; Stafleu,(Trigoniaceae)PulleFl. Surin.3(2): 175. 1951.
Trigonia hypoleuca var pubescens Warming, (Trigoniaceae) Mart. Fl. Bras. 13(2): 140.
1875. Type. Wullschlaegel8161 Suriname, fl, fr (holotype BR, isotypes IPEAN,
NY, VEN).
Trigonia xanthopila Garke, Linnaea 22: 51. 1849. Kegel 1177 Suriname, fl (holotype on
2 sheets as indicated by Garke, GOET).
LITERATURE CITED
LISTOFTAXA
NUMERICAL
LISTOF EXSICCATAE
INDEX