A study of Bixa (Bixaceae), with particular reference to the leaf undersurface indumentum as a diagnostic character

7 Pages·1994·4 MB·

by  Ruth E Dempsey

Bull. not. Hist. Mus. Land. (Bot.)24(2): 173-179 Issued24November 1994
A
ofBixa with
study (Bixaceae), particular
reference to the leaf undersurface
indumentum as a character
diagnostic
RUTH E. DEMPSEY and NANCY C. GARWOOD
DepartmentofBotany, TheNaturalHistory Museum, CromwellRoad, London SW75BD
CONTENTS
Introduction 173
Materialsandmethods 174
Results 175
Discussion 176
Appendix 178
References .179
.
SYNOPSIS. In the course ofpreparation ofan accountofPanamanian Bixa urucurana Willd. for the Seedling Flora
Project, differences in leafundersurface indumentum were found between adults and juveniles. We describe the
patternofvariationofindumentum betweenspecieswithin the genusand betweendevelopmentalstageswithin B.
urucurana. We conclude that leafundersurface indumentum is a usefuldiagnosticcharacterforthe adult stagesof
thegenusbut notthejuvenilestages.
INTRODUCTION more or less kidney-shaped fruit, there are other defining
features which separate them as species. Baer (1976) con-
tends that they are diagnosable on fruit shape when this is
Bixa is the only genus of the small woody family, Bixaceae more accuratelydefined. Other diagnosticcharacters include
which is widespread in tropical America and frequently fruit ornamentation, dehiscence, life-form, and the size and
cultivated throughout the tropics. The genus has often been density of trichomes on the undersurface of the leaf (see
treated as one highlyvariable species, Bixa orellana L. sensu Table 1 fora summary ofthese).
lato (Eichler, 1871; Kuntze, 1891; Warburg, 1895; Standley Theprimarytaxonomicdifficultyin the genusisseparating
& Williams, 1961; Robyns, 1967; Croat, 1978) or as five Bixa orellana and B. urucurana. There is a high degree of
separate species (Macbride, 1941; Baer, 1976; Molau, 1983). variation between cultivars of B. orellana, especially in
Primarilyfruit andindumentumcharacters have been usedto capsule shape since this has been the subject of artificial
distinguish different species. Linnaeus' (1753) protologue of selection. Baer (1976) discusses possible intermediates or
B. orellana simply described the capsules as 'chestnut-like'. hybrids between B. orellana and B. urucurana, known pre-
Willdenow (1809) described a second species, B. urucurana dominantly from Panama and Costa Rica. These apparently
from Brazil, definingit bythe 'white' scales densely covering have trichomes of intermediate size and density, relatively
the leafunderside. This species has frequently been reduced short, subspherical fruits with slender spines and intermedi-
in rank to a variety of B. orellana, especially in Central ate numbers ofseeds.
America (Kuntze, 1891; Standley & Williams, 1961; Robyns, The variation in leafindumentum isourprincipalconcern.
1967). In 1831, Don published a description ofB. platycarpa The undersurfaces of leaves of all five species of Bixa have
Ruiz & Pav. from Peru, which had flat, kidney-shaped fruits peltate scales, which are circular discs supported on short
with very few bristles. Triana (1858) described B. sphaero- obconical stalks about 20 |o,m long. They are red-brown in
carpa from Nouvelle-Grenade (now Colombia), having scaly colour and formed of an outer ring of cells with a central
leaf undersides and spherical capsules, but this was later circular area which is more densely pigmented. In B. uru-
reduced to a synonym of B. urucurana (Pilger, 1925; Baer, curana, the scales are large, with several concentric rings of
1976). Huber (1910) described B. arborea, having a kidney- cellsaround the centralcore, irregularinoutline andfunneli-
shaped, dorsallycompressed capsulewith awrinkledsurface. form. Its scales are also very densely packed, giving a
Finally, B. excelsa Gleason & Krukoff was described in tawny-brown appearance to the undersurface of the leaf.
Gleason (1934), and is characterized by reniform capsules (Willdenow's(1809) originaldescriptionofwhite scalesprob-
densely covered by glandular trichomes. Although the origi- ably referred to a shiny appearance, observed due to reflec-
nal descriptionsofthe three species restricted to the Amazon tion ofincident light at certain angles).
basin (B. arborea, B. excelsa, B. platycarpa) all refer to a Our study is based on herbarium material collected and
TheNaturalHistoryMuseum, 1994
174 R.E. DEMPSEYAND N.C. GARWOOD
Table 1 Summaryoflife form, fruitandscalecharactersofthe five Bixaspecies. (Baer, 1976; Molau, 1983; Macbride, 1941)
Species Fruitshape Spines Dehiscence Habit Trichome
l.s. t.s. diameter density
mm2
(|xm) (per )
B. arborea
STUDY OFBIXA (BIXACEAE) 175
Table2 Resultsoftwo-wayANOVA (Model II)ofeffectsofposition(edge, middleorvein)andleafonscaledensityanddiameter.Two
leavesoneachoffivejuvenilesandoneadultwere included.
Scalediameter Scaledensity
Totalscale Centralcore
Source DF MS MS MS
Leaf
176 R.E. DEMPSEY AND N.C. GARWOOD
mens (9.4-13.6 cm long x 5.0-9.7 cm wide) and are well Although there are too few specimens to form a reliable
within the published ranges for adults of the species cluster, the three Amazonian tree species (B. arborea, B.
(5.5-27 cm long x 3-18 cm wide), (Standley & Williams, excelsa and B. platycarpa) also cluster with Bixa orellana
1961; Robyns, 1967; Molau, 1983). (Fig. 2b).
C) Bixa orellana and B. urucurana have the least certain
specificstatus in the genus. On a plot ofmean scale diameter
versus mean scale density (Fig. 2a) specimens attributable to
these taxa cluster into two groups (see Appendix for list of DISCUSSION
specimens), the lower being B. orellana and the upper B.
urucurana, with two exceptions. Firstly, the Panamanian
BM
Fendler 328 specimen ('F' in Fig. 2a), the specimen of There is considerable variation in the indumentum of leaf
which consists ofa mature B. urucurana fruit and a separate undersurfaces in the genus Bixa. There are also several
small non-fruiting branch, clearly groups with B. orellana. different trends in this variation which need to be considered
Baer(1976) placed aspecimen fromthe same collection in B. andexplainedbefore the utilityofindumentumdifferencesas
urucurana. Secondly, ourjuvenile specimensofB. urucurana adiagnosticcharacterofspeciescan be evaluated.
from Panama and the unidentified Bixa sp. seedlings from Differencesbetweenpositionson the same leafwere found
BCI clearly group at the lower end ofthe B. orellana range. to be much less significant than other effects (e.g. leaf,
In contrast, the (adult) parent is clearly in the B. urucurana individual or age) for scale density and not significant for
group. scale diameter. Baer (1976) stated that scale density
South American specimens of both Bixa orellana and B. increased near veins or at the base ofthe leaf, but we found
urucurana group at the extremes ofthe ranges of both mean little evidence forthis in either adults orjuveniles.
scale diameterand density. The Central American specimens Both scale density and scale diameter are radically differ-
ofboth taxagroup towards the centre ofthe overall range of entbetween adultsandjuvenilestagesofBixaurucurana. On
indumentum characters. West Indian specimens of B. orel- average, adult scales are more than twice as large and about
lana group closely with South American representatives of fourtimesdenserthanthose ofthejuveniles. There is aslight
the same taxon. overlap, however, between the ranges of density in adults
20 r 20
15 15
00
.00
* 10 .. 10
E
&
a
CD 5 5
L L
40 r 40
30 30
o
20 20
1 o
CoO Z 10 o o C) 10
L
65 65
60 60
55 55
i v
50 50
03
CoO
45 45
CO
40
25 30 35 40 45 25 30 35 40
Leaf node Leaf node
Fig. 1 Relationshipofscalediameter,scaledensityandleafwidthtoleafnode injuvenilespecimens.
(a)Juvenile Garwood 1830D.
(b)Juvenile Garwood2085D.
(Opencirclesrepresentfullyexpandedleaves, filledtrianglesexpandingleavesandopentriangles 'borderline' leaves. Errorbarsforscale
densityanddiameterrepresentthestandarderrorofthe meanand,wheremissing, theerrorsaresmallerthanthesymbols. Forscale
diameter,the hatched line representsanon-significant linearregression. Thelinearregressionsforscaledensityaresignificant (p = 0.025).
Onlyfully-expandedleaveswere included inthe regressioncalculations.)
STUDY OF BIXA (BIXACEAE) 177
O
o
o
O00
G
E
\ o
o
o
Qo3
70 80 90 70 80 90 100 110 120
Diameter (pm) Diameter (pm)
Fig.2 Scalediameterversusscale densityin Bixaspp. Mean SE (N = 5 replicatesonone leafexcept forGarwoodspecimensmarkedby
stars).
(a) B. orellanaand B. urucurana. (FilledcirclesrepresentSouth Americancollections, hatchedcirclesWest Indiancollectionsandopen
circlesCentralAmericancollections. StarsrepresentourPanamaniancollections: A = Adult (N = 10replicatesfrom two leavesonone
individual),J= Juvenile (N = 10replicatesfrom twoleavesoneachoffiveindividuals) andS = Bixasp. seedling(N = 10replicatesfrom
twoleavesononeindividual). F = PanamanianspecimenFendler328.
(b) Five Bixaspecies. (OpencirclesandstarsrepresentB. orellanaand B. urucuranaspecimensfromFig. 2a;squares, B. arborea; pentagons,
B. platycarpa; diamond, B. excelsa).
(27-65 scales per mm2) and juveniles (2-35 scales per mm2). plotted against mean scale density (Fig. 2). The lower group
There is no overlap between scale diameter of the juveniles with small, sparse scales includes all the specimens of Bixa
(27.6-73.6 u,m) and the adults (82.8-147.2 |xm). orellana sensu stricto. The upper cluster of individuals with
The difference in total scale diameter between adults and large, relatively densely packed scales, contains specimens
juveniles is accounted for by the outer portion or annulus of whichwere originallyvariouslyidentified as B. orellanasensu
the scale: there is no significant difference in central core lato, B. urucurana, B. orellanavar. urucuranaor B. sphaero-
diameter between stages. In Bixa urucurana, the scales on carpa (see Appendix). We interpret these as B. urucurana
adult leaves have a number of concentric rings of colourless since they form a close-knit group on the basis of indumen-
cellsaroundthedarkcentralcore (Baer 1976). Evidently, itis tum characters and their fruit characteristics correspond,
an increase in the number of these rings which increases being subspherical, spiny and indehiscent. The problematic
overall scale size in the adults of this species. Although we specimen, Fendler 328 ('F' on Fig. 2a) which Baer identified
observed these annuli in juvenile specimens, they were as B. urucurana, clearly groups with B. orellana on the basis
generally very narrow, comprising perhaps only one or two of indumentum characters. However, there is a mature,
rings ofcells and were occasionally completely absent, espe- subspherical capsule on the herbarium sheet which looks like
cially in thevery young beach-collected seedlings. that typical of B. urucurana. On closer inspection, it was
The second trend within Bixa urucurana is that scale noticed that the vegetative shoot on the herbarium sheet was
density increases with node number in juvenile specimens, non-reproductive and had never borne flowers or fruit;
whilst there is no clear trend in scale diameter with node therefore, it must have been collected either from a com-
number. However, since we found a significant difference in pletely different part ofthe same plant oreven perhaps from
scale diameter between adults and juveniles, it would appear a different individual than the fruit.
that changes in this parameter (i.e. the increase in numberof TheSouthAmericanspecimensofthetwoshrubbyspecies,
rings of cells around the centre of the scale) must occur at Bixaorellana and B. urucurana, group at the extremesofthe
some point between the two stages ofplant development we range of indumentum characters (Fig. 2a). Although the
studied. Further study of a range of intermediate, sapling Central American specimens still form recognizable clusters
stages would be required to ascertain whether there is a with the conspecific specimens from South America, their
gradual increase in scale size with node. Scale diameterisnot indumentum characters tend to fall into the middle of the
low due to small leaf size, since our juvenile leaves are overallrangeforthetwospecies. ItwouldappearfromBaer's
generally larger than the adult leaves examined. (1976) estimates ofscale density and diameter (Table 1) that
Thirdly, the comparison between our specimens and those some ofthe Central American specimens fall into his 'inter-
of a sample from the BM herbarium shows that two distinct mediate'category. However, hisscale densityestimate for B.
clusters of taxa are formed when mean scale diameter is urucurana, 156/mm2, is inordinately high when compared to
178 R.E. DEMPSEYAND N.C. GARWOOD
our measurements: our highest mean value was 70 scales/ development. The stages in this process apparently encom-
mm2 and our highest single value was 85 scales/mm2. Unfor- pass all the variation observed in B. orellana and the Amazo-
tunately, because Baer does not give his sample sizes, it is nian species. That four of the five Bixa species and the
difficult to compare the measurements. The fact that speci- juveniles of B. urucurana have small sparse scales suggests
mensofthe twoCentral American species are much closerto that perhaps the large, densely packed scales of adult B.
each other than are those from South America, may suggest urucurana represent a more derived state.
that there is some degree ofintermediacy. Perhaps ifwe had The demonstration of the range of variation in indumen-
sampled more specimens, we would have observed more ofa tum characters through development within Bixa urucurana
continuum rather than two separate clusters. Since West clearlyshowsthatthe utmostcaution must beexercisedwhen
Indian specimens of B. orellana group with South American using them as diagnostic characters within the genus. It is
specimens ofthe same taxon, it appearsthat dispersal (either important primarily to ensure that the leaf material studied
naturallyorthroughcultivation) hastakenplacedirectlyfrom comesfrom a mature individual.
South America to the West Indies, bypassing Central
America.
We conclude that the differences between indumentum ACKNOWLEDGEMENTS. The Seedling Flora Project has been gener-
characters in B. orellana and B. urucurana adults distinguish ously supported by the National Science Foundation, the Natural
them as two separate species. This is reinforced by differ- Environment Research Council, the Smithsonian Tropical Research
encTehseinthfrreueitAcmhaarzaocnteirasn(stereeeTasbpleceie1s).cluster with Bixa orel- SIinsetrirtau,tep,rianncdipaTlhetecNhantiucraallaHssiissttoarnyt MinusPeaunma.maW,eINthRaEnkNAERduEarfdoor
lanaanddo notappearto beseparable from itonthe basisof permission to work in Panama, C. Humphries for comments on the
indumentum characters alone, although it is recognized that manuscript, and the many people on BCI and at STRI who helped
the sample studied here is small. the Project in various ways but are too numerous to thank here
It is apparent that Bixa urucurana manifests a broad range individually.
of scale diameter and density as it proceeds through its
APPENDIX
Tableofspecimensexamined.
Mean(SEM)
STUDY OF B1XA (BIXACEAE) 179
Garwood
(J)