TRIBES AND GENERA OF THE Lennart Andersson2 CINCHONEAE COMPLEX (RUBIACEAE)' ABSTRACT The phylogenetic relationships of taxa recently referred to the tribes Calycophylleae, Cinchoneae, and Hillieae were analyzed using cladistic methods. The analyses used morphological and micromorphological characters, and a selection of 60 species were used as terminal taxa. It was concluded that ail three tribes are supported as monophyletic but must be altered somewhat in circumscription. The tribes Cinchoneae and Hillieae are sister groups, and these two in turn form the sister group of the Calycophylleae. The tribe Calycophylleae should comprise the genera Alseis, Calycophyllum, Capirona, Ferdinandusa, Macrocnemum, Wittmackanthus, and the new genus Semaphyllanthe L. Andersson here segregated from Calycophyllum. These genera are ail supported as monophyletic, although the support is sometimes weak. The tribe Hillieae should comprise Balmea, Blepharidium, Cosmibuena, and Hillia. It was not tested whether or not these genera are monophyletic. The tribe Cinchoneae should comprise Cinchona, Joosia, Ladenbergia, Maguireocharis, Pimentelia, Remijia, Stilpnophyllum, and the new genus Cinchonopsis L. Andersson here erected for Cinchona amazonica. Dolicholobium occurs as a basal clade of the Cinchoneae but its position is unstable and it is possibly not related to this alliance at ail. Cinchona is supported as monophyletic if C. amazonica and C. henleana are excluded, the latter one being transferred to Ladenbergia. Ladenbergia is supported as monophyletic in approximately the traditional sense. Remijia is supported as monophyletic if Cephalodendron is included. Joosia and Stilpnophyllum are supported as monophyletic, while the remaining genera are monotypic. Six new combinations are made: Semaphyllanthe intonsa (Steyerm.) L. Andersson, S. megistocaula (K. Krause) L. Andersson, S. merumensis (Steyerm.) L. Andersson, S. obovata (Ducke) L. Andersson, S. spectabilis (Steyerm.) L. Andersson, S. venezuelensis (Steyerm.) L. Andersson, and Cinchonopsis amazonica (Standley) L. Andersson. The tribe Cinchoneae is often widely circum- scribed (see, e.g., de Candolle, 1830; Schumann, 1891, 1897; Robbrecht, 1988). In this sense it is a pantropical group comprising ail genera with multiovulate ovaries, ascendingly imbricate ovules, and winged seeds. The case was examined by An- dersson & Persson (1991), who concluded that a number of genera traditionally included in the Cin- choneae are not at ail closely related to Cinchona. Out of 50 genera referred to the Cinchoneae by Robbrecht (1988), 10 were transferred to the Hed- yotideae and six to the Condamineae. Disregarding four genera left as incertae sedis, the remainder were divided among four tribes. The paleotropical genera, except Dolicholobium, were referred to the tribe Coptosapelteae. Four genera, Alseis, Ca- lycophyllum, Schizocalyx, and Wittmackanthus, were referred to the new tribe Calycophylleae. Hil- lia was left in a monotypic tribe of its own, the Hillieae. Thirteen genera were left in the Cincho- neae: Capirona, Cephalodendron, Cinchona, Cosmibuena, Dolicholobium, Ferdinandusa, Joosia, Ladenbergia, Macrocnemum, Maguireo- charis, Pimentelia, Remijia, and Stilpnophyl- ' This study was supported by a grant (B-BU 08997 Research Council. 2 Dept. of Systematic Botany, Univ. of GCteborg, Carl lum. The relationships between these four tribes were not resolved in strict consensus although a majority of equally parsimonious solutions sug- gested that the Coptosapelteae are possibly the sister group of the three neotropical tribes, which in turn may form a monophyletic group. The tribes Cinchoneae and Hillieae are probably sister groups, forming together the sister group of the Calyco- phylleae. The investigation of Andersson and Persson was based in large part on data from the literature, using genera as terminal taxa. The aim of the present study is to reach higher resolution regard- ing circumscription and interrelationships of the neotropical tribes and genera of the Cinchoneae complex, using original observations throughout and species as terminal taxa. Two initial assumptions are based on the findings of Andersson & Persson: (1) the neotropical tribes Calycophylleae, Cinchoneae, and Hillieae form a monophyletic group, possibly including also the genera Balmea and Blepharidium; (2) the sister group of this complex is the tribe Coptosapelteae. Neither of these assumptions was fully supported, -307 to L. Andersson) from the Swedish Natural Science Skottsbergs Gata 22, S-413 19 Giteborg, Sweden. N. MISSOURI BOT. GARD. 82: 409-427. 1995. Annals of the Missouri Botanical Garden but they are accepted here in lack of a better hypothesis. The genus Schizocalyx was included in the Ca- lycophylleae by Andersson & Persson but, the fruit and seed characters now having been examined, it appears that its closest relative is Bathysa in the Rondeletieae. It is therefore excluded from this study. MATERIALS AND METHODS Observations of morphological features were made on herbarium material, occasionally supple- mented by pickled material. Species descriptions were used as a basis for coding and were based on numerous collections; a list of representative ma- terial studied is given in Appendix 1. Most vege- tative characters were observed on dry material, supplemented by label data. Vegetative buds, flow- ers, and seeds were boiled in water before dissection and observation. Leaf blade anatomy was observed in material cut by hand from boiled leaves and examined under the interference contrast microscope without stain- ing. Observations on the sculpturing of the inside of exotesta cells were made with the aid of the scan- ning electron microscope (SEM), using material treated in the following way. Rehydrated seeds were stored in 70% ethanol and transferred to tap water before treatment. The outer exotesta wall was oxidized by treatment in 15% HNO:, on a boiling water bath for 5 min. The seeds were then rinsed twice in distilled water and treated with a sonicator at 8 microns. No treatment time was found that suited ail materials, and a series of treatments varying between 1 and 20 seconds was therefore tried for each sample. After sonication the seeds were rinsed in water, transferred briefly to 96% ethanol, and left to dry on a slide. Gen- erally, only one seed sample from each species was studied. Pollen characters were studied with the aid of the SEM, using the methods described by Anders- son (1993). Generally, only one sample per species was studied. The cladistic analyses were performed using the PAUP software package (Swofford, 1993). Nu- merous preliminary runs were made to guide the selection of taxa and characters as well as the coding of characters. The final runs, on which the following discussion is based, were divided into two steps. Step I was designed to reveal major subdivisions in the complex, i.e., mainly aiming at establishing tribal delimitations. For this a matrix was used comprising a small number of representatives (Fig. 1) from ail "generic level" groups found in the preliminary runs. In this first series Hymenodic- tyon floribundum was used as outgroup, being a taxon mostly included in the Coptosapelteae in the investigation by Andersson & Persson (1991), and one for which a complete data set was available. Four characters were deleted in Step I: 2, 43, 47, and 57. Step II was designed to achieve better resolution and a more detailed hypothesis concerning the "Cinchona clade," clade 60 of Figure 1. For this purpose a matrix was used containing ca. 10 spe- cies each of the larger genera and ail species of the smaller genera for which sufficient data were available (Fig. 2). Two species found in Step I to belong to basal clades of the sister group were used as a joint outgroup, Ferdinandusa elliptica and Macrocnemum roseum. Thirteen characters were deleted in Step II: 13, 14, 16, 29, 31, 32, 33, 35, 39, 46, 56, 59, and 60. The 60 terminal taxa were selected arbitrarily from ca. 120 species actually investigated in order to minimize the amount of missing data. The selection of characters was based on ex- perience from an ongoing species level revision, and ail recorded characters found to vary among species and being shared by at least two species were included in preliminary runs. Characters found in preliminary runs to be highly homoplasious were then excluded. All characters and states are listed in Table 1. PAUP was run with the following optional heu- ristic search settings: addition sequence = simple, 1 tree held at each step, TBR branch swapping performed, MULPARS option in effect, steepest descent option not in effect, MAXTREES = 1000, multistate taxa interpreted as polymorphism. Quantitative characters (4, 11, 27, 30, 48, 58) were treated as ordered, others as unordered. After a first run with unweighted characters, the char- acters were reweighted by the maximum value of the rescaled consistency index. Successive rew- eighting was then performed until reaching a con- stant number of trees and identical consensus trees in two successive runs. In step two, monophyly of the outgroup was secured by reinforcing a mono- phyly constraint tree. A complete character matrix is given in Table 2; also included in this table is a list of acronyms with corresponding species names and authorities. The results of the cladistic analyses were analyzed with the aid of the MacClade package (Maddison 410 Volume 82, Number 3 1995 Andersson Cinchoneae Complex >> i wn F o< O< O< z wZ u u m cv w m o m ` mcp < O w N u � w � i � i i W� � V � m m t� $� � U � a � � O i � 31 c20 m) 2 smaller trees, or shrubs 2-phyllotaxy 1 decussate 2 whorled 3-stipule persistence 1 caducous 2 persistent 4-stipules: intrapetiolar fusion 1 free 2 partly fused 3 fused throughout 5-stipule posture in bud 1 held flat back-to-back 2 clasping to contorted 3 overlapping 6 - domatia 1 absent 2 tuft or pouch (-with-tuft) domatia present 3 pit or crypt domatia present 7-density of secondary leaf veins 1 < 15 pairs per 5 cm of midrib 2 > 15 pairs per 5 cm of midrib Inflorescence morphology 8-terminal inflorescence 1 present 2 absent 9-partial inflorescences ("principal inflorescence branches") 1 lax, dichasial or mainly so 2 congested into headlike clusters, dichasial or mainly so 3 � purely monochasial 4 thyrsoid 5 raceme- or spike-like 6 of solitary flowers 10-bracts and bracteoles 1 present, plane and � triangular 2 present, � sacciform 3 absent Floral morphology 11-flower size 1 small, corolla tube + lobe <10 mm long 2 medium, corolla tube + lobe 10-25 mm long 3 large, corolla tube + lobe 25-55 mm long 4 very large, corolla tube + lobe >55 mm long 12-heterostyly 1 absent 2 present 13-protogyny 1 absent 2 present 14-calyx shape 1 tubular at base, with lobes 0.5-2 x the length of tube 2 tubular at base, with lobes <0.5 x the length of tube (or lobes sometimes obsolete) 3 tubular at base, with lobes > 2 x the length of tube 4 � closed in young bud, opening by a slit ("spa- thaceous") 5 of free lobes 6 reduced to minute teeth or totally obsolete 15-colleter placement inside calyx 1 in a row at base 2 solitary or in small groups at sinuses 3 in groups or solitary toward base opposite si- nuses 4 scattered � all over 5 numerous on lobe margins at base - coded as missing when colleters are lacking 16-calycine semaphylls 1 absent 2 present at least in some flowers of an inflores- cence 17 corolla aestivation 1 valvate 2 contorted 3 imbricate 4 open 5 reduplicate 18- corolla shape 1 campanulate 2 salverform 3 urceolate 4 funnel-shaped 19-corolla color 1 white to cream, sometimes pale greenish 2 pink to reddish or purplish 20 corolla indumentum outside 1 absent or sparse 2 sericeous 3 + densely hirtellous 21 - corolla tube indumentum inside 1 absent 2 hirsute above stamen attachments, with long, rigid hairs 3 � hirtellous or villosulous to villose above sta- men attachments 4 � densely pubescent below stamen attachments 22-corolla tube surface inside 1 smooth 2 papillose above stamen attachments or through- out 3 papillose only below stamen attachments 4 coarsely verruculose 23 linear thinnings in corolla tissue between sinus base and filament attachment 1 absent 2 present 24-corolla lobe indumentum adaxially 1 glabrous 2 shortly papillose-hirtellous along margins 413 Annals of the Missouri Botanical Garden 3 villous along margins, with long, thin-walled hairs 4 � densely villous or hirsute ail over, with thick- walled hairs 25-terminal, wedge-shaped part of corolla lobes 1 absent or inconspicuous 2 conspicuous 26-marginal, petaloid appendages of corolla lobes 1 absent 2 present 27-filament attachment 1 to basal � of corolla tube 2 around middle (1�- of corolla tube) 3 to distal �V of corolla tube 28-filament length 1 long (� equaling anther or longer) 2 short 3 long in brevistylous, short in macrostylous flow- ers 29-filament indumentum 1 absent 2 filaments hirsute to villous 30-anther attachment 1 dorsifixed around middle 2 dorsifixed near base 3 basifixed 31-anther presentation 1 anthers included to semi-exserted 2 anthers wholly exserted 32-anther shape 1 elongate to linear (length/width ratio - 2) 2 button-shaped (length/width ratio <2) 33-anther base 1 obtuse 2 sagittate 34-apical connective process 1 absent 2 present 35-dorsally widened and reinforced connective 1 absent 2 present 36-disc indumentum 1 absent 2 present 37-style indumentum 1 absent 2 hirsute or puberulous distally or throughout 3 hirsute or puberulous at the very base, other- wise glabrous 38-stigma branches 1 straight, long (on the average > 15% of style length) 2 straight, short (on the average <15% of style length) 3 strongly recurved 4 absent, stigma � capitate 39-placenta 1 attached centrally by a lamella 2 attached apically by a stipe Fruit morphology 40-septicidal dehiscence 1 present, basipetal 2 present, acropetal 3 absent 41 -loculicidal dehiscence 1 absent 2 present, partial 3 present, � complete 42-apical, beaklike process in mature fruit 1 absent 2 present 43-endocarp 1 thin (<0.4 mm thick) 2 thick 44-calyx presence 1 calyx present (but often much eroded) in fruit 2 calyx abscised in fruit Seed morphology 45-hilum position 1 central or subcentral on seed body 2 marginal 3 on wing 46-seed wing shape 1 wing completely (but sometimes partly narrow- ly) surrounding seed body 2 wing interrupted at seed body 3 rudimentary and irregularly developed 47-seed wing margin 1 entire, irregularly lobed, or � dented 2 deeply laciniate 3 densely ciliate/fimbriate 48-seed length (including wing) 1 small (< 4.0 mm) 2 medium (4.0-11 mm) 3 long (> 11 mm) Vegetative anatomy 49 raphides 1 absent 2 present 50-adaxial leaf epidermis 1 1-layered 2 2- to several-layered 51-adaxial leaf cuticle 1 thin 2 thick (approaching half of epidermal cell depth, or thicker) Pollen 52-aperture type 1 colporate with distinct colpus 2 colporate with indistinct colpus 3 porate 414 Table 1. Continued. Volume 82, Number 3 1995 Andersson Cinchoneae Complex Table 1. Continued. 53-colpus ends 1 acute 2 obtuse to truncate 54-os 1 clearly demarcated with regular, often slightly thickened margin 2 poorly demarcated with irregular margin 55-tectum perforation type 1 reticulate (lumen/muri ratio > 1.0) 2 foveolate (lumen/muri ratio <1.0) 3 imperforate 56-lumen differentiation into two distinct size classes 1 absent 2 present 57-latitudinal differentiation of lumina 1 absent, or lumina smaller toward colpi 2 lumina larger toward colpi to their posture in bud. They are either held flat back-to-back, and are then usually lingulate to obovate and herbaceous in texture, or they are clasping to contorted, forming a cone over vege- tative buds, and then more or less triangular and usually brownish and more or less coriaceous in texture. When the stipules are fused throughout both inter- and intrapetiolarly to form a calyptra, this character does not apply and was coded as missing data. In Joosia the branching of partial inflorescences is predominantly monochasial, more or less bora- goid shoot chains arising as lateral branches from a basal dichasial branching. In Joosia aequatoria and J. dielsiana, with more or less congested in- florescences, the boragoid branches are often short and few-flowered. In these cases it is dubious wheth- er inflorescence branching should be coded as di- chasial or monochasial, occasional monochasial branchings being common in cymes of the Rubi- aceae. Based on observations of a number of col- lections of J. dielsiana with longer inflorescence branches it was concluded that this ambiguous sit- uation has arisen through reduction of longer mon- ochasial branch systems. The inflorescence branch- es of J. aequatoria and J. dielsiana are therefore coded as monochasial. In Remijia inflorescences are typically thyrsoid. In several species, like R. pilosinervula, where the thyrses are few-tiered, reduction series occur, from several-tiered to 1-tiered thyrses. In the ex- treme such inflorescences may appear to be highly condensed dichasia. In such cases these inflores- cences were coded as thyrsiform, however. 58-columella length 1 long (> 0.5 �m) 2 medium (0.15-0.4 pim) 3 short (< 0.15 p�m) Seed anatomy 59-wall thickenings of inner tangential wall 1 meshwork or broad bands 2 plate with large pores 3 walls not thickened, or thickening an undiffer- entiated sheet 60-spinules or verruculae on secondary wall inside 1 absent 2 present When the available flowering collections are few it may be impossible to decide whether the species is heterostylous. Macrostylous flowers in heteros- tylous species are not strikingly different from flow- ers of monomorphic species, and in some of the heterostylous species examined the number of ma- crostylous plants may exceed that of brevistylous ones by one to ten. The following coding practice was followed. In genera where heterostyly was doc- umented to be common (Cinchona, Joosia, Lad- enbergia, and Remijia), the species were coded as heterostylous if the brevistylous morph was en- countered, otherwise as missing data. In other gen- era, where heterostyly has never been documented, species were coded as homostylous on the basis of single observations of long-styled flowers. In Dolicholobium the flowers are functionally unisexual (on monoecious plants), with the non- functional organs vestigial. Because this character state is autapomorphic in the present data set, it was not included in the analysis and D. oblongi- folium was coded as homostylous, the difference in style development between flower morphs being obviously non-homologous with heterostyly in other genera. The presence of colleters on the inside of the calyx is highly unpredictable, even within species. When present, however, the number and place- ment of colleters is fairly constant within species. Therefore, states were assigned to placement and grouping of colleters, while absence of colleters was coded as missing data. Corolla aestivation in some species of Cosmi- buena was described as quinquncial by Taylor 415 416 Annals of the Missounri Botanical Garden �, 6 cj d e r I �^ I II * E ^^^ ^ ^ ^ ^ _ Mcii < -- - -- -- .-- .2 s lil� ll, . ^ ^^ ^ - - V - - - - - - - - - - ---- - - - m- - NNNN 1 41 N NN � (,y !' -"r1^ O I u , 3 3 " - ph p a N1 < N N - - - - - - - - N Cu N N N N N N Cu N N " < 11 ��(s �� �T �� N3 Nt� N~ NT N) -' t^ ^ ^^' c � g---- N N--- N N ~ ~ ~ ~ ~ a ('CINNN . Q Li yW ~IU P w l� NNNNN10NN 1 -c N N" !'- � �tl ||------------------- V 3 0 I M O s e w .2- tN-- ---N- NNN N N NN N N+- NN N CNN "1^ 44 4N0 ^^ >�^l' �0N------ -----N- sa u t ,-& N . � 0 --CMCM- <'4CM<'J' CM---.-CM-CM-. - - �O -* C,->e->C-- -- - - - - --tCOIlClCItXlN N NN - - -N -- N ------ S--- -- - - --- ----- - - -- --C >0 N I - �- " - C- �- ^- �- ^ - M �M N .-\� " � �~ - C--O - n�.<0 - e>M <0CM -.-- M.-C>INC CMCM CM CM CM - - - - -- -N - a CM- - - - - -C----CMCM CM <0 CC- -CMC- M-CMC- CM - C -C - - - CMC M- 0 C- - - M C- - C- N - ---- -- --- . N ---.-------.--CM- CM CMCMCMCMCMCMCMCMCMCMCMCMN---->'.CMCMCMCMCMCMCM - - - - - - - - - - - ---- -- ----- --- -- -------- --- -- - - ---N( N NN----- -----CMCMNNN f N N NCMCMCM -CM-CM-- -CC - - - N N N CMCMCMCM.f --- ---- N-fNN----- M ----- ----.--e>..CM - - - - - - - - -e> N N N N N N--CMCMCMCM--- CMNN e-CM-.-.CMC-NM -- -- - -- -- -----------CM .C--.-C--. 0-f.uP -N N " ' nf` .n - hnNM C f Annals of the Missouri Botanical Garden (1994). However, according to accepted usage (e.g., Stearn, 1983) quinquncial is a special case of im- bricate. In Hillia, Cosmibuena, and Dolicholob- ium the petals occur in two layers (typically three in the outer and two in the inner layer), being typically contorted in each layer. Aestivation is therefore coded as contorted in these taxa. In Cinchona filament length differs considerably between brevistylous and macrostylous flower morphs. This is coded as a separate character state under character 28. In cases where only macros- tylous flowers were seen this character is coded as missing data. Stigma branch length (character 38) in heter- ostylous species was calculated on the basis of the macrostylous morph. In cases where only the brev- istylous morph was observed, style length of the macrostylous morph was approximated to equal corolla tube length and the value was accepted if well separated from the class limit; if close to the class limit, the character was coded as missing data. Pollen characters were discussed and docu- mented in detail by Andersson (1993). RESULTS AND DISCUSSION Throughout this discussion, clades will be des- ignated by the numbers (Figs. 1 and 2) of the edges supporting them. Terminal taxa will be referred to by their traditional names, on which the acronyms used in the trees are based; a number of new names and combinations are introduced in the Synopsis. Only selected characters will be discussed here; complete apomorphy lists for the trees of Figures 1A and 2A are given in Appendix 2. The Step I analysis run with unweighted char- acters resulted in 24 equally parsimonious trees with a length of 286 steps and an overall consis- tency index of 0.490. Successive reweighting gave nine equally parsimonious trees 289 steps long (calculated on the basis of unweighted characters) and with a consistency index of 0.484. Using Ar- chie's (1989: 267) empirical formula, homoplasy excess ratio (HER) for the tree of Figure lA was estimated at 0.587. Both with weighted and unweighted characters, a basal dichotomy was found between clades 45 and 60 (Fig. lA). Clade 45 is not supported by any unique and consistent apomorphies, but by three rare character states, long filaments (with parallelisms in Pimentelia and Stilpnophyllum), anthers fully exserted (with reversai in Capirona), and anthers button-shaped (with reversals in Ca- pirona and Wittmackanthus). Clade 45 contains ail genera (except Schizocalyx) referred to the Calycophylleae by Andersson & Persson (1991), plus three genera referred by them to the Cincho- neae: Capirona, Ferdinandusa, and Macrocne- mum. Macrocnemum and Capirona are further nested in clade 45 by the presence of bristly hairs in the corolla tube and on the filaments, and by the interruption of the seed wing at the seed body. Ail are excluded from clade 60 by the lack of the apomorphic pollen characters foveolate exine and obtuse colpus ends. The conclusion seems well founded, therefore, that Capirona, Ferdinandusa, and Macrocnemum should be transferred to the Calycophylleae. Calycophyllum in the traditional sense of, for example, Bacigalupo (1952) and Steyermark (1964, 1974) is divided between two clades, one (52) comprising the type species, C. candidissimum, and two closely related species, the other one (54) comprising the so-called "Calycophyllum obova- tum group" (Steyermark, 1964). Calycophyllum megistocaulum (= C. acreanum Ducke), consid- ered by Steyermark (1964) to be closely related to the "C. obovatum" group, is nested in clade 50 by virtue of its huge size, and small flowers with a campanulate corolla, but its position within the clade is not resolved. Capirona is nested in clade 50 on the basis of, among other things, huge size and presence of calycine semaphylls. It is nest- ed in clade 51 also on the basis of the verruculose inside of its exotesta cells. Capirona is well char- acterized by several unique autapomorphies such as intrapetiolar stipules (stipules fused throughout intrapetiolarly and deeply divided interpetiolarly) and its unique exine structure (Andersson, 1993). This, taken together with the fact that clade 52 is likewise supported by several unique synapomor- phies, makes it rather unattractive to merge Ca- pirona with Calycophyllum. In consequence, Ca- lycophyllum will have to be split into two genera, Calycophyllum s. str. (clade 52) and Semaphyl- lanthe (clade 54), described as a new genus in the Synopsis. Calycophyllum s. str. (clade 52) is supported as monophyletic by three unique synapomorphies, i.e., sacciform bracts and bracteoles, calyx reduced to minute teeth, and filaments inserted at the si- nuses. It is further supported by having corolla lobes imbricate in bud (parallelisms in Wittmack- anthus and Blepharidium). Semaphyllanthe (clade 54) is supported only by more or less strongly homoplasious characters. Alseis (clade 48) is well supported as monophy- letic by five unique synapomorphies: protogynous flowers, corolla open in bud (reversed in at least one species excluded from this analysis), corolla 418 Volume 82, Number 3 1995 419 Andersson Cinchoneae Complex urceolate, puberulous style (reversed in several spe- cies excluded from this analysis), and placenta at- tached apically by a stipe. It is further supported by two rare character states that are more or less constant in the genus: subspicate (reduced thyr- siform) inflorescences (parallelisms in Wittmack- anthus and Hymenodictyon), and filaments at- tached close to the corolla base (parallelism in Calycophyllum intonsum). Its sister-group rela- tionship (clade 47) to the Calycophyllum group is supported by long recurved stigma branches (nu- merous reversais in the Calycophyllum group) and verruculose thickenings on the inner tangential wall of exotesta cells (reversai in Calycophyllum in- tonsum). Ferdinandusa (clade 59) is supported as mono- phyletic by three unique synapomorphies: colleters placed in a continuous row along the base on the inside of the calyx, connective dorsally widened and reinforced, and lumina of the exine differen- tiated into two distinct size classes. It is further supported by one rare character state, large seed size (parallelisms in species of Ladenbergia and Remijia). Macrocnemum (clade 57) is supported as mono- phyletic by three unique synapomorphies: corolla reduplicate in bud, corolla tube minutely papillose inside below the stamen attachments, and fruits loculicidally dehiscent throughout. It is further sup- ported by three rare apomorphic character states: corolla lobes marginally minutely papillose (paral- lelisms in Ladenbergia and Remijia), thinnings in the corolla tube tissue running down from the si- nuses to the stamen attachments (parallelism in Cinchona), and inner tangential wall of the exotesta cells with sheetlike secondary thickening with pores (parallelism in Joosia dichotoma). The sister-group relationship between Macroc- nemum and Wittmackanthus (clade 56) is sup- ported only by homoplasious characters (reddish corolla, small seeds, foveolate exine). Steyermark & Kirkbride (1975) concluded that the closest relative of Wittmackanthus is Calycophyllum. However, Wittmackanthus lacks the critical syn- apomorphies of the Calycophyllum group, such as internally verruculose exotesta cells and curved style branches. Its position within the Calycophyl- leae seems rather uncertain. Clade 60, which includes the remaining genera of the Cinchoneae and Hillia, is supported as monophyletic by a single unique synapomorphy in the runs of Step I, obtuse to truncate colpus ends (further transformed in Blepharidium and Doli- cholobium, where pollen grains are porate). Step II (Fig. 2) generated more than 1000 min- imal trees in the run with unweighted characters, 234 steps long and with overall consistency indices of 0.490. After successive reweighting 569 mini- mal trees were found, 235 steps long (as calculated with unweighted characters) and with overall con- sistency indices of 0.579. The HER for the tree of Figure 2A was estimated at 0.485. The differences in topology between the trees from Step I (Fig. 1) and those from Step II (Fig. 2) are striking. These differences seem to depend mainly on the fact that different outgroups are used. Since a closer outgroup is used in Step II, and since the ingroup contains a fuller selection of terminal taxa, edge lengths are shorter on the av- erage. One is therefore justified in putting greater confidence in the results of Step II than those of Step I (see, e.g., Penny et al., 1992). The ingroup (clade 44) is supported as mono- phyletic by four unique synapomorphies: included anthers, elongate to linear anthers, long connivent stigma branches, and obtuse to truncate coipus ends. It is further supported by two homoplastic character states: short filaments (reversed in Pi- mentelia and Stilpnophyllum) and foveolate tec- tum (reversed in Hillia parasitica and with further parallel transformations to psilate in Blephari- dium, internally in Cosmibuena grandiflora, and in Dolicholobium). Dolicholobium comes out as a basal clade (40) in Step II analyses, in contrast to those of Step I, where it cornes out in a more distal position within a basal paraphyletic complex. Its position was vari- able also among the preliminary runs. This uncer- tainty may indicate that the assumption that it belongs to the Cinchoneae complex is faulty. Both the run with unweighted and that with weighted characters indicate a basal dichotomy between one clade (45) comprising Balmea, Ble- pharidium, Cosmibuena, and Hillia, and another one (48) comprising Cephalodendron, Cinchona, Dolicholobium, Joosia, Ladenbergia, Maguireo- charis, Pimentelia, Remijia, and Stilpnophyl- lum. This result is interpreted as support for the recognition of an extended tribe Hillieae (clade 45) that is the sister group of the tribe Cinchoneae (clade 48). The tribe Hillieae (clade 45) is supported by one unique and consistent synapomorphy, fruit with an apical, beaklike prolongation, and also by the rare character state calyx shed in fruit (parallelisms in Cephalodendron and Wittmackanthus). The "central core" of the tribe (clade 46) is furthermore supported by presence of raphides. The phylogeny of this tribe (excluding Blepharidium) was studied in detail by Taylor (1994). Annals of the Missouri Botanical Garden The tribe Cinchoneae (clade 48) is supported as monophyletic by an externally sericeous corolla (reversais in Stilpnophyllum grandifolium and Maguireocharis). If Dolicholobium is excluded (clade 49) it is supported by another two unique and consistent synapomorphies, corolla lobes val- vate in bud and os of the pollen aperture poorly defined. The genera Maguireocharis, Pimentelia, and Stilpnophyllum, together with Cinchona amazon- ica, form a monophyletic group (clade 77) that is the sister group of the rest (clade 50) of the Cin- choneae. Clade 77 is supported as monophyletic by two synapomorphies, corolla tube hirsute inside above stamens and corolla lobes hirsute ail over adaxially. It is also set aside from the rest of the Cinchoneae by lacking the unique and consistent synapomorphy of clade 50, heterostylous flowers. The genus Stilpnophyllum (clade 80) is supported as monophyletic by its unusual leaf venation, with secondary veins very thin and closely set. Clade 79 is supported as monophyletic by a campanulate corolla, but since this character may easily be erroneously interpreted in very short-tubed flowers, it should not be overemphasized. Clade 78 is sup- ported only by homoplastic characters. Cinchona amazonica must obviously be removed from the genus Cinchona (clade 71), in which it was sup- posedly originally placed based mainly on its acrop- etal fruit dehiscence. Recognizing that the reso- lution of clade 77 may not be particularly stable, I suggest that it be referred to a genus of its own, Cinchonopsis, formally described in the Synopsis. The sister group relationship of Cinchona and Joosia (clade 70) is supported by the joint posses- sion of reddish flowers, a character with several parallelisms and also reversais within Joosia. It is not supported in ail solutions (Fig. 2C), and the two genera may corne out as successive basal branchings of clade 50. Both genera are excluded from clade 51 by lack of two unique synapomor- phies. Joosia (clade 74) is supported as monophyletic by four unique and consistent synapomorphies: per- sistent stipules, predominantly monochasial partial inflorescences, corolla lobes with submarginal, showy appendages, and inner tangential exotesta walls with sheetlike thickening. It is further supported by absence of inflorescence bracts (parallelism in Dolicholobium) and hairy disk (parallelisms in Cin- chona calisaya and C. amazonica). The segre- gation of Joosia dichotoma as a distinct subgenus, Sectocalyx, as suggested by Steyermark (1975), would leave the autonymous subgenus Joosia para- phyletic. Steyermark seems to have based his de- cision entirely on the occurrence in J. dichotoma of a unique autapomorphy, the spathaceous calyx. Cinchona (clade 71) is supported as monophy- letic by two unique and consistent synapomorphies, corolla lobe margins villous with thin-walled hairs and filaments distinctly different in length in brevi- and macrostylous flowers. It is further supported by the corolla tube with longitudinal thinnings (par- allelism in Macrocnemum) and acropetal fruit de- hiscence (consistent but with numerous parallelisms in Ladenbergia and Remijia). Cinchona hen- leana, traditionally (e.g., Standley, 1930, 1931a; Steyermark, 1974) included in Cinchona on the basis of its acropetally dehiscent fruits, is nested within the Ladenbergia clade and should be re- moved (see further below). The sister-group relationship (clade 51) between Ladenbergia and Remijia is supported by two unique and consistent synapomorphies, corolla lobes with the apical part ridged and wedge-shaped and corolla lobe margins papillose. These two character states are conceivably functionally coupled, which may weaken the support for the clade. Ladenbergia (clade 60) is supported as mono- phyletic by only two homoplasious characters, stip- ules fused at base (with several reversais and par- allelisms, and a further transformation to fused throughout in clade 69) and large seeds (parallelism in clade 58). Cinchona henleana is nested within the Ladenbergia clade. It seems, intuitively, close- ly related to, if at ail distinct from, two other species traditionally (e.g., Standley, 1930) treated here, L. muzonensis (Goudot) Standley and L. hooker- iana (Wedd.) Standley. These two were separated as section Muzonia by Weddell (1849) and later elevated to generic rank by Osorio (1874). Judging from the present analysis, the recognition of Mu- zonia, as advocated by Sullivan & Steyermark (1983) would, however, force a fragmentation of the remaining part of Ladenbergia and seems, therefore, to be ill advised. The sections recognized by Weddell (1849) under the synonymous name Cascarilla seem to be in fair agreement with the present results, but are not really put to the test due to only partial correspondence in selection of taxa. The genus Remijia (clade 52) is supported as monophyletic by one synapomorphy, terminal in- florescence absent, which is consistent but with numerous parallelisms. Clade 53, comprising ail species except R. pedunculata and R. purdieana (clade 59), is further characterized by a unique synapomorphy, thyrsoid inflorescences (further transformed to capitate in Cephalodendron glo- bosum). Clade 59 is supported as monophyletic by 420 Volume 82, Number 3 1995 Andersson Cinchoneae Complex marginally ciliate seed wings. The apically split capsule valves (fruit with partial loculicidal dehis- cence), traditionally emphasized as a distinctive character of Remijia, is a highly homoplasious one. The genus Cephalodendron is nested within the Remijia clade and should be regarded as an apo- morphic derivative of this genus; recognition of Cephalodendron would force the pulverization of Remijia. SYNOPSIS Tribe Calycophylleae L. Andersson & C. Pers- son, Pl. Syst. Evol. 178: 89. 1991. TYPE: Calycophyllum DC. Trees or rarely shrubs. Raphides absent. Stipules usually clasping to contorted in bud, rarely held flat back-to-back, or calyptrate. Corolla lobes con- torted, imbricate, open, or reduplicate in bud; fil- aments long, often bristly hirsute at base, anthers button-shaped to elongate, rarely linear, dorsifixed near middle, or basifixed; ovary multiovulate, with distinct placentae attached to septum by a lamella, or apically by a stipe. Pollen 3-colporate, the colpi with acute ends and os distinctly demarcated, or rarely 3-porate, exine usually reticulate, rarely � foveolate. Fruit a usually septicidal capsule with a short loculicidal fissure at apex of valves, or (Ma- crocnemum) loculicidal. Seeds lenticular or pla- noconvex, with exotestal wing, hilum located cen- trally or marginally on seed body, or on wing, exotesta cells with bandlike thickenings on inner tangential wall, or (in Macrocnemum) the thick- ening sheetlike with pores. Alseis Schott in Sprengel, Syst. Veg. 4(2): 404. 1827. TYPE: Alseis floribunda Schott in Sprengel. Alseis is recognized by its subspicate (reduced thyrsiform) inflorescences, protogynous flowers with urceolate corolla having the lobes usually open in bud, and its pendent placentae. It is a neotropical genus supposed (Andersson, 1992) to comprise about 16 species. It is badly in need of taxonomic revision, however. Calycophyllum DC., Prodr. 4: 367. 1830. TYPE: Macrocnemum candidissimum M. Vahl = Calycophyllum candidissimum (M. Vahl) DC. Eukylista Benth., J. Bot. (Hooker) 5: 230. 1853. TYPE: Eukylista spruceana Benth. = Calycophyllum spruceanum (Benth.) J. D. Hook. ex Schumann. Calycophyllum is recognized by its calyptrate stipules and sacciform bracts and bracteoles, calyx (except semaphylls) reduced to minute teeth (to sometimes totally obsolete), corolla lobes that are imbricate in bud, filaments inserted on the corolla at the sinuses, corolla tube bristly hirsute inside in the upper part, and glabrous filaments. It is a genus of three species distributed from northern Central America and the Greater Antilles to northern Ar- gentina and Paraguay. Capirona Spruce, J. Proc. Linn. Soc., Bot. 3: 200. 1859. TYPE: Capirona decorticans Spruce. Monadelphanthus Karsten, FI. Columb. 1: 67. 1860. TYPE: Monadelphanthus floridus Karsten = Ca- pirona decorticans Spruce. Loretoa Standley, Field Mus. Nat. Hist., Bot. Ser., 11: 221. 1936. TYPE: Loretoa peruviana Standley = Capirona decorticans Spruce. Capirona is recognized by its intrapetiolar, per- sistent stipules, corolla with lobes contorted in bud and tube conspicuously plicate, filaments inserted near the base of the corolla tube and interconnected at base by a callosity, basifixed anthers, pollen with reticulate exine with very narrow muri, wide lu- mina, and numerous free, rodlike structures within the lumina, and by seeds with the hilum located on the wing. It is a monotypic genus (Andersson, 1994) distributed more or less throughout the Am- azon basin and the Guianas. Semaphyllanthe L. Andersson, gen. nov. TYPE: Semaphyllanthe obovata (Ducke) L. Anders- son = Calycophyllum obovatum (Ducke) Ducke = Warszewiczia obovata Ducke. Calycophyllo similis sed differt stipulis liberis, bracteis planis, calyce distincto, corolla aestivatione contorta, fi- lamentis ad basin hirsutis, ad medium corollae tubo, vel basin versus affixis. Trees. Leaves decussate; stipules interpetiolar, not fused at base, � triangular, clasping in bud, caducous. Inflorescences terminal and lateral, cy- mose, clustered to form � corymbose synfloresc- ences at branch tips. Flowers small, homostylous, normally 4-merous; calyx � cup-shaped, with dis- tinct or indistinct lobes, some flowers usually (ex- cept in S. megistocaula) with one of the lobes expanded into a large semaphyll; corolla � cam- panulate, tube � equaling the lobes, or shorter, glabrous outside, densely bristly hirsute inside above stamen insertion, lobes contorted in bud, glabrous; stamens attached to corolla tube in central part or lower, filaments longer than anthers, bristly hirsute in lower part; anthers button-shaped, dorsifixed near middle, long exserted; stigma branches con- 421 Annals of the Missouri Botanical Garden nivent or widely recurved. Capsules septicidal with basipetal dehiscence, crowned by the persistent calyx. Seeds winged, hilum located � centrally on seed body, exotesta cells with bandlike, � verru- culose secondary thickenings on inner tangential wall. SPECIES INCLUDED: Semaphyllanthe intonsa (Steyerm.) L. Anders- son, comb. nov. Basionym: Calycophyllum intonsum Steyerm., Mem. New York Bot. Gard. 10(5): 186. 1963. Semaphyllanthe megistocaula (K. Krause) L. Andersson, comb. nov. Basionym: Remijia megistocaula K. Krause, Bot. Jahrb. 40: 319. 1908. Semaphyllanthe merumensis (Steyerm.) L. Andersson, comb. nov. Basionym: Calyco- phyllum merumense Steyerm., Mem. New York Bot. Gard. 23: 228. 1972. Semaphyllanthe obovata (Ducke) L. Anders- son, comb. nov. Basionym: Warszewiczia obovata Ducke, Notizbl. Bot. Gart. Berlin- Dahlem 11: 475. 1932. Semaphyllanthe spectabilis (Steyerm.) L. An- dersson, comb. nov. Basionym: Calycophyl- lum spectabile Steyerm., Mem. New York Bot. Gard. 10(5): 188. 1963. Semaphyllanthe venezuelensis (Steyerm.) L. Andersson, comb. nov. Basionym: Calyco- phyllum venezuelense Steyerm., Mem. New York Bot. Gard. 10(5): 189. 1963. Semaphyllanthe differs from Calycophyllum in having free (vs. calyptrate) stipules, plane (vs. sac- cate) bracteoles, a distinct tubular calyx (vs. one reduced to minute teeth), contorted (vs. imbricate) corolla aestivation, filaments inserted in center of corolla tube or lower (vs. at top), and filaments bristly hirsute toward base (vs. glabrous through- out). It is heavily centered on the Guiana shield and in the white sand campinas of the northern Amazon basin. Only the aberrant species S. meg- istocaula occurs outside of this area, in the south- western Amazon. Semaphyllanthe megistocaula is included in the genus for practical reasons only: it shares the (plesiomorphic) diagnostic character states with the rest of the species, but is not reliably associated with them in the cladistic analysis. Ferdinandusa Pohl, Pl. Bras. Icon. Descr. 2: 8. 1828. TYPE: Based on Ferdinandea Pohl. Ferdinandea Pohi, Flora 10: 153. 1827, non Ferdinan- da Lagasca. TYPE: Not designated. Aspidanthera Benth., J. Bot. (Hooker) 3: 217. 1841. TYPE: Aspidanthera rudgeoides Benth. = Ferdi- nandusa rudgeoides (Benth.) Wedd. Gomphosia Wedd., Ann. Sci. Nat. Bot., S�r. 3, 10: 14. 1848. TYPE: Not designated. Ferdinandusa is recognized by its free, con- torted stipules, its more or less cochlear corolla lobes that are contorted in bud, its anthers with a dorsally widened and reinforced connective, its re- ticulate exine with two distinct size classes of lu- mina, and its large seeds. It is a neotropical genus heavily centered in the Amazon basin and is sup- posed to contain ca. 25 species (Andersson, 1992). It seems badly in need of a taxonomic revision, however. Macrocnemum P. Browne, Civ. Nat. Hist. Ja- maica 165. 1756. TYPE: Macrocnemum ja- maicense L. Lasionema D. Don, London Edinburgh Philos. Mag. & J. Sci. 2: 377. 1833. TYPE: Lasionema roseum (Ruiz L�pez & Pav�n) D. Don = Cinchona rosea Ruiz L�pez & Pav�n = Macrocnemum roseum (Ruiz L�pez & Pav�n) Wedd. Macrocnemum is recognized by its reduplicate corolla lobes which are minutely hirtellous along the margins, its corolla tube which is minutely papillose inside below the filament attachments and which has linear thinnings running down from the sinuses to the filament bases, its loculicidal cap- sules, and its small seeds with the hilum located terminally or subterminally on the basiscopic wing. It is distributed in southern Central America, on Jamaica, and in subandean South America south to central Bolivia. In spite of its having ca. 14 presently recognized species (Andersson, 1992), the number of good species is probably less than half a dozen. Tribe Hillieae Bremek. ex S. Darwin, Taxon 25: 603. 1976. TYPE: Hillia Jacq. Trees or shrubs, rarely subshrubs, often epi- phytic or chasmophytic. Raphides usually present (absent in Blepharidium). Stipules clasping to con- torted in bud, or (usually) held flat back-to-back. Corolla lobes contorted or (Blepharidium) imbri- cate in bud; filaments short, glabrous, anthers lin- ear, dorsifixed near middle or near base, or basi- fixed; ovary multiovulate, with distinct placentae attached to septum by a lamella, ovules ascendingly imbricate. Pollen 3-colporate, colpi with obtuse ends, os distinctly demarcated, or pollen 3-porate, exine reticulate, or foveolate to � psilate. Fruit a usually septicidal capsule, apical portion (above calyx scar) prolonged and � beaklike, calyx shed after an- thesis. Seeds lenticular or planoconvex to ellipsoid, 422 Volume 82, Number 3 1995 423 Andersson Cinchoneae Complex with exotestal wing (rudimentary in Hillia), in Hil- lia also with long exotestal trichomes, hilum located centrally or marginally on seed body, exotesta cells with bandlike thickenings on inner tangential wall. The tribe Hillieae is recognized mainly by its large woody capsules with the apical portion more or less prolonged and beaklike in mature fruit, and from which the calyx is shed soon after anthesis. The central core is formed by the genera Balmea, Cosmibuena, and Hillia. These are further char- acterized by the presence of raphides and contorted aestivation. Hillia Jacq., Enum. Syst. Pl. 3. 1760. TYPE: Hillia parasitica Jacq. Ravnia Oersted, Vidensk. Meddel. Dansk Naturhist. For- en. Kj0benhavn 1852: 49. 1852. Hillia subg. Rav- nia (Oersted) C.M. Taylor, Selbyana 11: 30. 1989. Hillia is characterized primarily by its rudi- mentary seed wing and presence of long exotestal hairs. It comprises 24 species distributed from southern Mexico through Central America and the West Indies to southern Brazil and Bolivia. The genus was monographed by Taylor (1994). Cosmibuena Ruiz L�pez & Pav�n, FI. Peruv. 3: 2. 1802, nom. cons., non Cosmibuena Ruiz L�pez & Pav�n, FI. Peruv. Prodr. 10. 1794, nom. rejic. TYPE: Cosmibuena obtusifolia Ruiz L�pez & Pav�n, typus cons. = Cosmi- buena grandiflora (Ruiz L�pez & Pav�n) Rus- by = Cinchona grandiflora Ruiz L�pez & Pav�n. Buena Pohl, Pl. Bras. Icon. Descr. 1: 8. 1826, non Cav. 1800. TYPE: Cosmibuena grandiflora (Ruiz Lupez & Pav�n) Rusby = Cinchona grandiflora Ruiz L�pez & Pav�n. Cosmibuena is distinguished from Hillia by its stipules being fused intrapetiolarly at base, its well- developed seed wing, and the absence of exotestal trichomes. It comprises four species and is distrib- uted from southern Mexico to central Bolivia and southeastern Brazil. It was revised by Taylor (1992). Some of the cladistic analyses performed by Taylor (1994) indicate that Cosmibuena may be para- phyletic when Hillia is separated. Balmea Martinez, Bull. Torrey Bot. Club 69: 438. 1942. TYPE: Balmea stormae Martinez. Balmea differs from Cosmibuena mainly in hav- ing free stipules, a cymose, many-flowered inflo- rescence, and a deep red to dark purplish corolla. It is a monotypic genus with the single species, B. stormae, ranging from the state of Michoac�n in central Mexico to northern Guatemala. Blepharidium Standley, J. Wash. Acad. Sci. 8: 59. 1918. TYPE: Blepharidium guatema- lense Standley. Blepharidium agrees with other genera of the Hillieae in having large, woody capsules with a beaklike apical prolongation, and in having a ca- ducous calyx. It differs from other members of the tribe in the absence of raphides, in having clasping stipules, corolla with imbricate aestivation, and an- thers dorsifixed around the middle. It has been considered (Standley, 1940) to comprise two spe- cies, but B. mexicanum Standley is hardly specif- ically distinct from B. guatemalense. The genus is distributed from southern Mexico to northern Guatemala. Tribe Cinchoneae DC., Ann. Mus. Natl. Hist. Nat. 9: 217. 1807, as "Cinchonaceae." TYPE: Cinchona L. Trees or shrubs. Raphides absent. Stipules held flat back-to-back in bud, or (in species with ver- ticillate leaves) held to form a cone and then mar- ginally overlapping, free or intrapetiolarly fused at base, or calyptrate. Corolla lobes valvate in bud; filaments usually short or rather short, but often rather long in brevistylous morphs of heterostylous species, glabrous, anthers elongate to linear, dor- sifixed near base; ovary multiovulate, with distinct placentae attached to septum by a lamella, ovules ascendingly imbricate. Pollen 3-colporate, colpi with obtuse to truncate ends, os poorly demarcated, exine foveolate. Fruit a septicidal capsule, apical portion not prolonged, calyx persistent. Seeds len- ticular or planoconvex, with exotestal wing, hilum located centrally or marginally on seed body, ex- otesta cells with bandlike thickenings on inner tan- gential wall. Pimentelia Wedd., Hist. Nat. Quinquinas 94. 1849. TYPE: Pimentelia glomerata Wedd. Pimentelia is strongly reminiscent of Stilpno- phyllum in having strongly "resinous" buds and young shoots, ail inflorescences lateral, corolla strongly hairy inside in the upper part of the tube and ail over the adaxial surface of the lobes, the hairs being coarse and thick-walled, and in having small capsules and small seeds. It differs, however, in having widely (vs. very narrowly) spaced lateral veins with a distinct tertiary venation in between. Annals of the Missouri Botanical Garden Stilpnophyllum J. D. Hook., in Benth. & J. D. Hook., Gen. Pl. 2: 33. 1873. TYPE: Stilp- nophyllum lineatum J. D. Hook. Stilpnophyllum differs from Pimentelia mainly in leaf venation, having a large number (> 25 pairs) of fine, closely set secondary veins. It is a genus of four species (Andersson, 1994) distributed in the Andes of Ecuador and Peru. Cinchonopsis L. Andersson, gen. nov. TYPE: Cinchonopsis amazonica (Standley) L. An- dersson = Cinchona amazonica Standley. A Cinchona differt flore parvo corolla alba campanulata tubo distali parte et lobis in tota superficiei grosse pubes. centibus. A Maguireocharide, Pimentelia, et Stilpno- phyllo differt praesentia inflorescentia terminali et capsula dehiscentia basifugi. Trees. Leaves decussate; stipules interpetiolar, shortly fused across petiole at base, lingulate to triangular, held pressed back-to-back in bud, ca- ducous. Inflorescences terminal and lateral, cy- mose. Flowers small, homostylous, normally 5-merous; calyx � cup-shaped, with distinct lobes; corolla � campanulate, white, tube � equaling the lobes, or shorter, � puberulous outside, densely bristly pubescent inside above stamen insertion, lobes valvate in bud, densely pubescent throughout on adaxial side with thick-walled hairs; stamens attached to corolla tube in central part, filaments � equaling anthers, glabrous; anthers elongate, dorsifixed near base, with a small apical connective process, semi-exserted; stigma branches connivent. Capsules septicidal with acropetal dehiscence, crowned by the persistent calyx. Seeds winged, hilum located � centrally on seed body, exotesta cells with bandlike, smooth secondary thickenings on inner tangential wall. SPECIES INCLUDED: Cinchonopsis amazonica (Standley) L. Anders- son, comb. nov. Basionym: Cinchona ama- zonica Standley, Publ. Field Columbian Mus., Bot. Ser., 8: 334. 1931. Standley (1931b) compared Cinchona amazon- ica to C. micrantha Ruiz L�pez & Pav�n and seems not to have been aware of the striking dif- ferences in floral structure between C. amazonica and other species of Cinchona. The decisive char- acter for his placement of his new species was probably the acropetally dehiscent capsules. How- ever, it differs from Cinchona in having small, homostylous flowers with a white, campanulate co- rolla, which is coarsely and densely pubescent in- side in the distal part of the tube and throughout the adaxial surface of the lobes, and in the absence of thinnings on the corolla tube. Cinchonopsis dif- fers from Maguireocharis, Pimentelia, and Stilp- nophyllum in having terminal inflorescences and acropetally dehiscent capsules. The only species seems to be widely distributed in the central and western Amazon basin and has been recorded from Peru, Colombia, Brazil (Amazonas and Roraima), and Venezuela. Joosia Karsten, FI. Columb. 1: 9. 1858. TYPE: Joosia umbellifera Karsten. Joosia is recognized by its persistent stipules, the showy marginal appendages on the corolla lobes, fruits with valves that are usually deeply split lo- culicidally and with the halves often conspicuously coiled. The inflorescences are predominantly mon- ochasial but this is sometimes not readily seen when they are congested into headlike clusters. Joosia is a genus of seven species (Andersson, 1994) distributed from Panama along the Andes to north- ern Bolivia; it is heavily centered in the Ecuadorean Andes. Cinchona L., Sp. Pl. 172. 1753. Gen. Pl. ed. 5 79. 1754. TYPE: Cinchona officinalis L. Kinkina Adans., Fam. PI. 2: 147. 1763. TYPE: Based on Cinchona L. Cinchona [sect.] Quinquina Endl., Gen. PI. 556. 1838. TYPE: Cinchona officinalis L. (lectotype, desig- nated by Andersson, 1994). Cinchona is recognized, above ail, by the mar- ginally villose corolla lobes, the hairs of which are thin-walled and somewhat widened distally. It is further characterized by having terminal inflores- cences, reddish (pink to purple) flowers, by the corolla lobes not being wedge-shaped at the apex and by the linear thinnings present in the corolla tube. The acropetal dehiscence of the capsule pre- viously used to distinguish Cinchona on the one hand and Ladenbergia and Remijia on the other (e.g., Schumann, 1891), seems to be a constant feature of Cinchona but is not a reliable diagnostic feature; it occurs in Cinchonopsis and many spe- cies of Remijia and Ladenbergia as well. Cin- chona is a genus of 20 to 25 species distributed from Costa Rica and the coastal mountains of Ven- ezuela along the Andes to central Bolivia; it is heavily centered in the Andes of Ecuador and northern Peru. Ladenbergia Klotzsch in Hayne, Getreue Dars- tell. Gew. 14: tab. 15. 1846. TYPE: Lad- enbergia moritziana Klotzsch in Hayne (lec- totype, designated by Andersson, 1994). 424 Volume 82, Number 3 1995 Andersson Cinchoneae Complex Cinchona [sect.] Cascarilla Endl., Gen. PI. 556. 1838. Cascarilla (Endl.) Wedd., Ann. Sci. Nat. Bot., S�r. 3, 10: 10. 1849. TYPE: Ladenbergia macrocarpa (M. Vahl) Klotzsch in Hayne = Cinchona macro- carpa M. Vahl (lectotype, designated by Andersson, 1994). Cascarilla [sect.] Pseudoquina Wedd., Hist. Nat. Quin- quinas 79. 1849. TYPE: Not designated. Cascarilla [sect.] Carua Wedd., Hist. Nat. Quinquinas 83. 1849. TYPE: Cascarilla carua Wedd. = Lad- enbergia carua (Wedd.) Standley (lectotype, des- ignated by Andersson, 1994). Cascarilla [sect.] Calyptrata Wedd., Hist. Nat. Quin- quinas 89. 1849. TYPE: Cascarilla macrocarpa (M. Vahl) Wedd. = Cinchona macrocarpa M. Vahl = Ladenbergia macrocarpa (M. Vahl) Klotzsch in Hayne (lectotype, designated by Andersson, 1994). Cascarilla [sect.] Muzonia Wedd., Hist. Nat. Quinquinas 89. 1849. Muzonia (Wedd.) N. Osorio, Estudio Quinas Est. Unid. Colombia, ed. 2, 24. 1874. TYPE: Cascarilla muzonensis (Goudot) Wedd. = Cincho- na muzonensis Goudot = Ladenbergia muzonensis (Goudot) Standley. Ladenbergia is recognized by the presence of terminal inflorescences, white flowers, corolla lobes with a wedge-shaped terminal portion which is mi- nutely papillose, and, usually, by large seeds (>: 11 mm long). It comprises ca. 40 species (An- dersson, 1992) and has a wide distribution com- prising southern Central America and most of hu- mid tropical South America; it has conspicuous centers of diversity in the Guiana Highlands and the Northern Andes. Remijia DC., Biblioth. Universelle Sci., Sci. Arts 41: 155. 1829. TYPE: Remijia ferruginea (A. St. Hil.) DC. = Cinchona ferruginea A. St. Hil. (lectotype, designated by Andersson, 1994). Cephalodendron Steyerm., Mem. New York Bot. %,ard. 23: 228. 1972. TYPE: Cephalodendron globosum Steyerm. (holotype). Remijia is recognized by the absence of terminal inflorescences, usually thyrsiform inflorescences, white flowers, and by corolla lobes with the terminal part wedge-shaped and minutely papillose. It com- prises ca. 40 presently recognized species (An- dersson, 1992), but is partly in need of revision. It is widely distributed in tropical South America, but heavily centered on the western Guiana shield and adjacent parts of the Amazon basin. Literature Cited Andersson, L. 1992. A provisional checklist of neo- tropical Rubiaceae. Scripta Bot. Belg. 1: 1-200. . 1993. Pollen characteristics of the tribes Ca- lycophylleae, Cinchoneae, and Hillieae (Rubiaceae). Nordic J. Bot. 13: 405-417. -- . 1994. Tribes Cinchoneae, Calycophylleae, and Coptosapelteae. In: L. Andersson & C. M. Taylor, Rubiaceae pt. 2. Flora of Ecuador (edited by G. Harling & L. Andersson) 50: 1-112. - & C. Persson. 1991. Circumscription of the tribe Cinchoneae (Rubiaceae)-A cladistic approach. Pl. Syst. Evol. 178: 65-94. Archie, J. W. 1989. Homoplasy excess ratios: New indices for measuring levels of homoplasy in phylo- genetic systematics and a critique of the consistency index. Syst. Zool. 38: 253-269. Bacigalupo, N. M. 1952. Sinopsis preliminar del g�nero "Calycophyllum" (Rubiaceae). Darwiniana 10: 145- 156. Candolle, A. P. de. 1830. Prodromus Systematis Na- turalis Regni Vegetabilis, IV. Treuttel & Wiirtz, Par- is. Maddison, W. P. & D. R. Maddison. 1992. MacClade: Analysis of Phylogeny and Character Evolution. Ver- sion 3.0. Sinauer Associates, Sunderland, Massachu- setts. Osorio, N. 1874. Estudio sobre las Quinas de los Estados Unidos de Colombia, ed. 2. Echeverria Hermanos, Bogota. Penny, D., M. D. Hendy, M. D. & M. A. Steel. 1992. Progress with methods for constructing evolutionary trees. TREE 7: 73-79. Robbrecht, E. 1988. Tropical woody Rubiaceae. Opera Bot. Belg. 1: 1-271. Schumann, K. 1891. Rubiaceae. In: A. Engler & K. Prantl (editors), Die natirlichen Pflanzenfamilien 4(4): 1 -156. Engelmann, Leipzig. - . 1897. Rubiaceae. In: A. Engler & K. Prantl (editors), Die natirlichen Pflanzenfamilien, Nach- tr�ge zum II.-IV. Teil (pp. 309-316). Engelmann, Leipzig. Standley, P.C. 1930. The Rubiaceae of Colombia. Publ. Field Columbian Mus., Bot. Ser. 7: 3-175. - . 1931a. The Rubiaceae of Venezuela. Publ. Field Columbian Mus., Bot. Ser. 7: 341-486. - . 1931b. Studies of American plants-V. Publ. Field Columbian Mus., Bot. Ser. 8: 295-398. - . 1940. Studies of American Plants-X. Publ. Field Mus. Nat. Hist., Bot. Ser. 22: 65-129. Stearn, W. T. 1983. Botanical Latin, 3 rev. ed. David & Charles, Newton Abbot. Steyermark, J. A. 1964. Rubiaceae. In: B. Maguire & J. Wurdack, The botany of the Guayana Highland- Part IV. Mem. New York Bot. Gard. 10(5): 186- 278. - 1974. Rubiaceae. In: T. Lasser (editor), Flora de Venezuela 9 (1-3). Instituto Bot�nico, Caracas. . 1975. Revision of the genus Joosia (Rubi- aceae). Brittonia 27: 251-262. - & J. H. Kirkbride, Jr. 1975. The genus Witt- mackanthus (Rubiaceae). Ann. Missouri Bot. Gard. 62: 504-509. Sullivan, G. A. & J. A. Steyermark. 1983. Recognition of the genus Muzonia. Amer. J. Bot. 70 (abstracts): 132. Swofford, D. L. 1993. PAUP: Phylogenetic Analysis Using Parsimony, version 3.1.1. Computer program distributed by the Illinois Natural History Survey, Champaign, Illinois. Taylor, C. M. 1992. Revision of Cosmibuena (Rubi- aceae). Ann. Missouri Bot. Gard. 79: 886-900. - . 1994. Revision of Hillia (Rubiaceae). Ann. Missouri Bot. Gard. 81: 571-609. Weddell, H.-A. 1849. Histoire Naturelle des Quinquinas ou Monographie du Genre Cinchona. Victor Masson, Paris. 425 Annals of the Missouri Botanical Garden Appendix 1. Representative material examined. Alseis eggersii Standley: Dodson & Gentry 12466 (SEL, US); Eggers 15738 (S). A. floribunda Schott in Spreng.: Hatschbach 46834 (GB); Regnell 111. 1524 (S). A. yu- catanensis Standley: Contreras 950 (S); Lundell 17955 (S). Balmea stormae Martinez: Hinton 7894 (F, MO, US); Rzedowski 34593 (NY). Blepharidium guatema- lense Standley: Contreras 9383 (S); Jones & Facey 3433 (NY). Calycophyllum candidissimum (M. Vahl) DC.: Budowski 33 (NY); Bunting 8250 (NY). C. intonsum Steyerm.: Berg et al. in Prance 19493 (S); Prance 3738 (S). C. megistocaulum (Krause) C. M. Taylor: Croat 20768 (AAU); Ducke 24414 (S). C. multiflorum Griseb.: Malme 1070B, 3006 (S). C. obovatum (Ducke) Ducke: Clark & Maquirino 7656 (GB); Ducke 266 (NY). C. spruceanum (Benth.) Hook. f.: Holm-Nielsen et al. 21056 (AAU); Krukoff 5416 (S). C. venezuelense Steyerm.: Maguire et al. 46789 (NY). Capirona decorticans Spruce: Mori & Bolten 8413 (S); Palacios 3225 (GB). Cephal- odendron globosum Steyerm.: Maguire 42467 (NY); Nee 31171 (NY). Cinchona amazonica Standley: Fer- ndndez & Sanoja 4964 (NY); Fosberg 38838 (NY). C. barbacoensis Karsten: Pittier 517 (US); van der Werff et al. 12028 (GB, MO). C. calisaya Wedd.: Krukoff 11374 (NY), Rusby 2347 (NY). C. lancifolia Mutis: Fosberg 19570 (S), 20057 (US). C. macrocalyx Pav. ex DC.: Madsen 75475, 86152 (AAU). C. officinalis L.: Eriksson et al. 222 (GB); Hart 762 (A). C. pubescens Vahl: Harling & Andersson 22436 (GB); Prieto AP-18 (S). Cosmibuena grandiflora (Ruiz L�pez & Pav�n) Rus- by: Maguire et al. 53574 (S); Krukoff 10418 (S). Dol- icholobium oblongifolium A. Gray: A. C. Smith 5109, 7296 (S). Ferdinandusa elliptica Pohl: Hatschbach 44786 (GB). F. gouc'otiana Schumann: Tillett et al. 45184 (S). F. uaupensis Spruce ex Schumann: Berry 2168 (NY), Clark 7245 (NY). Hillia parasitica Jacq.: Shafer 3297 (S); Tollsten 141 (GB). Hymenodictyon floribundum (Hochst. & Steud.) Robinson: "M. R. " 2285 7 (NY); Taylor 3399 (S). Joosia aequatoria Steyerm.: Har- ling & Andersson 23643 (GB). J. dichotoma (Ruiz L�pez & Pav�n) Karsten: Schunke 1100 (NY), 5442 (S). J. dielseana Standley: Steere 8026 (NY); Steyermark 54262 (F). J. umbellifera Karsten: Grant 9983 (S); von Snei- dern 5005 (S). Ladenbergia amazonensis Ducke: Wil. liams 14493 (NY). L. carua (Wedd.) Standley: Krukoff 11021 (NY); Rusby G-1 (NY). L. crassifolia (DC.) Stan- dley: Camp E-4793 (NY), Prieto AP-26B (NY). L. lam- bertiana (C. Martius) Klotzsch in Hayne: Maguire 32864 (NY); Tillett & Tillett 45276 (NY). L. macrocarpa (Vahl) Klotzsch in Hayne: Camp E-4322 (NY); Jaramillo 9352 (AAU). L. moritziana Klotzsch in Hayne: Steyermark 56925 (NY). L. oblongifolia (Mutis) L. Andersson: Camp E-43 (NY), E-55 (S). L. pavonii (Lambert) Standley: Dodson & Gentry 9716 (F); Gentry et al. 54751 (GB). L. sericophylla Standley: Asplund 18914 (S); Fosberg 22474 (NY). L. undata Klotzsch in Hayne: Steyermark 56107, 56178 (NY). Luculia intermedia Hutch.: Rock 7032 (S); Tsai 55159 (S). Macrocnemum cuencanum Standley: Prieto CP-14 (NY); Steyermark 52709 (F). M. jamaicense L.: Proctor 9999 (NY). M. roseum (Ruiz L�pez & Pav�n) Wedd.: Palacios 2768 (GB); Williams 5468 (S). Maguireocharis neblinae Steyerm.: Maguire 37346 (NY). Pimentelia glomerata Wedd.: Hodge 6017 (US); Solomon 9290 (GB). Remijia amazonica Schu- mann: Krukoff 7967 (S); Prance et al. 4923 (S). R. ferruginea (A. St.-Hil.) DC.: Regnell 111.90 (S). R. fir- mula (C. Martius) Wedd.: Silva & Pinheiro 4207 (NY). R. pacimonica Standley: Clark 6602 (NY). R. pedun- culata (Karsten) Fliickiger: Fosberg & Fassett 21840 (S); Jaramillo 6936 (GB). R. pilosinervula Steyerm.: Pipoly et al. 7189 (NY); Steyermark & Wurdack 713 (NY). R. purdieana Wedd.: Fosberg & Fassett 21776 (NY). R. vaupesiana Steyerm.: Fr�es 21225 (NY). Stilp- nophyllum grandifolium L. Andersson: Neill & Palacios 9538 (MO); Palacios et al. 8563 (GB). S. oellgaardii L. Andersson: Oellgaard 74918 (AAU). Wittmackan. thus stanleyanus (R. M. Schomb.) Kuntze: Gentry et al. 19052 (MO). Appendix 2. Lists of apomorphies. The two lists that follow use the following format: 7 3:1� 2 (0.500); 11:1� 3 [2](0.545). This should be interpreted thus: edge 7, character 3 transformed from state 1 to state 2 (consistency index 0.500); character 11 transformed from state 1 to state 3 [2 steps](consistency index 0.545). Double arrows (> >) indicate that the change is unambiguous, single arrows (>) that it is ambiguous. The number of steps is given only when distinct from 1. An asterisk indicates that this character was excluded during search in this particular run. Figure 1A.- 2 1:2> >1 (0.125). 4 16:1> >2 (0.250); 36:1�2 (0.333); 60:2�>>1 (0.500). 5 46:1�2 (0.500). 7 3:1>>2 (0.500); 11:1>>3 [2X(0.545); 16: 1 �2 (0.250); 30:1> >3 [2X(0.429); 31:2�> > 1 (0.500); 32:2> > 1 (0.333); 41:2�> > 1(0.200); 45:1> > 3 (0.333); 50:1 �2 (0.091). 8 19:1>�2 (0.300); 27:2�1 (0.500); 34:1�2 (0.200); 50:1�2 (0.091). 10 1:1>>2 (0.125). 12 21:2>>1 (0.857). 13 19:2>>1 (0.300). 14 1:2�>>1 (0.125); 50:1�>>2 (0.091). 15 1:2�1 (0.125); 9:1>�5 (0.625); 16:1�2 (0.250); 20:1�3 (0.500); 27:2�3 (0.500); 29:2>1 (0.333); 32:2> > 1 (0.333); 38:2> >4 (0.571); 44:1> >2 (0.250); 50:1�2 (0.091). 16 1:2�1 (0.125); 11:2�3 (0.545); 51:1�2(0.222). 17 1:2>>1 (0.200). 18 14: 3�1 (0.583). 19 19:1�2 (0.300); 41:1>2 (0.200); 50:2�>>1 (0.091). 20 4:1�2 (0.400); 11:3�4 (0.545); 14:3> >2 (0.583); 45:1 > 3 (0.333); 52:1> > 2 (0.750); 58:2>>3 (0.375). 21 1:2>>1 (0.125); 5:1> >2(0.500); 17:2> >3(0.571); 30:2> >1(0.429); 41:1 >2 (0.200); 55:2> >3 (0.800). 22 9:1> >2 (0.625); 21:1> >4 (0.857); 41:1 > 2 (0.200); 44:1> >2 (0.250). 23 14:2�3 (0.583); 22:1�2 (1.000); *43:1�2 (0.600); *47:1�2 (0.667); 51:1>2 (0.222). 24 6:2�1 (0.500); 58:2�3 (0.375). 25 1:2�1 (0.125). 26 *2:1>>2 (1.000); 5:1>>3 (0.500); 9:1�4 (0.625). 27 *47:1�3 (0.667). 28 9:1�4 (0.625); 46:1> >2 (0.500). 29 6:1> >2 (0.500); *43: 1 > > 2 (0.600); 51:1>2 (0.222). 30 6:1 > > 3 (0.500). 32 14:1>>4 (0.583). 33 19:1>>2 (0.300); 50:1>>2 (0.091). 34 41:2>3 (0.200). 35 4:1�>>2 (0.400); 8:2>>1 (0.333); 36:1>>2 (0.333); 51:2>1 (0.222). 36 15:2> >3 (0.500); 34:1�>2 (0.200). 37 20:2�1 (0.500). 38 4:1�>>2 (0.400); 11:1�>>2 (0.545); 14: 1 � 2 (0.583). 39 19:1> >2 (0.300); 20:2> > 1 (0.500); 34:1 �2 (0.200). 40 4:1 �2 (0.400); 10:1 �3 (0.667); 14:1 > 2 (0.583); 30:2> >3 (0.429); 58:2> > 3 (0.375). 41 9:1>>6 (0.625); 11:3>>4 (0.545); 14: 3> >5 (0.583); 30:2> >3 (0.429); *43:1> >2 (0.600); 45:1�2 (0.333); 46:1�3 (0.500). 42 9:1�5 (0.625); 11:2> > 1 (0.545); 17:2>> 1 (0.571); 18:2>>3 (0.400); 19:1 > > 2 (0.300); 33:1 > > 2 (0.500); 34:1> > 2 426 Volume 82, Number 3 1995 Andersson Cinchoneae Complex (0.200); 38:2>>4 (0.571); 40:1>>3 (0.333); 41:1>3 (0.200); 45:1>>3 (0.333). 45 28:2> >1 (0.667); 30: 2>>1 (0.429); 31:1>>2 (0.500); 32:1>>2 (0.333); 41:1>2 (0.200); 42:2�> > 1 (0.500); 44:2�> > 1 (0.250); 58:2>>1 (0.375). 46 21:1>>2 (0.857); 29:1>2 (0.333); 46:1>2 (0.500). 47 11:2>>1 (0.545); 18: 2>1 (0.400); 38:2>1 (0.571); 60:1�>>2 (0.500). 48 9:1>>5 (0.625); 13: 1>>2 (1.000); 17:2�4 (0.571); 18:1>3 (0.400); 27:2>>1 (0.500); 37:1>>2 (1.000); 38:1>3 (0.571); 39:1>>2 (1.000); 45:1>>3 (0.333). 49 58:1>>2 (0.375). 50 1:2>>1 (0.125); 14:3>>2 (0.583); 46:2>1 (0.500). 51 4:1>>3 [2](0.400); 6:1 > 2 (0.500); 15:2>4 (0.500). 52 10:1 > 2 (0.667); 14:2> >6 (0.583); 17:2> >3 (0.571); 27:2> >3 (0.500); 29:2>>1 (0.333). 53 6:2>>3 (0.500); 38:1>>3 (0.571); 58:1>>2 (0.375). 54 14:2>>1 (0.583); 16: 1>>2 (0.250). 55 *47:1>>2 (0.667); 51:1>>2 (0.222). 56 4:1�>>2 (0.400); 17:2>3 (0.571); 19:1>>2 (0.300); 48:2�1 (0.714); 55:1 �2 (0.800). 57 5:2>>1 (0.500); 14:3>>2 (0.583); 17:3>5 (0.571); 22:1> >3(1.000); 23:1> >2(0.500); 24:1> >2(0.750); 40:1 > > 3 (0.333); 41:2> >3 (0.200); 45:1 > 3 (0.333); 58:1>>2 (0.375); 59:1>>2 (0.667). 58 8:1>>2 (0.333). 59 15:2>> 1 (0.500); 35:1>>2 (1.000); 48: 2�3 (0.714); 56:1> >2 (1.000). 60 5:2� 1 (0.500); 11:2>>3 (0.545); 27:2>>3 (0.500); 49:1>2 (0.500); 50:1>>2 (0.091); 53:1>>2 (1.000). 61 55:1>>2 (0.800). 62 15:2>3 (0.500); 33:1>>2 (0.500). 63 8:1>>2 (0.333); 14:3>1 (0.583); 38:2>>1 (0.571); 49:2>1 (0.500); 52:1>3 (0.750). 64 20:1 > >2 (0.500); 27:3> >2 (0.500); 42:2> > 1 (0.500); 44:2> > 1 (0.250). 65 11:3>>2 (0.545); 17:2>>1 (0.571); 40:1>2 (0.333); 50:2>1 (0.091); 51:1>2 (0.222); 52:3>1 (0.750); 54:1�>>2 (1.000). 66 12:1�2 (1.000). 67 24:1> >2 (0.750); 25:1> >2 (1.000). 68 4:1 >2 (0.400); 14:1> >2 (0.583); 40:2> 1 (0.333). 69 6:1> >2 (0.500). 70 50:1>>2 (0.091). 71 8:2>>1 (0.333); 48:2>>3 (0.714); 51:2>1 (0.222). 72 11:2>>4 [2](0.545). 73 8:2>>1 (0.333); 15:2>3 (0.500); 51:2>1 (0.222). 74 19:1> >2 (0.300); 23:1> >2 (0.500); 24:1> >3 (0.750); 28:2�3 (0.667). 75 50:1�2 (0.091). 76 3:1�2 (0.500); 9:1> >3 (0.625); 10:1> >3 (0.667); 26:1> >2 (1.000); 36:1� 2 (0.333); 40:2>1 (0.333); 41:1>2 (0.200); 59:1>2 (0.667). 77 34:1>>2 (0.200); 59: 2>3 (0.667). 78 14:1>3 (0.583); 21:1>>3 (0.857); 24:1�4 (0.750). 79 6:1�2 (0.500); 11:2�>>1 (0.545); 18:2>1 (0.400); 50:1 >2 (0.091). 80 14:3> 1 (0.583); 28:2� 1 (0.667); 40:2> 1(0.333). 81 7:1> >2 (1.000); 18:1>2 (0.400); 48:2>>1 (0.714); *57:1>2 (1.000). Figure 2A.-1 *14:2>3 (0.517); 19:1>>2 (0.375); 50:2>> 1 (0.273). 2 4:1> >2 (0.375); 45:2>3 (0.667); 52:1> >2 (0.667). 3 9:1> >6 (1.000); *14:2>5 (0.517); *33:2>1 (0.500); 43:1>>2 (0.500); *46:1>>3 (0.667); 55:2>>1 (0.857). 4 1:2>1 (0.250); 5:1>>2 (0.667); 7:2>>3 (1.000); 30:2>1 (0.600); 52:1>>3 (0.667); 55:2>>3 (0.857). 5 9:4>>2 (1.000); *14: 1>>2 (0.517); 21:1>>4 (1.000); 41:1>>2 (0.286); 44:1�>>2 (0.500). 6 8:1�>>2 (0.600). 7 40:1�>>2 (0.333); *46:1> >2 (0.667). 9 4:1> >3 [2](0.375); 50: 2>>1 (0.273). 10 43:1>>2 (0.500). 11 51:2>>1 (0.200); 58:2>>3 (0.375). 12 4:1>>2 (0.375); 50: 2>>1 (0.273). 13 *14:1>>3 (0.517). 14 1:2>>1 (0.250). 15 58:2>>3 (0.375). 16 *14:1>>2 (0.517); 15:2>>4 (0.429); 40:1>>2 (0.333). 17 *14:1>>2 (0.517); 58:2> >3 (0.375). 18 6:1 >2 (0.600); 57:1> >2 (0.333). 20 58:2>>3 (0.375). 21 *14:1>>3 (0.517); 22:1>>2 (0.750). 22 51:2>1 (0.200). 23 37:1>>2 (1.000). 24 1:2>>1 (0.250). 25 *14:1>>2 (0.517); 48:2� 3 (0.692). 26 36:1 �2 (0.500). 28 *14:1>�2 (0.517). 29 43:1>>2 (0.500). 31 11:2>>3 (0.739). 32 *14:1>>4 (0.517). 33 34:1>>2 (0.250). 34 34: 1> >2 (0.250); *59:2>3 (0.667). 35 4:1> >2 (0.375); 8:2>>1 (0.333); *14:1>3 (0.517); 28:1>2 (0.500); 36:1> > 2 (0.500); 40:1> >2 (0.333); 51:2� 1 (0.200). 36 34:1 > 2 (0.250). 37 20:2> > 1 (0.333). 38 4:1 > > 2 (0.375); 11:1>2 (0.739); *14:1>2 (0.517). 39 *14: 1>3(0.517); 19:1> >2 (0.375); 20:2> > 1 (0.333); 34: 1>>2 (0.250); 40:1>>2 (0.333); 50:2>>1 (0.273). 40 4:1>>2 (0.375); 10:1>>3 (0.500); *14:1>2 (0.517); 30:2> >3 (0.600); 52:1> >3 (0.667); 58:2> >3 (0.375). 41 5:1� 2 (0.667); 8:2>>1 (0.333); *14: 1>3 (0.517); *35:1>>2 (1.000); 48:2>>3 (0.692); 50:2>>1 (0.273); 51:1>>2 (0.200); *56:1>>2 (1.000); 58:2>>1 (0.375). 42 4:1>>2 (0.375); 11: 3>>2 (0.739); 17:2>>5 (1.000); 19:1>>2 (0.375); 22:1> >3 (0.750); 23:1> >2 (0.500); 24:1> >2 (0.750); *29: (1.000); 40:1> >3 (0.333); 41:2> >3 (0.286); 45: 1�3 (0.667); *46:1�2 (0.667); 48:2�>>1 (0.692); *59:1>>2 (0.667). 44 1:1>2 (0.250); 15:1>2 (0.429); 28:1> >2 (0.500); 30:1>2 (0.600); *31:2>> 1 (1.000); *32:2> > 1 (1.000); 38:2>1 (0.600); 53:1 > >2 (1.000); 55:1>2 (0.857). 45 27:2>>3 (1.000); 42:1>>2 (1.000); 44:1>�2 (0.500). 46 8:2�1 (0.333); *14: 1>2 (0.517); 15:2>3 (0.429); *33:1>2 (0.500); 38: 1>2(0.600); 49:1�2 (1.000). 47 11:3�4 (0.739); 30:2>3 (0.600); 41:2> > 1 (0.286); 45:1>2 (0.667); 58:2>3 (0.375). 48 20:1>>2 (0.333); 41:2>>1 (0.286). 49 1:3>>2 (0.739); 17:2>>1 (1.000); 54: 1>>2 (1.000). 50 8:2>1 (0.333); 12:1>>2 (1.000). 51 6:1�2 (0.600); 24:1�2 (0.750); 25:1>>2 (1.000). 52 8:1>2 (0.333); 51:1>>2 (0.200). 53 9:1 �4 (1.000). 54 6:2�1 (0.600); 50:2�1 (0.273). 55 19:1>2 (0.375). 56 2:1>>2 (1.000); 5:1>>3 (0.667); *14:1>>2 (0.517); 15:2>3 (0.429). 57 11:2>3 (0.739); 41:1>>2 (0.286). 58 37:1>>3 (1.000); 48:2�3 (0.692). 59 47:1�>>3 (1.000). 60 4:1�>>2 (0.375); 48:2�3 (0.692). 61 11:2�3 (0.739). 62 6:2>1 (0.600). 63 50:2�>>1 (0.273). 64 4:2>1 (0.375); 22:1>>2 (0.750); 51:1>>2 (0.200). 65 11:3�4 (0.739). 66 43:1�2 (0.500); 51:1>2 (0.200). 67 47:1�>>2 (1.000). 68 4:2�>>1 (0.375); 6:2> > 1 (0.600). 69 4:2> >3 (0.375); *14:1>2 (0.517). 15:2> > 1 (0.429); 22:1> >4 (0.750); 57:1> >2 (0.333). 70 15:2>3 (0.429); 19:1>>2 (0.375). 71 23:1>>2 (0.500); 24:1> >3 (0.750); 28:2> >3 (0.500); 40:1> >2 (0.333). 72 6:1>>3 (0.600). 73 6:1>2 (0.600); 50: 2>>1 (0.273); 51:1>>2 (0.200). 74 3:1>>2 (1.000); 9:1 �3(1.000); 10:1 �3(0.500); 26:1�2(1.000); 36:1 > > 2 (0.500); 41:1>2 (0.286); *59:1>2 (0.667). 75 41:2>3 (0.286). 76 19:2>>1 (0.375); 50:2>>1 (0.273). 77 21:1>>3 (1.000); 24:1>>4 (0.750); 51: 1>>2 (0.200). 78 6:1>>2 (0.600); 11:2>1 (0.739); 28:2> 1 (0.500). 79 15:2>3 (0.429); 18:2> > 1 (1.000). 80 7:1> >2 (1.000); 48:2� 1 (0.692); 57:1>2 (0.333). 427