Peabody Museum of Natural History Yale University New Haven, CT 06511 (Received 30 June 1990) Abstract Angiosperm ovules and carpels have numerous characters and character states, not all of which have been utilized in phylogenetic analyses. A character analysis of these two organs is performed, including description of characters and states, and character polarization to identify the ancestral states. In addition, the distributions of these states are examined in the Magnoliidae and putatively basal orders of Hamamelidae, Caryophyllidae, Dilleniidae, Rosidae, Alismatidae (sensu Cronquist 1981), and Liliiflorae (Sensu Dahlgren et al. 1985). These support two hypotheses: the ancestral ovules were orthotropous, bitegmic and crassinucellar, based on current terminology; and ancestral carpels had ascidate morphology and one or two ovules. Separate structural hypotheses for the evolution of ovules and of carpels are interpreted and indicate that anatropous ovules and plicate (conduplicate) carpels are derived. Finally, these analyses suggest that since the dicots Chloranthaceae, Saururaceae, Piperaceae and Amborellaceae and the monocots © Copyright 1991 by the Peabody Museum of Natural History, Yale University. All rights reserved. No part of this publication, except brief quotations for scholarly purposes, may be reproduced without the written permission of the Director, Peabody Museum of Natural History. ISBN No. 0-912532-25-4 Posti la Number 208 23 August 1991 Angiosperm Ovules and Carpels: Their Characters and Polarities, Distribution in Basal Clades, and Structural Evolution David Winship Taylor Dioscoreales (e.g., Stemonaceae and Smilacaceae) have all the suggested ancestral states, they may be most similar to the morphology of the ancestral angiosperm. Key Words Angiosperms, ancestral morphology, carpel, ovule, structural evolution. Introduction Morphological features of ovules and carpels have had a long history of use for understanding the relationships of angiosperms. These applications have been from the level of intrafamilial relationships (e.g., Moeliono 1970; Tobe and Raven 1986, 1987) to the level of angiosperm origin and superfamilial relationships (e.g., Leinfellner 1950; Bailey and Swamy 1951; Philipson 1974, 1977; Dahlgren 1980; Dahlgren et al. 1981). These studies show the potential usefulness of these organs for phylogenetic analyses. This usefulness arises from their generally conservative nature and their numerous characters and character states. In addition, it is possible to polarize the ovule characters with outgroup comparison to identify the ancestral states (e.g., Stebbins 1974). Yet the characters of ovules and carpels are Currently underutilized in phylogenetic studies. The current terminology for ovules appears to conceal characters (Bouman 1984) that may be of phylogenetic and evolutionary Peabody Museum of Natural History Yale University New Haven, CT 06511 (Received 30 June 1990) Abstract Angiosperm ovules and carpels have numerous characters and character states, not all of which have been utilized in phylogenetic analyses. A character analysis of these two organs is performed, including description of characters and states, and character polarization to identify the ancestral states. In addition, the distributions of these states are examined in the Magnoliidae and putatively basal orders of Hamamelidae, Caryophyllidae, Dilleniidae, Rosidae, Alismatidae (sensu Cronquist 1981), and Liliiflorae (sensu Dahlgren et al. 1985). These support two hypotheses: the ancestral ovules were orthotropous, bitegmic and crassinucellar, based on current terminology; and ancestral carpels had ascidate morphology and one or two ovules. Separate structural hypotheses for the evolution of ovules and of carpels are interpreted and indicate that anatropous ovules and plicate (conduplicate) carpels are derived. Finally, these analyses suggest that since the dicots Chloranthaceae, Saururaceae, Piperaceae and Amborellaceae and the monocots © Copyright 1991 by the Peabody Museum of Natural History, Yale University. All rights reserved. No part of this publication, except brief quotations for scholarly purposes, may be reproduced without the written permission of the Director, Peabody Museum of Natural History. ISBN No. 0-912532-25-4 Posti la Number 208 23 August 1991 Angiosperm Ovules and Carpels: Their Characters and Polarities, Distribution in Basal Clades, and Structural Evolution David Winship Taylor Dioscoreales (e.g., Stemonaceae and Smilacaceae) have all the suggested ancestral states, they may be most similar to the morphology of the ancestral angiosperm. Key Words Angiosperms, ancestral morphology, carpel, ovule, structural evolution. Introduction Morphological features of ovules and carpels have had a long history of use for understanding the relationships of angiosperms. These applications have been from the level of intrafamilial relationships (e.g., Moeliono 1970; Tobe and Raven 1986, 1987) to the level of angiosperm origin and superfamilial relationships (e.g., Leinfellner 1950; Bailey and Swamy 1951; Philipson 1974, 1977; Dahlgren 1980; Dahlgren et al. 1981). These studies show the potential usefulness of these organs for phylogenetic analyses. This usefulness arises from their generally conservative nature and their numerous characters and character states. In addition, it is possible to polarize the ovule characters with outgroup comparison to identify the ancestral states (e.g., Stebbins 1974). Yet the characters of ovules and carpels are Currently underutilized in phylogenetic studies. The current terminology for ovules appears to conceal characters (Bouman 1984) that may be of phylogenetic and evolutionary 2 Angiosperm Ovules and Carpels: Their Characters and Polarities significance. In contrast, the terminology for carpels is quite small. This both obscures differences in general carpel types as well as makes detailed comparisons between taxa difficult. This study presents a character analysis of ovules and carpels. Such an analysis not only provides lists of useful characters for phylogenetic analysis, it also suggests hypotheses for structural homologies and ancestral characters states (Neff 1986, Bryant 1989). It is realized that this is only an initial list, aS new characters will continue to be proposed. In addition, the characters are polarized by outgroup comparison, developmental transformation series, and association between characters from different organs as appropriate. Finally, the distribution of these character states is documented in the families at the base of the angiosperm phylogeny (Magnoliidae and basal orders of Hamamelidae, Caryophyllidae, Dilleniidae, Rosidae, Alismatidae [sensu Cronquist 1981], and Liliiflorae [sensu Dahlgren et al. 1985]). These data and analyses are used to suggest two groups of hypotheses. The first hypotheses concern the structure of the ancestral ovules and carpels of angiosperms (based on the polarizations), whereas the second group deals with the structural evolution of ovules and carpels (based on the polarizations, and structural and developmental transformation series). Lastly, | speculate how these data affect our understanding of the hypothetical ancestral angiosperm and the most basally placed monocot. Materials and Methods A total of 77 families of angiosperms were examined in this study. These were chosen to represent and document the diversity of the major basal clades of angiosperms. Sixty- seven of these families were from dicot taxonomic groups circumscribed by Cronquist (1981), and include all the families of the Magnoliidae; the basal two orders of the Hamamelidae (Trochodendrales and Postilla 208 Hamamelidales); the basal two orders of the Caryophyllidae (Polygonales and Plumbaginales) with the addition of the putatively ancestral Phytolaccaceae, Molluginaceae and Caryophyllaceae of the Caryophyllales; the basal order of the Dilleniidae (Dilleniales), and the members of the basal order of the Rosidae (Rosales) which had crassinucellar ovules. Two potentially ancestral groups of monocots were also examined. The first is the ancestral order of Alismatidae (Alismatales), sensu Cronquist (1981), and the second group is the basal order Dioscoreales as defined by Dahlgren et al. (1985). In addition, | examined a number of fossil and extant taxa that are thought to be the sister-group or otherwise cladistically close to angiosperms, based on the recent analyses by Crane (1985, 1988) and Doyle and Donoghue (1986, 1987). These taxa included Bennettitales, Caytonia, corystosperms, glossopterids, Gnetales and Pentoxylon. Data compiled for the angiosperm families are found in the Appendices. General data, compiled from Cronquist (1981) and Davis (1966) are presented for each family. Under each family are entries for representative genera, using information available from the literature. Ovule characters were usually collected from medial sections taken at the embryo sac stage, in order to avoid variation (e.g., additional curvature) introduced during subsequent development of the seed. Carpel characters were obtained from both longitudinal and transverse sections of young carpels, when available. The genera listed under each family document the extent of the variation of each character. In some cases, several genera in a family had identical suites of character states, and these states are represented by one genus in the Appendices. In addition, | attempted to secure data for both the ovules and carpels for the same genus. To describe the ovules and carpels, additional characters and states were created when appropriate. Characters with quantitative measurements, such as ovule angle, were divided into character states 3 Angiosperm Ovules and Carpels: Their Characters and Polarities based on the natural breaks found in the data. No doubt further work will subdivide and more stringently define some of the characters and states used, especially those which are shown to have had multiple evolutionary origins. Additional characters are likely to be added, particularly those relating to developmental morphology. Polarization was determined by: outgroup comparison (Watrous and Wheeler 1981); developmental transformation series (Nelson 1978; Crisci and Stuessy 1980; Kraus 1988); and association of character states from different organs (Crisci and Stuessy 1980). Detailed outgroup comparison, even in ovules where structural homologies to the outgroups exist, is difficult due to the variability in the outgroups and because the exact relationships of the outgroups to the angiosperms is not certain (Crane 1988). Because of this variation in states, the outgroup ovule states for the angiosperm clade were determined by four methods and topologies. The first two were analyzed using the method of Maddison et al. (1984) for resolved cladograms. The first cladogram used was that of Crane (1985, fig. 22). It has Gnetales as the sister-group to angiosperms and a clade composed of Bennettitales and Pentoxylon as the next outgroup. This clade is followed sequentially, towards the base, by corystosperms, Caytonia and glossopterids. The second cladogram, from Doyle and Donoghue (1986), has the angiosperm sister- group composed of a clade with three taxa. These include Gnetales and Pentoxylon as terminal sister-groups and with Bennettitales as a sister-group to those; this clade is followed sequentially, towards the base, by Caytonia and glossopterids. The last two methods were considered since the relationships of the outgroups are uncertain (Crane 1988). One method considered common as ancestral (from the six taxa in the cladogram of Crane 1985), which may be considered a legitimate method when relationships are unresolved (but see discussion in Maddison et al. 1984). The last method was that of de Queiroz (1987). This is essentially a modification of the outgroup Postilla 208 substitution method of Cantino (1982, Donoghue and Cantino 1984), and was developed for determining outgroup states when the outgroup relationships are unresolved. In this case it was based on the consensus cladogram of Crane (1988) which shows a polychotomy between Gnetales, Bennettitales, Pentoxylon, Caytonia and glossopterids. Statistical tests for association between characters (Sporne 1977) are based on a Chi squared 2 by 2 and 2 by C contingency tests (Snedecor and Cochran 1967). In these comparisons only one monocot group (Dioscoreales) was used. The matrix of presence and absence of the states from the two characters was filled at the family level. Thus a family might have combinations of states not found in any single extant taxon. In addition, if the family had more than one character state, the occurrence was recorded in more than one place in the matrix. Results Ovules and carpels each have unique suites of characters that are treated in separate parts below. Each part is subdivided into three other categories: characters and states; character polarization; and distribution of ancestral states in basal angiosperms. The characters and states are discussed first, and this section is supplemented by tables defining the characters and states, and figures showing many of the states. Next is a discussion of character state polarity based on the methods discussed above. In the final section, the distributions of the states in the classification of Cronquist (1981) and the phylogeny of Donoghue and Doyle (1989) are examined. Ovules Although the terminology dealing with ovules is extensive (see Summaries of Maheshwari 1950; Bocquet and Bersier 1960), it has been noted that considerable morphological variability is included within single descriptive 4 Angiosperm Ovules and Carpels: Their Characters and Polarities Postilla 208 Table 1 Definitions of ovule characters and states. Many of these states are illustrated in Figure 1. Ovule Morphology 1. Relative position of the chalaza to the micro- pyle: 0 opposite, at the opposite or nearly op- posite end from the micropyle end (e.g., Fig. 1A); 1 lateral, distal, laterally, midway to distally between the two ends of the ovule (e.g., Fig. 1G); 2 lateral, proximal, laterally, midway to beside the micropyle (e.g., Fig. 1D). 2. The angle between a line along the funiculus through the chalaza and a line from the micro- pyle through the embryo sac: 0 orthoangle, 0 to 5° (e.g., Fig. 1B); 1 semiangle, 6 to 89°; 2 hemiangle, 90 to 159° (e.g., Fig. 1A); 3 ana- angle, 160 to 200° (e.g., Fig. 1C); 4 circuman- gle, coiled greater than 200° (e.g., Fig. 1). 3. Complete symmetry of all portions of ovule with a section through funiculus and embryo sac: 0 symmetrical, all portions symmetrical (e.g., Fig. 1B); 1 asymmetrical, some portions asymmetri- cal (e.g., Fig. 1C). 4. Levels of symmetry of section through funiculus and embryo sac: 0 outer symmetry, 2 or more integuments, nucellus and embryo sac symmet- rical in size, shape and general orientation (e.g., Fig. 1B); 1 inner symmetry, only one integu- ment, nucellus and embryo sac symmetrical in size, shape and general orientation (e.g., Fig. 1C); 2 nucellar symmetry, nucellus and embryo sac symmetrical in shape (e.g., Fig. 1F); 3 sac symmetry, embryo sac symmetrical in shape; 4 lacking symmetry, not symmetrical (e.g., Fig. 1E). 5. Relative position of attachment of nucellus to micropyle end: 0 opposite, at the opposite or nearly opposite end from the micropyle (e.g., Fig. 1A); 1 lateral, distal, extending to side of embryo sac near distal end; 2 lateral, proximal, extending to side of embryo sac near proximal (micropyle) end (e.g., Fig. 1D). Ovule Integument Morphology 6. Continuity between the integument and funicu- lus: 0 apo, outermost integument equally free abaxially and adaxially, has no continuity with the funiculus (e.g., Fig. 1A); 1 semi, outermost integument unequally free with abaxial (outer) portion adjacent to the funiculus having 5 to 50% continuity with the funiculus; 2 hemi, outer- most integument unequally free with abaxial (outer) portion adjacent to the funiculus having partial continuity which is greater than 50% with the funiculus (e.g., Fig. 1C); 3 syn, outermost in- tegument unequally free with abaxial (outer) portion beside the funiculus having complete continuity with the funiculus, outer integument missing beside the funiculus (e.g., Fig. 1H). 7. Integument number: 0 bitegmic, nucellus sur- rounded by two integuments and the outer in- tegument may be free or have continuity with the funiculus (e.g., Fig. 1A); 1 unitegmic, nucel- lus surrounded by one integument and the in- tegument may be free or have continuity with the funiculus; 2 ategmic, nucellus not surround- ed by any integuments, naked; 3 multitegmic, nucellus surrounded by three or more integu- ments and the outermost integument may be free or have continuity with the funiculus. 8. Integuments forming micropyle: 0 inner, only by inner, which may project past outer (e.g., Fig. 1A); 1 both, both form similar-sized pore (e.g., Fig. 1C); 2 both zigzag, both form similar- sized pore but the hole through the outer is dis- placed from the hole formed by the inner integ- ument (e.g., Fig. 1D); 3 outer, only by outer, in- ner of shorter length. Nucellus Morphology 9. Division of archesporial cell and layers above embryo sac: 0 multiparietal, division of the ar- chesporal cell to produce a parietal cell that di- vides to form at least two layers; 1 uniparietal, division of the archesporial cell to produce a pa- rietal cell that may form up to one layer, 2 none, no parietal cell. 10. Division of the nucellar epidermis to form layers above the embryo sac: 0 multiepidermal, divi- sion of the epidermis to produce more than one layer, forming a nucellar cap; 1 uniepidermal, division of the epidermis to produce only one layer, forming a nucellar cap; 2 none, no divi- sion of the epidermis. terms (Bocquet 1959; Bocquet and Bersier 1960; Davis 1966; Bouman 1984). The following list of characters attempts to document the extensive variation in ovule morphology, particularly that masked by current terminology. In the future additional characters will undoubtedly be proposed, including a suite based on developmental traits (e.g., Bouman 1984). Characters and States The characters and their states are summarized in Table 1. The characters can be assigned to one of the following suites of characters: general ovule morphology (characters 1-5), integument morphology (characters 6-8) and nucellar morphology (characters 9, 10). 5 Angiosperm Ovules and Carpels: Their Characters and Polarities General ovule morphology includes five characters that show: the position of the attachment of the chalaza and nucellus relative to the micropyle; the angle of the ovule relative to the funiculus; and the symmetry of the ovule and parts (Table 1). Character 1 shows the position of the chalaza attachment relative to the micropyle. This can vary from opposite (Fig. 1A), lateral, distal (Fig. 1G) to lateral, proximal (Fig. 1D). Character 2 records the angle between a line from the funiculus through the chalaza and a line from the micropyle through the embryo sac. The states are divided by the natural breaks recorded in ovule angles, and include orthoangle (Fig. 1B), semiangle, hemiangle (Fig. 1A), ana-angle (Fig. 1C) or circumangle (Fig. 11). Character 3 indicates the symmetry of the entire ovule and the states are symmetrical (Fig. 1B) or asymmetrical (Fig. 1C). The next character, 4, records the general symmetry of the various layers of the ovule and the states are outer symmetry (Fig. 1B), inner symmetry (Fig. 1C), nucellar symmetry (Fig. 1F), sac symmetry or lacking symmetry (Fig. 1E). Finally, character 5 indicates the position of the nucellar connection to the ovule relative to the micropyle. The states are opposite (Fig. 1A); lateral, distal; or lateral, proximal (Fig. 1D). Thus an ovule considered in conventional terms to have campylotropous morphology (Fig. 1D, F) would have the chalaza attached proximal, have an ana-angle, and would have been asymmetrical overall with at least the embryo sac symmetrical. Yet the two could vary in the internal symmetry (Sac or nucellar) and the nucellar attachment site (proximal or opposite). There are three characters describing variation in integument morphology: continuity (a term of preference over fusion, Sattler 1978) between the integument and funiculus (character 6), integument number (character 7) and integumentary composition of the micropyle (character 8). The states of character 6 are given as prefixes to be attached to the term for integument number (character 7) and include: the states with the Postilla 208 outermost integument apo, totally free (Fig. 1A); semi, mostly free; hemi, mostly continuous (Fig. 1C); or syn, missing alongside the micropyle (Fig. 1H). The integument number includes the commonly used bitegmic (Fig. 1A), unitegmic, ategmic or multitegmic states. Lastly, the micropyle may be formed by the inner integument (Fig. 1A), both integuments (Fig. 1C), both integuments arranged in a zigzag (Fig. 1D) and the outer integuments. Thus an ovule considered to have anatropous morphology (Fig. 1C, H) with the existing terminology could have hemibitegmic morphology and a micropyle formed from both integuments (Fig. 1C) or synbitegmic morphology and a micropyle formed from the inner integument (Figure 1D). The last two characters deal with nucellar morphology which indicate whether the archesporial cell divides and the number of cell layers it produces (character 9), and division and extent of cell layers derived from the nucellar epidermis (character 10). These distinctions have been discussed by Davis (1966). The states for the archesporial cell are multiparietal, uniparietal and none; and those for the nucellar epidermis are multiepidermal, uniepidermal and none. Character Polarization The ovule character states were polarized with outgroup comparison (Watrous and Wheeler 1981). The distribution of character states was examined for the six outgroups (Table 2) determined by Crane (1985) and/or Doyle and Donoghue (1986, 1987). Most of these taxa have also previously been suggested to be the specific sister-group to angiosperms (see Stebbins 1974; Retallack and Dilcher 1981; Crane 1985; Doyle and Donoghue 1986). For defining the morphology of the outgroups, | have accepted the recent interpretations that: Bennettitales and Pentoxylon are bitegmic (Crane 1985), there is a lack of fusion between the outer integument and funiculus in corystosperms (G. J. Retallack, personal communications 1988); and the outer envelopes of Gnetales ovules 6 Angiosperm Ovules and Carpels: Postilla 208 Their Characters and Polarities Fig. 1 Diagrammatic representations of basic ovule types showing their evolutionary relationships based on the proposed ancestral character states and developmental transformation series. (See text for further discussion.) The ovule in B is considered ancestral. The specific character states (based on characters 1-8 and their codes in Table 1) for each ovule are A) 02000000; B) 00000000; C) 23110201; D) 23132002; E) 13142102; F) 23120201; G) 24140202; H) 23110300; /) 24110300. (D-G redrawn from Bocquet and Bersier 1960). and outer integuments of angiosperms are not The polarization is complicated because homologous (Crane 1985). In addition, a there is some variability in states between hypothetical bitegmic ovule, which would be these taxa, and the relationships of the considered homologous to angiosperms, was outgroups to angiosperms are equivocal. Two created for Caytonia (Doyle and Donoghue recent cladistic analyses (Crane 1985, Doyle 1986), glossopterids (Retallack and Dilcher and Donoghue 1986) did not produce the 1981) and Gnetales. These are coded under same topology (Crane 1988) and the topology the respective taxa as ‘‘hypothetical’”’ of trees with the same length or several steps (Table 2). longer suggest that any of these outgroup U Angiosperm Ovules and Carpels: Their Characters and Polarities Table 2 Postilla 208 Distribution of ovule states in fossil and extant outgroups. ? = no data. Abbreviations for each character: character 1, attachment of chazala: opp = opposite, prox = proximal; character 2, angle: ana = ana-angle, hemi = hemiangle, ortho = orthoangle, semi = semiangle; character 3, complete symmetry: asym = asymmetrical, sym = symmetrical; character 4, levels of symmetry: inner = inner symmetry, outer = outer symmetry; character 5, attachment of nucellus: opp = opposite; characters 6,7, continuity of integument and funiculus, and integument number: apobi = apobitegmic, apouni = apounitegmic, synbi = synbitegmic; character 8, micropyle: both = both, inner = inner; character 9, division of archesporial cell: multi or mult = multiparietal; character 10, division of nucellar epidermis: multi = multiepidermal, none = none. GENERAL OVULE MORPHOLOGY TAXON 1 2 3 4 Bennettitales opp ortho-semi sym outer Caytonia opp ortho sym inner —hypothetical prox ana asym inner corystosperms opp hemi sym outer glossopterids opp ortho sym inner —hypothetical opp ortho sym outer Gnetales opp ortho sym inner —hypothetical opp ortho sym outer Pentoxylon opp ortho-semi sym outer INTEGUMENT NUCELLAR MORPHOLOGY MORPHOLOGY 5 6,7 8 9 10 Rer. opp = apobi inner multi = ?multi 1,2 opp apouni — inner ? ? 3 opp — synbi both ? ? opp = apobi inner ? ? 4,5 Opp apouni — inner multi ?none 6 opp = apobi both multi ?none opp apouni inner multi = multi 7 opp apobi inner multi = multi opp = apobi inner ?multi ? 1 1. Crane 1985; 2. Crepet 1974; 3. Doyle and Donoghue 1986; 4. Thomas 1933; 5. Retallack, personal communication, 1988; 6. Gould and Delevoryas 1977; 7. Takaso and Bouman 1986. taxa could be the sister-group. Because of this variability | determined the ancestral state for the internode below the angiosperms using the four methods described in the Materials and Methods section above. In all four of these polarizations | also separately examined the empirical codings and hypothetical codings (Table 3). For most characters, these four polarizations produced the same results (Table 3). Characters 1-3, 5, 6 and 9 are all polarized the same, suggesting the ancestral ovule had the following character states: chalaza opposite the micropyle; an orthoangle (less than 5°; total symmetry (based on a section through the funiculus and embryo sac); the nucellus attached opposite the micropyle; the outermost integument equally free; and a multiparietal nucellar morphology (Fig. 1B). If one uses current terminology only orthotropous (Fig. 1B) ovules have such a morphology. Character 4 refers to the symmetry of the various layers of the ovule. Empirical coding is either outer or inner symmetry, and hypothetical coding denotes outer symmetry. This equivocal polarization is an artifact because unitegmic ovules cannot have outer symmetry. The plesiomorphic state of character 7 is either unitegmic or bitegmic (empirical coding) or bitegmic alone (hypothetical coding). Until recently (Les 1988), no one has seriously suggested that the unitegmic condition was ancestral (e.g., Bocquet and Bersier 1960; Cronquist 1968, 1988; Takhtajan 1969, 1980). This is based both on the distribution of the state in the angiosperms (ingroup commonality) and the developmental transformation series that shows that the unitegmic state arose independently in several lineages (Maheshwari 1950; Bouman 1984). Due to these different justifications, | consider the bitegmic state as ancestral. Therefore the ancestral state for character 4 is outer and for character 7 is bitegmic. In character 8, the condition where the inner integument forms the micropyle is ancestral in all the polarizations except the topology of Doyle and Donoghue (1986) and the topology of Crane (1988) with the hypothetical coding. Since only the inner integument can form the micropyle in unitegmic ovules and the inner integument 8 Angiosperm Ovules and Carpels: Their Characters and Polarities Table 3 The ancestral carpel states based on four methods and topologies (see Materials and Methods section) using the documented empirical (e) and hypothetical (h) states. (See text for further discussion.) A Based on the topology of Crane (1985). B Based on the topology of Doyle and Donoghue (1986). C Based on the taxa in Crane (1985). D Based on the topology of Crane (1988). ? = equivocal polarization between two states; | = insufficient data for method. CHARACTERS 1 23 455s 6 89 10 AL er 10.0 OF S22 On 0). 20m 10 0 h 0 0 0 0 0 0 0 0 0 0 is EY OF Oy Wee OF §@) -7 - O) Oo) 0 i OO OO 0 OF O # 0 CS gO OO 2° 0) oO 2 Oo 0 ny Oy 0 OO. Oi 0) 0) 10) 0 Dy 2 WM © 0 2? OF @ #0) | | ny O 0 @ © © @ @W 2 | | forms the micropyle in all known bitegmic outgroups (Bennettitales, corystosperms and Pentoxylon), this state will be accepted as ancestral. Lastly, there is little data on the division of the nucellar epidermis (character 10). Examination of bennettitalean ovules (Crepet 1974) and data from the literature on Gnetales (e.g., Takaso and Bouman 1986) suggest that there is a division to form at least one layer and thus is ancestral. Discussion of Polarization. Authors have previously used other data to polarize angiosperm states. Presently the literature is in conflict about which type, anatropous or orthotropous, is ancestral for angiosperms. In general, angiosperm phylogenists (e.g., Cronquist 1968, 1988; Takhtajan 1969, 1980; Crane 1985; Doyle and Donoghue 1986) have concluded that anatropous is ancestral, based in its Occurrence in putatively ancestral angiosperm taxa (Magnoliales) and its common distribution in all angiosperms (Davis 1966). This conflicts with the common view of ovule morphologists (e.g., see reviews in Takhtajan 1980 and Bor 1978) who believe that the orthotropous ovule is ancestral based on development. Postilla 208 A dissenting morphological view is that of Bor (1978) who points out that the angle of the anatropous ovules is not due to the bending followed by fusion of an orthotropous ovule. The angle of an anatropous ovule is initiated by the formation of the integuments on one side of the ovule primordium. Actually, there are three related processes: change in the placement of the chalaza from an opposite position to proximal position with the concurrent formation of a raphe; an increase of the continuation between the funiculus and the outermost integument; and an increase of the angle to approximately 180°. The question is whether or not there are similarities in the developmental programs of the two ovule types which show transformations from one type to the other. Similar developmental transformation series can be found in ovules of Chloranthaceae (Vijayaraghavan 1964), Molluginaceae (Narayana and Lodha 1963) and Potamogetonaceae (Takaso and Bouman 1984). They all have initiation of the ovule terminally on the ovule primordium yet the ovules in the latter two families become curved, at different stages of development. The ovules at the embryo sac stage of development in Potamogetonaceae have a semiangle and are apobitegmic, whereas the ovules in Molluginaceae have an ana-angle and are hemibitegmic. Thus the three ovules share the same early developmental transformation stages, but the latter two have additional stages to those found in Chloranthaceae. Based on these data, orthotropous type ovules, as found in Chloranthaceae, are ancestral if no neoteny or other shifts in developmental timing have occurred. It also appears that ovules with distal chalazas and semi- to hemicontinuation are more similar to orthotropous ovules than those with proximal attachment and syncontinuation. Further evidence is found in species of the families Barclayaceae, Lardizabalaceae and Platanaceae. In these families, the ovules in a single carpel have angles ranging from ortho to semi, and at least in Platanaceae the attachment is opposite to distal (personal 9 Angiosperm Ovules and Carpels: Their Characters and Polarities observation). Thus developmental data is compatible with the polarizations from outgroup Comparison. In conclusion, an ovule with opposite relative position of the chalaza, orthoangle, total and outer symmetry of ovule, an opposite relative position of the nucellar attachment, apobitegmic integuments, inner integument micropyle and multiparietal nucellus (Fig. 1B) is ancestral for angiosperms. These states are found in some bitegmic ovules of Bennettitales and Pentoxylon. Character States in Basal Angiosperms The distributions of ovule character states are shown by representative genera for each family (Appendix 1). Many of the major clades of basal angiosperms have families with the ancestral states mentioned above. In Magnoliidae, six families from four orders (Amborellaceae, Chloranthaceae, Saururaceae, Piperaceae, Barclayaceae and Lardizabalaceae) have the ancestral states. These states are also found in two families of basal Hamamelidae (Platanaceae, Hamamelidaceae). Ancestral states are found in the order Polygonales of Caryophyllidae (Polygonaceae) and nearly all of these states are found in Crassulaceae (Rosidae). Although none of the taxa in Dilleniales (Dilleniidae) have the ancestral states, some members of the closely related Flacourtiaceae (Hickey and Wolfe 1975) have them. Lastly, two families (Stemonaceae and Smilacaceae) from the basal monocots examined have taxa with the ancestral state. The distribution suggests that ancestral character states are found in most clades but are rarely common. If the distributions of these states are plotted on the cladogram of Donoghue and Doyle (1989) a similar result is obtained with the ancestral states found scattered in a number of clades. Both these distributions suggest that either the polarization is fundamentally incorrect or that the most common character states (basically anatropous morphology) are homoplasous and have evolved multiple times. Postilla 208 Carpels Unlike the relatively well-defined characters for ovules, only general morphological types of carpel have been defined (see Leinfellner 1950; Bailey and Swamy 1951; Weberling 1981; van Heel 1981). Some problems arise because even the definition of the carpel itself is in dispute (Sattler 1974; Sattler and Perlin 1982). In some configurations carpels have ovules attached to a gynoecial appendage (Sattler 1974), while in others they are attached in a cauline or axillary position (e.g., Moeliono 1970; Sattler and Lacroix 1988). In this study a carpel is considered as a gynoecial ‘appendage which encloses ovule(s)” (Sattler and Perlin 1982) which may have the ovules attached to the appendage, or in a cauline or axillary position to it. The list of characters documents some of the variation that exists in carpel morphology. Again, it is likely that developmental characters will be important for future descriptions (van Heel 1981, 1983, 1984). Characters and States The characters that describe carpels can be divided into two suites (Table 4; Fig. 2): general carpel morphology and anatomy; and ovule placement. The former includes a character for carpel morphology (character 11) and one for anatomy (character 12). The character for carpel morphology (character 11) uses terms that are in general usage, but | define them in a topological manner so that they can usually be identified without developmental data. The terms refer to the actual developmental closure of the carpel and not to a hypothetical evolutionary suture. It should be noted that developmental data are a preferable adjunct since redifferentiation of epidermal cells in the carpel has been well documented (e.g., Siegel and Verbeke 1989). There are six character states for carpel morphology. The first is ascidate, in which the ovules are attached proximal to the closure (Fig. 2A-C, I-L). The second is ascoplicate, in which some ovules are attached proximal and some adjacent to the closure (Fig. 2F, G). The 10 Angiosperm Ovules and Carpels: Their Characters and Polarities Postilla 208 Table 4 Definition of carpel characters and states. Most states illustrated in Figure 2. General Carpel Morphology and Anatomy 11. Carpel morphology: 0 ascidate, placenta of the ovule(s) attached proximal to the closure of the carpel either on the gynoecial appendage, axillary or terminally (e.g., Fig. 2A); 1 ascopli- cate, placentae of some of the ovules are prox- imal to the closure of the carpel and some are adjacent to the closure of the carpel (e.g., Fig. 2F); 2 plicate A, placentae of the ovule(s) at- tached adjacent to the closure of the carpel and on the gynoecial appendage, but the base below the attachment of the ovules is ascidate (Fig. 4G); 3 plicate B, placentae of the ovule(s) attached adjacent to the closure of the carpel and on the gynoecial appendage and closure reaches base of carpel (Fig. 2H); 4 ascidate, notched, as in ascidate, but in which the outer portion of the carpel has a notch in the admedi- al (ventral) surface that does not penetrate the locule at the level of the attachment of the ovules (e.g., Fig. 2E); 5 ascidate, split, as in as- cidate, but in which the proximal part of the pla- centa is split and the two parts thereby formed are contiguous with the base and margins of the closure. (Thus the placenta could be con- sidered proximal to the closure or marginal at the base of the closure; e.g., Fig. 2D.) 12. Vasculature to the ovule(s): 0 one trace, one (usually ventral) trace in the base of the carpel, from which the ovular trace(s) originate. (The one trace state includes the situation in which the trace branches and each branch extends to an ovule.); 1 two trace, two (usually ventral) traces in the base of the carpel from which the ovular traces originate; 2 multiple trace, more than two traces in the base of the carpel from which the ovular traces originate; 3 residual trace, the remains of the receptacular stele goes directly to a single ovule. Ovule Attachment 13. Position of attachment of the placentae in the carpel: 0 basal, at or near the base of the loc- ule or near the attachment of the carpel to the receptacle (e.g., Fig. 2C); 1 apical, at or near the apex of the locule (e.g., Fig. 2A); 2 lateral, marginally, along the lateral margins or sub- margins of the closure of the carpel (e.g., Fig. 2H); 3 lateral, admedially, not along the clo- sure, but towards the central axis of the flower or inflorescence (e.g., Fig. 2B); 4 lateral, exme- dially, not along the closure, but opposite clo- sure or away from the central axis of the flower or inflorescence (e.g., Fig. 21); 5 lateral, radial, not along the margin, but along the radial walls (e.g., Fig. 2J); 6 chaotic, arranged in an appar- ently random manner in the carpel (Fig. 2K). 14. Ovular orientation with respect to the plane of symmetry of the carpel. The plane of symmetry is a radial plane originating from the central axis of the flower or inflorescence that evenly bi- sects the carpel, usually along the closure of the carpel or the pollen-tube conducting tissue. If the carpel is terminal, the plane is the one that evenly bisects the carpel and dissects the lateral closure of the carpel: 0 planar, ovule at- tached in the plane (Fig. 2A); 1 nonplanar, ovule not attached in the plane (Fig. 2B); 2 planeless, this occurs when the carpel is termi- nal, has no lateral closure and is radially sym- metrical so no single plane is possible (Fig. 2L). 15. Number and arrangement of ovules: 0 single, ovule one (e.g., Fig. 2A); 1 double, alternate, ovules two, at different levels; 2 double, oppo- site, ovules two, paired at the same level (e.g., Fig. 2D); 3 few, alternate, ovule 3 to 6, at dif- ferent levels (e.g., Fig. 2H); 4 few, opposite, ovules 3 to 6, pairs of ovules at the same level (e.g., Fig. 2B); 5 few, clustered, ovules 3 to 6, ovules at the same level; 6 many, alternate, ovules more than 6, at different levels (e.g., Fig. 2K); 7 many, opposite, ovules more than 6, pairs of ovules at the same level; 8 many, clustered, ovules more than 6, ovules at the same level. next two are plicate or conduplicate: plicate A has the ovules attached adjacent to the closure, but the base of the carpel below the closure is ascidate or cup-shaped (Fig. 4G); plicate B has a form similar to plicate A, but the closure extends to the base of the carpel (Fig. 2H). The last two states are special cases of ascidate: ascidate, notched and ascidate, split. In the first the ovules are attached proximal to the closure yet having a notch on the admedial surface which does not penetrate the locule at the level of ovule attachment (Fig. 2E). In ascidate, split, the proximal part of the placenta is split and the two parts formed are contiguous with the base of the closure (Fig. 2D). The character for vasculature to the ovule(s) has four states and refers to the number of traces at the carpel base. The trace states may be one, two, multiple or residual. In this last state, the remains of a receptacular stele extend into a single ovule. 11 Angiosperm Ovules and Carpels: Postilla 208 Their Characters and Polarities © Pers) | ook jhe ae Ss) Reon bos | gL Tech) &) © CaS Eee cee KOK) © ee © Fig. 2 Diagrammatic representations of basic carpel types showing most of the morphological, ovule attachment, ovule orientation, and ovule number and arrangement character states. The shading shows the closure, the portion of the carpel that closes during ontogeny. The diagrammatic cross-sections of the carpels are orient- ed so that the right side of the longitudinal section is down. Specific character states (based on characters 11, 13-15 and their codes in Table 4) are A) 0100; B) 0314; C) 0000; D) 5312: E) 4312; F) 13(lower) 2(upper) O(lower) 2(upper)4; G) 13(lower) 2(upper)14; H) 3213; /) 0401; J) 0516; K) 0616; L) 0020. 12 Angiosperm Ovules and Carpels: Their Characters and Polarities The second suite of characters includes: the position of ovule attachment in the carpel (character 13); the orientation of the ovule in respect to the plane of carpel symmetry (character 14); and the number and arrangement of the ovules to each other (character 15). The character states for position of attachment including three previously described states (e.g., Cronquist 1988): basal (Fig. 2C), apical (Fig. 2A), and lateral, marginal, which is along the closure (Fig. 2H). Additional states include lateral, admedial which is not along the closure but towards the central axis of the flower (Fig. 2B); lateral, exmedial which is away from the central axis of the flower (Fig. 2); lateral, radial which is along the radial walls (Fig. 2J) and chaotic which is in an apparently random manner in the carpel (Fig. 2A). The states for the character of ovule orientation in respect to plane of symmetry are as follows: planar, in the plane (Fig. 2A); nonplanar, out of the plane (Fig. 2B); and planeless, for carpels with a terminal ovule and radial symmetry (Fig. 2L). The last character, number and arrangement of ovules, has nine states. These states are: single, double and alternate; double and opposite; few and alternate; few and opposite; few and clustered; many and alternate; many and opposite; and many and clustered. These states are based on numbers enclosing natural modes in the taxa studied. Based on this character suite, an ascidate carpel may have the ovules attached: apically and planar (Fig. 2A); lateral; admedially and nonplanar (Fig. 2B); radially and nonplanar (Fig. 2J) or basally and aplanar (Figure 2L). Plicate carpels are usually marginal and nonplanar (Fig. 2H). Ascoplicate carpels have a mixture of a number of states, with the lower ovule(s) admedial and planar (Fig. 2F) or admedial and nonplanar (Fig. 2G); and the upper ovule(s) marginal. Character Polarization Polarization to identify the ancestral carpel states is difficult due to the uncertain homologies between angiosperm carpels and reproductive organs Postilla 208 of related taxa. Thus it is not possible to use outgroup Comparison, as knowledge of such structural homologies is needed. As a result, several secondary and less reliable methods were used to polarize the carpel characters. The potential problems with character association (e.g., Stebbins 1974; Crisci and Stuessy 1980) and developmental transformation series (e.g., Kraus 1988) are realized and the results should not be considered as robust as those from the ovule analysis. In the sections below, | discuss the polarity of carpel characters based on: 1) association of characters and 2) developmental morphology and transformation series. In the first analysis, the frequency of six polarized ovule characters were examined to see if they were correlated with carpel characters. These characters include: position of funicular attachment; angle of ovule to funiculus; symmetry of the entire ovule; position of nucellar attachment; degree of continuation of the outermost integument; and division of the archesporial cell to form a parietal cell. The ovule characters were chosen because they were unequivocally polarized with outgroup comparison (see above). The mode of micropyle formation was not included in the comparison since it did not appear to be independent from other characters. First, all unitegmic ovules must necessarily have this state. In addition, there is a significant correlation between both apo and syncontinuation to inner integument (Table 5). The correlation to syncontinuation occurs since the micropyle cannot be formed by the outer integument, because a portion of the outer integument is missing along the funiculus. A few synbitegmic ovules are exceptions and have the micropyle formed by both integuments and the funiculus (e.g., Magnoliaceae). Two carpel characters were chosen: general carpel morphology and number of ovules in each carpel. These were chosen because they contrast the two major hypotheses of the carpel composition: that the carpel is composed of a sterile gynoecial appendage and a fertile meristematic region 13 Angiosperm Ovules and Carpels: Their Characters and Polarities usually in the axillary or cauline position (an ascidate carpel usually with one or two ovules; e.g., Moeliono 1970) or is a megasporophyll, basically a fertile leaf with a conduplicate or peltate structure (a plicate carpel usually with few or many ovules). The conduplicate carpel is considered to be similar to a folded vegetative leaf, without an adaxial meristem, whereas the peltate carpel is similar to a peltate leaf having an adaxial meristem (or cross-zone). To examine these hypotheses, the frequencies of the ascoplicate, plicate A and plicate B carpel character states were combined and contrasted with the combined ascidate states. In addition, the frequencies of the families with ‘‘few” and ‘“‘many’”’ ovules states were combined to contrast with the combined ‘‘single”’ and ‘‘double’’ states. The initial comparison between the ancestral ovule states and the carpel states (Table 6) shows that in most comparisons a higher percentage of ancestral ovule states are found in ascidate carpels and carpels with low numbers of ovules than are found in ascoplicate or plicate carpels and carpels with few or many ovules. Only two states, opposite attachment of the nucellus and existence of a parietal cell, were found more often in the carpels with more than two ovules though only by a few percent. These observations suggest that ascidate carpels with one or two ovules may be ancestral. The levels of significance of these correlations were examined (Table 6). Ten of the correlations between ascidate and one or two ovules were positively, though not significantly, correlated with ovule state. Only two (opposite attachment of nucellus and existence of parietal cell) were negatively correlated (Table 6), though these had the smallest Chi squared values. This suite of generally positive correlations to the ancestral ovule states again suggests that the ancestral states may have been ascidate and had one or two ovules. In addition, several highly significant ingroup correlations were found. Ascidate morphology was significantly correlated to carpels with one or two ovules, whereas ascoplicate and plicate morphology were Postilla 208 Table 5 Statistical analysis of the correlation within ovules for the families in Appendix 1 excluding Alismatales. Comparison between continuation and micropyle type in bitegmic ovules. APOBITEGMIC HEMIBITEGMIC SYNBITEGMIC Inner — 10 (62%) 16 (47%) 25 (78%) Both 6 (38%) 18 (53%) 7 (23%) 0.01 > p > 0.005 significantly correlated to those with greater than two ovules (Table 7A). These data provide evidence that the ascidate carpels with one or two ovules are real entities, and the data in the Appendices show that ascidate carpels with high numbers of ovules are rare. The frequency of combinations of basic carpel states within families was also examined (Table 7B). Each family was examined to see what type of carpel morphology was found and whether there was only one type, referred to as alone or were several types (in combination). The highly significant results show that the families with ascidate carpels are most likely to have only ascidate morphology. In contrast, families with ascoplicate carpels are almost always found in combination with either ascidate or plicate morphology. Lastly, plicate carpels are found alone in only half the families with this carpel type, whereas in other families they are associated with other morphologies. These results are again consistent with the notion that the ascidate morphology is ancestral. Another potentially useful method of polarizing characters is the use of developmental transformation series based on developmental morphology (Nelson 1978; Crisci and Stuessy 1980; Kraus 1988), though there are obvious limitations to this method. The fundamental questions are: do carpels with different morphologies share developmental stages and are developmental transformation series found between types? Fortunately, the growing number of studies of carpel development provides a preliminary understanding of the developmental stages of 14 Angiosperm Ovules and Carpels: Their Characters and Polarities Table 6 Postilla 208 Statistical analysis of the correlation between polarized ancestral ovule characters and select carpel characters (morphology and ovule number) for the families in Appendix 1 and 2 excluding Alismatales. A Position of attachment of funiculus. B Angle of ovule to funiculus. C Symmetry of the entire ovule. D Position of attachment of nucellus. E Continuation of outermost integument. F Formation and division of parietal cell. A opposite other ascidate 10 (71%) 49 (58%) nonascidate 4 (29%) 36 (42%) 0.50 > p > 0.25 B orthoangle other ascidate 10 (67%) 49 (58%) nonascidate 5 (33%) 35 (42%) O75 = =70'50 g symmetrical other ascidate 10 (63%) 49 (59%) nonascidate 6 (37%) 34 (41%) 0.75 > p > 0.50 D opposite other ascidate 54 (61%) 5 (50%) nonascidate 35 (39%) 5 (50%) 0.75 > p > 0.50 E apo other ascidate 14 (70%) 45 (57%) nonascidate 6 (30%) 34 (43%) 050'>ip =20/25 F parietal no parietal ascidate 48 (58%) 4 (40%) nonascidate 35 (42%) 6 (60%) 0.50 > p> 0.25 the various carpel types (e.g., van Heel 1981, 1983, 1984: Endress 1972; Erbar 1983; Tucker 1959, 1981; Tucker and Gifford 1966). Some carpels have a U-shaped primordium with a meristematic area that produces the ovules either within or between the tips of the primordium (Fig. 3Aa). The mature carpels that Table 7 1 or 2 ovules >2 ovules 1 or 2 ovules >2 ovules 1 or 2 ovules >2 ovules 1 or 2 ovules >2 ovules 1 or 2 ovules >2 ovules 1 or 2 ovules >2 ovules opposite other 9 (56%) 48 (44%) 7 (44%) 62 (56%) 0.50 > p > 0.25 orthoangle other 9 (53%) 48 (44%) 8 (47%) 61 (56%) 0.50 > p > 0.25 symmetrical other 9 (56%) 48 (44%) 7 (44%) 62 (56%) 0.50 > p > 0.25 opposite other 50 (45%) 7 (47%) 61 (55%) 8 (53%) p > 0.90 apo other 14 (61%) 43 (42%) 9 (39%) 60 (58%) 0:25: =" p:=0N0 parietal no parietal 49 (43%) 5 (42%) 64 (57%) 7 (58%) p > 0.90 develop have an ascidate morphology (e.g., Fig. 3Ab) with basal to apical attachment. (The latter placentation occurs when the carpels are laterally attached to an elongate receptacle.) The second, and apparently most common type of primordium, has the shape of a Statistical analysis of the correlation within carpels for the families in Appendix 1 and 2 excluding Alismatales. A A selected pair of carpel characters (morphology and number of ovules). B Between carpel morphology and whether they are found alone or in combination within families. A 1 or 2 ovules >2 ovules B Alone In combination ascidate 49 (77%) 15 (23%) ascidate 40 (74%) 14 (26%) ascoplicate 4 (27%) 11 (73%) p < 0.005 ascoplicate 1 (8%) 12 (92%) p < 0.005 plicate 6 (21%) 23 (79%) plicate 14 (54%) 12 (46%) 15 Angiosperm Ovules and Carpels: Their Characters and Polarities circular or elliptical cup (Fig. 3B, Il, Ill) and variable placement of the placental regions. Some have the meristematic area that produces the ovules confined to the admedial end that is toward the floral apex (Fig. 3Bla). Others have the meristematic area both at the admedial end and along portions of the rim of the cup (Fig. 3Blla). Finally, the remainder apparently have the meristematic area only along the rim (Fig. 3Bllla). The mature carpels developing from the first type, with the meristematic area only at the end (Fig. 3Bla), are similar to those seen in Figure 3Blb. Primordia with the fertile areas on the admedial end and the rim (Fig. 3Blla) produce carpels of two types. One type has a distal closure and a period of intercalary growth producing a mature carpel with an ascidate morphology similar to Figure 2U and K; in addition, some of the ovules would have an admedial placement. The other type has an ascoplicate morphology (Fig. 3Bllb; Fig. 2G). Finally, primordia with the fertile area restricted to the rim (Fig. 3Bllla) mostly produce carpels with a plicate A type morphology (e.g., Fig. 3Blllb). If there is a distal closure and intercalary growth, they may have an ascidate morphology similar to Figure 2l, J, K, but the ovules will not be admedially placed. The final type has a U-shaped primordium with the fertile region along the margins of the closure (Fig. 3Ca). The mature carpel would have a plicate B type morphology (Fig. 3Cb). It is actually unclear whether this type exists as many plicate carpels have a cross-zone (e.g., van Heel 1981; Erbar 1983; Tucker and Gifford 1966). Such a cross-zone may be reduced to a few cells (Tucker and Gifford 1966) and may not be recognizable. Plicate B carpels are apparently found in Canellaceae, Saururaceae, Ranunculaceae, Lardizabalaceae, Paeoniaceae, and Trilliaceae. Carpel Transformations. |n all these types, the greatest development of the carpel primordium occurs on the portion distal to the center of the apical meristem (except, perhaps, Cercidiphyllum, but see van Heel 1987). In later development the separate Postilla 208 ea | Oe (0) ©) (©) BIIl [25 © poo ©) DEVELOPMENT Fig. 3 Comparison of the types of carpel primordia (a) and representative mature morphologies (b). In the former (a) the shading shows the region of the primordium from which the ovules develop, whereas in the latter (b), it shows the closure, the portion of the carpel that closes during ontogeny. The diagrammatic cross-sections of the carpels are oriented so that the right side of the longitudinal section is down. Based on developmental transformation series, Type A has evolved into Type B and ultimately to Type C. (See text for further discussion.) 16 Angiosperm Ovules and Carpels: Their Characters and Polarities Postilla 208 meristem or the cup fully develops. The few studies that have compared development of these types suggest that they also all share a separate meristematic origin of the gynoecial appendage and the ovules (e.g., Hagerup 1934, 1936, 1938). Thus Type A is the simplest as seen in Illicium (Robertson and Tucker 1979). Type BI is an elaboration of Type A with the connation of the gynoecial and ovule meristems (see Leins 1972 and Sattler 1974, 1978 for the developmental basis of connation and adnation between floral parts) which results in a gynoecium such as that in Chloranthaceae (as figured by Endress 1987). Type BIl has the same developmental steps as BI with the additional stage of extension of the ovule meristem along the closure as can be seen in a number of genera in Ranunculaceae (van Heel 1981). Type BIll is developmentally identical to Bll (e.g., Swamy 1949), except for the inability to produce ovules medially, though the cross-zone still exists. Finally, Type C appears to be the total integration of the two meristems with the inhibition of the medial portion of the cross- zone, though it is impossible to tell from the current data whether its cell division completely ceases or whether a few cells are produced (e.g., Tucker 1981). The similarity of the Types A and B and the proposed developmental transformation series would support the hypothesis that ascidate carpels are ancestral. If this is correct, one might expect to find phylogenetic examples of this developmental transformation series. Ideally, the should exhibit a basic transformation (Fig. 3) from the Type A developmental pattern to the Type C pattern (Hagerup 1934, 1936, 1938, though | disagree with the comparisons to gymnosperms). Unfortunately, independent analyses of the relationships of the taxa are tentative and only potential phylogenetic evidence for portions of this transition series exists, based on the variability of certain families. For example, van Heel’s (1981) work on Ranunculaceae shows examples of the BI type, which produces a mature ascidate morphology, as well as carpels of the BIl type, which produces an ascoplicate morphology. In addition, some of the taxa have carpels of the BIll type with a mature plicate A morphology. Study of Drimys winteri shows that the mature carpel is ascoplicate with two rows of ovules (Tucker 1959). The primordium is the Bll type, but the origin of the submarginal region that makes the fertile portion of the rim may be from the subapical initial (which forms the gynoecial appendage) and the adaxial initials or both. The primordium can be considered transitional between type BIl and type BIll because the ovules originate from the submarginal region and the portion of the submarginal region produced on the admedial side of the carpel may originate from the adaxial initials. Also of interest is the observation that the plicate A carpel of Drimys lanceolata has poorly defined meristematic regions along the rim and in a very restricted portion of the admedial side (Tucker and Gifford 1966). This could be considered to be an example of how the transition between Type BIll and Type C would occur. Finally, the fertile region of the Type C primordium has been suggested to have originated from a distal (adaxial) meristem (Hagerup 1938). Acer has a syncarpous, bicarpellate ovary with parietal placentation. The carpels have plicate B morphology and Type C development. Hagerup (1938) has suggested, based on detailed ontogenetic work, that the placentae originate from a portion of the floral meristem adaxial to, or more terminal than, the gynoecial appendage. If this were indeed so, and even Hagerup considers it equivocal, it would indicate that carpels with Type C development have two separate meristems, one for the gynoecial appendage and one for the placentae. These meristems would be directly homologous to those found in other developmental types. Discussion of Polarization. The preliminary evidence suggests that ascidate carpels with one or two ovules are ancestral. This contrasts to the current hypotheses that indicate that plicate carpels with many, marginally-placed ovules are basic. In either case, the adaxial (admedially placed) ee ee eee ee eee eee — — eee 17 Angiosperm Ovules and Carpels: Their Characters and Polarities placental meristem of a carpel with a sterile gynoecial appendage is structurally homologous with the adaxial (admedially placed) cross-zone of a plicate carpel, and with the adaxial (admedially placed) meristem that produces the marginal placentae of a conduplicate carpel. This understanding of the developmental morphology does away with the concept of “congenital fusion’ and replaces it with the concept that the formation of many ascidate and conduplicate carpels is due to the integration of the two meristematic areas. This concept of integration of meristems (see Sattler 1974, 1978) has been already documented for other floral organs (e.g., Leins 1972, Sattler 1974, 1978). Although there appear to be examples of the developmental transformation between carpel primordia types, the direction of the series can be disputed. Is Type A or Type C ancestral? The distribution of the types in the ancestral angiosperms is informative. Type C primordium has only been found in a few taxa thus far; Type A and Type BI are found throughout the basal angiosperms. In addition, ascidate carpels are more commonly found alone in families as opposed to the ascoplicate and plicate morphologies (Table 7). These distributions are at least suggestive that A or BI is ancestral. Both Type A and Type BI produce mature carpels with ascidate morphology and usually low numbers of ovules. Based on these observations, a carpel having ascidate morphology and one or two ovules can be considered ancestral and all ascoplicate and plicate carpels derived. This hypothesized polarity is suggestive but should be corroborated with further work. Character States in Basal Angiosperms The distributional data on the carpel states are compiled in Appendix 2. The ancestral states of ascidate morphology and single or double ovules are common in Magnoliidae, Caryophyllidae, Rosidae and Dioscoreales. These states are also found somewhat less commonly in Hamamelidae, Dilleniales and Alismatales. These data show that the proposed Postilla 208 ancestral carpel states are commonly found in every basal clade. Thus the common carpel character states (ascidate morphology, lateral attachment and nonplanar orientation, and one or two ovules in an alternate arrangement which is similar to Fig. 2D) also included the polarized ancestral states. When the distribution of states is examined on the tree of Donoghue and Doyle (1989), it also appears that the trees are more parsimonious when the ancestral states are considered to be ascidate morphology and one or two ovules. General Discussion Outgroup comparison, with which data from developmental transformation series is compatible, suggests that the ovule of ancestral angiosperm had certain characteristics. This hypothesized ovule had: the funiculus and nucellus attached opposite the micropyle; an orthoangle; symmetry throughout all layers of the ovule; no fusion between the funiculus and the adjacent integument; a micropyle formed from the inner integument only; division of the archesporial cell; and probably two integuments. Based on conventional terms, such an ovule would be considered orthotropous, bitegmic and crassinucellar. A structural hypothesis of ovule evolution is possible based on the polarization of characters, the developmental transformation series, and distribution of ovule types in families (Fig. 1). This hypothesis suggests that the ancestral ovule (Fig. 1B) gave rise to ovules that only differed in angle (Semi- or hemiangles) and attachment of the chalaza (opposite or distal; Fig. 1A). From this type evolved two later types of ovules: those with proximal funicular attachment, an ana-angle, inner symmetry and hemicontinuation (anatropous, the most common type in the basal angiosperms; Fig. 1C) as well as those with proximal funicular attachment, an ana- angle, sac symmetry and hemicontinuation (orthocampylotropous morphology; Fig. 1D). From the former (Fig. 1C) arose the ovules 18 Angiosperm Ovules and Carpels: Their Characters and Polarities with syncontinuation (anatropous; Fig. 1H), and nucellar symmetry (anacampylotropous; Fig. 1F). These ovules with synbitegmic integuments and anatropous morphology (Fig. 1H) produce the ovules with circinangle (circinanatropous; Fig. 11). Orthocampylo- tropous (Fig. 1D) and anacampylotropous (Fig. 1F) give rise to their respective amphitropous morphologies (Fig. 1E, G). Several ovule characters are not shown in Figure 1. The polarized states of characters 7, 9, and 10 suggest that all ovules with abitegmic integuments are derived from ovules with the bitegmic state, and all ovules that lack a parietal cell, with or without division of the nucellar epidermis (pseudocrassi- nucellar and tenuinucellar nucelli), are derived from ovules with the crassinucellar state. The above is a structural hypothesis, in which the steps may have occurred in parallel in different clades. The distributions of the ancestral and derived states (Appendix 1) strongly suggest that anatropous (and other suggested derived types) ovules are homoplasous. It was found in the survey of basal angiosperms (Appendix 1) that ovules considered anatropous are quite variable. Thus ‘‘anatropous” ovules vary in their angle, symmetry, and continuation of the funiculus and outermost integument. This variability would support the hypothesis that they originated multiple times. One hypothesis for the multiple origins of ‘‘anatropous”’ ovules is that the micropyle can be directed away from the pollen tube transmission tissue. Work on ovule development could further test this hypothesis. The hypothesis suggests that the curvature and late initiation of the outermost integument along the funiculus are derived. Comparative study of the genetics and expression of the developmental programs, perhaps by sampling the proteins expressed during ovule development, and developmental morphology of ovules with no curvature (e.g., orthotropous) with those which have a higher angle (€.g., anatropous) from the same family could test the hypothesis. In addition, phylogenetic analysis of clades with variable states could show whether the Postilla 208 structural hypothesis is corroborated in specific cases by the phylogenetic relationships. The polarizations of the carpel states are less robust than those of the ovules. Yet the polarizations consistently suggest that the ascidate morphology and one or two ovules per carpel are the ancestral states. The developmental data suggest a transformation series between the ascidate to plicate state. Based on these data the following evolutionary structural relationships between general carpel types are suggested (Fig. 4). The ancestral carpel with its ascidate morphology may have had the ovule basally (Fig. 4A), laterally (Fig. 4B) or apically (Fig. 4C) attached, and the orientation either planar with one ovule or nonplanar with two, opposite ovules. According to this model, carpels having ascoplicate morphology with both lateral, marginal attachment, and lateral admedial arrangement (of a single ovule in planar orientation; Fig. 4D) originated from carpels with the ascidate morphology and basal or apical ovule attachment (Fig. 4A, C). The position of the marginal ovules may have been either alternate or opposite. If the medial ovule and cross-zone in this carpel type was lost, the resulting carpel has a plicate B morphology (Fig. 4F). Two forms may have evolved from the ascidate carpel with lateral ovules (Fig. 4B), one having ascidate morphology and few or many ovules (Fig. 46), and the other having plicate morphology, but an ascidate base (Fig. 4G). Again, attachment of the ovules may have been opposite or alternate. The plicate types (Fig. 4F, G) may have also originated from each other. All other carpel types (e.g., those in Fig. 2) are considered derived from one of the above types. The distributions of the ancestral and derived states (Appendix 2) suggest that ascoplicate and plicate ovules may have evolved multiple times. The placement of the ovules and the morphology of the carpel base is quite variable in carpels having the ascoplicate morphology. Similar variation is found in carpels with plicate morphology, especially the marginal to submarginal 19 Angiosperm Ovules and Carpels: Postilla 208 Their Characters and Polarities Y Fig. 4 Evolutionary relationships of the basic carpel types based on the proposed ancestral character states and developmental transformation series. (See text for further discussion.) The shading shows the closure, the portion of the carpel that closes during ontogeny. The diagrammatic cross-sections of the carpels are oriented so that the right side of the longitudinal section is down. The types found in A, B, and C are equally ancestral, whereas the remaining are considered derived. placement of ovules along the margins of the The correlation of these two carpel types closure (e.g., Bailey and Swamy 1951; Puri with high ovule numbers may suggest an 1952, 1961; Canright 1960; Eames 1961). This evolutionary explanation for their multiple hypothesis suggests that this variability is due origins and frequency. In this survey, carpels to multiple origins of ascoplicate and plicate with ascidate morphology rarely have high carpels. ovule numbers. The study of additional 20 Angiosperm Ovules and Carpels: Their Characters and Polarities characters, such as exact placement of the ovules along the closure, also may clearly show that the derived carpels are not structurally similar and may have evolved due to selection for high seed number. Further tests of the hypothesis are possible with experimental morphogenetic analysis. Based on the polarized characters and the developmental morphology of the primordia, a transition series of carpel types have been suggested (Fig. 3A-C). The most divergent types are carpels with ascidate morphology and basal ovule attachment (Type A development) and with plicate B morphology and lateral, marginal ovule attachment (Type C development). Carpels possessing ascidate and basal states have placentae that originate from a separate meristem, distal to the meristem forming the gynoecial appendage. This morphology must be homologous with a carpel having plicate and lateral, marginal states in which the placenta may originate from the marginal area of the gynoecial appendage. Hagerup (1938) has already suggested that the placental regions in both morphologies originate from a distal (abaxial) meristem. This suggestion would indicate that the gynoecial appendage of a carpel showing plicate morphology is a composite structure produced by the integration of two meristems. Such a structural homology between the two carpel types is testable by following cell lineages in the gynoecial appendage and the region producing the placentae to see whether they have separate origins. Tracking of cell lineages is possible with clonal analysis as described by Poethig (1987). If the cell lineages for the placentae are found to have a separate origin from the gynoecial appendage (particularly in carpels with plicate morphology), this hypothesis would be supported. Importantly, results supporting this hypothesis would cast serious doubts on the concept that the angiosperm carpel is a megasporophyll with laminar ovules. The new concept would suggest that the ovule carpel complex is best interpreted as a short shoot with the gynoecial appendage equivalent to a bract or bracteole, and the ovule being the Postilla 208 apical portion of an axillary bud or terminal apex. Thus an angiosperm carpel and ovule system would potentially be homologous with the bract-bracteole-terminal ovule system found in the outgroups such as corystosperms and gnetophytes. Further work on the origin and development of the gynoecial and placental tissue will be necessary before complete understanding of the origin and evolution of the angiosperm carpel can be achieved. Another test would be a phylogenetic analysis of the genera in families that show a number of primordial and mature carpel morphologies (e.g., Ranunculaceae). If the phylogeny showed congruence with the suggested structural transformation, the hypothesis would be corroborated. The results of the above analysis and hypotheses of ovule and carpel structural evolution have interesting implications for basal angiosperm evolution. Currently, there is growing equivocal evidence (Walker and Walker 1984; Friis et al. 1986; Hamby and Zimmer 1990) that families with diminutive flowers may be closest to the ancestral angiosperm groundplan (Taylor 1988; Hickey and Taylor 1989; Taylor and Hickey 1990). The ovule and carpel data support these suggestions by indicating that only the piperalean families and Amborella have the ancestral states of both organs. In addition, there continues to be debate as to which group is ancestral in the monocots: Alismatales (Cronquist 1981) or Dioscoreales (Dahlgren et al. 1985). These data suggest that only members of Dioscoreales have the complete collection of ancestral ovule and carpel states. Acknowledgments | sincerely thank Leo J. Hickey for encourage- ment and lively discussions on this project, for his aid in drawing the figures and for the support from his research funds from the Peabody Museum of Natural History, Yale University. | also thank Toby and Leon Gold for financial support which made this project 21 Angiosperm Ovules and Carpels: Postilla 208 Their Characters and Polarities successful. Lastly, | thank Drs. Peter R. Crane, | comments that have greatly improved this William L. Crepet and Peter F. Stevens for work. Literature Cited Bailey, |. W. and B. G. Swamy. 1951. The conduplicate carpel of dicotyledons and its initial trends of specialization. Am. J. Bot. 38:373-79. Bocquet, G. 1959. The campylotropous ovule. Phytomorphology 9:222-27. Bocquet, G. and J. D. Bersier. 1960. La valeur systématique de l'ovule: développements tératologiques. Arch. Sci. 13:475-96. Bor, J. 1978. A note on anatropy versus orthotropy. Phytomorphology 28:219-24. Bouman, F. 1984. The ovule, p. 123-57. In A. M. Johri (ed.), Embryology of Angiosperms. Berlin, Springer- Verlag. Bryant, H. N. 1989. An evaluation of cladistic and character analysis as hypothetico-deductive procedures, and the consequences for character weighting. Syst. Zool. 38:214-27. Canright, J. E. 1960. The comparative morphology and relationships of the Magnoliaceae. Ill Carpels. Am. J. Bot. 47:145-55. Cantino, P. D. 1982. Affinities of the Lamiales: a cladistic analysis. Syst. Bot. 7:237-48. Crane, P.R. 1985. Phylogenetic analysis of seed plants and the origin of angiosperms. Ann. Missouri Bot. Gard. 72:716-93. 1988. Major clades and relationships in the ‘“‘higher’’ gymnosperms, p. 218-72. In C. B. Beck (ed.), Origin and Evolution of Gymnosperms. Columbia Univ. Press, New York. Crepet, W.L. 1974. Investigations of North American cycadeoids: the reproductive biology of Cycadeoidea. Palaeontographica Palaophytol. Abt. B 148:144-59. Crisci, J. V. and T. R. Stuessy. 1980. Determining primitive character states for phylogenetic reconstructions. Syst. Bot. 5:112-35. Cronquist, A. 1968. The Evolution and Classification of Flowering Plants. New York, Houghton Mifflin. 1981. An Integrated System of Classification of Flowering Plants. New York, Columbia Univ. Press. 1988. The Evolution and Classification of Flowering Plants. 2nd ed. New York, Columbia Univ. Press. Dahigren, R. M. T. 1980. A revised system of classification of the angiosperms. J. Linn. Soc., Bot. 80:91- 124. Dahigren, R. M. T., H. T. Clifford and P. F. Yeo. 1985. The Families of the Monocotyledons. Berlin, Springer-Verlag. Dahigren, R. M. T., S. Rosendal-Jensen and B. J. Nielsen. 1981. A revised classification of the angiosperms with comments on correlation between chemical and other characters, p. 149-204. In D. A. Young and D. S. Seigler (eds.), Phytochemistry and Angiosperm Phylogeny. New York, Praeger Publishers. Davis, G.L. 1966. Systematic Embryology of the Angiosperms. New York, John Wiley and Sons. Donoghue, M. J. and P. D. Cantino. 1984. The logic and limitations of the outgroups substitution approach to cladistic analysis. Syst. Bot. 9:192-202. Donoghue, M. J. and J. A. Doyle. 1989. Phylogenetic analysis of angiosperms and the relationships of Hamamelidae. In P. R. Crane and S. Blackmore (eds.), Evolution, Systematics, and Fossil History of the Hamamelidae, Volume 1: Introduction and ‘Lower’ Hamamelidae. Systematics Association Special Volume 40A:17-45. Oxford, Clarendon Press. Doyle, J. A. and M. J. Donoghue. 1986. Seed plant phylogeny and the origin of angiosperms: an experimental cladistic approach. Bot. Rev. (Lancaster) 52:321-431. 1987. The origin of angiosperms: a cladistic approach, p. 17-49. In E. M. Friis, W. G. Chaloner and P. R. Crane (eds.), The Origins of Angiosperms and Their Biological Consequences. Cambridge, Cambridge Univ. Press. Eames, A. J. 1961. Morphology of the Angiosperms. New York, McGraw-Hill. Endress, P. K. 1972. Zur vergleichenden entwichlungsmorphologie, embryologie und systematik bei Laurales. Bot. Jahrb. Syst. 92:331-428. 22 Angiosperm Ovules and Carpels: Postilla 208 Their Characters and Polarities 1987. The Chloranthaceae: reproductive structures and phylogenetic position. Bot. Jahrb. Syst. 109: 153-226. Erbar, C. 1983. Zum karpellbau einiger Magnoliiden. Bot. Jahrb. Syst. 104:3-31. Friis, E. M., P. R. Crane and K. R. Pedersen. 1986. Floral evidence for Cretaceous chloranthoid angiosperms. Nature 320:163-64. Gould, R. E. and T. Delevoryas. 1977. The biology of Glossopteris: evidence from petrified seed-bearing and pollen-bearing organs. Alcheringa 1:387-99. Hagerup, O. 1934. Zur abstammung einiger angiospermen durch Gnetales und coniferae. K. Dan. Vidensk. Selsk. Biol. Medd. 11(4):3-83. 1936. Zur abstammung einiger angiospermen durch Gnetales und Coniferae. Il. Centrospermae. K. Dan. Vidensk. Selsk. Biol. Medd. 13(6):3-59. 1938. On the origin of some angiosperms through the Gnetales and the Coniferae. Ill. The gynaecium (sic) of Salix cinerea. K. Dan. Vidensk. Selsk. Biol. Medd. 14(4):3-34. Hamby, R. K. and E. A. Zimmer. 1990. Ribosomal RNA as a phylogenetic tool in plant systematics. In D. E. Soltis, P. Soltis and J. J. Doyle (eds.), Plant Molecular Systematics. Routedge, Chapman and Hall. Heel, W. A., van. 1981. A S.E.M.-investigation on the development of free carpels. Blumea 27:499-522. 1983. The ascidiform early development of free carpels, a S.E.M.-investigation. Blumea 28:231-70. 1984. Variation in the development of ascidiform carpels, a S.E.M.-investigation. Blumea 29:443-52. 1987. Note on the morphology of the male inflorescences in Cercidiphyllum (Cercidiphyllaceae). Blumea 32:303-09. Hickey, L. J. and D. W. Taylor. 1989. Reexamination of leaf architectural characters of basal angiosperms and their sister-groups: implications for the origin and relationships of angiosperms. Am. J. Bot. Suppl. 76(6):245. Hickey, L. J. and J. A. Wolfe. 1975. The bases of angiosperm phylogeny: vegetative morphology. Ann. Missouri Bot. Gard. 62:538-89. Kraus, F. 1988. An empirical evaluation of the use of the ontogeny polarization criteria in phylogenetic inference. Syst. Zool. 37:106-41. Leinfellner, W. 1950. Der bauplan des synkarpen gynozeums. Oesterr. Bot. Z. 97:403-36. Leins, P. 1972. Das Karpell im ober- und unterstandigen gynoeceum. Ber. Dtsch. Bot. Ges. 85:291-94. Les, D. H. 1988. The origin and affinities of the Ceratophyllaceae. Taxon 37:326-45. Maddison, W. P., M. J. Donoghue and D. R. Maddison. 1984. Outgroup analysis and parsimony. Syst. Zool. 33:83-103. Maheshwari, P. 1950. An Introduction to the Embryology of Angiosperms. New York, McGraw-Hill Publishing Co. Moeliono, B. M. 1970. Cauline or Carpellary Placentation Among Dicotyledons. Vols. |, ll. Assen, Netherlands, Van Gorcum and Co. Narayana, H. S. and B. C. Lodha. 1963. Embryology of Orygia decumbens Forsk. (Aizoaceae). Phytomorphology 13:54-59. Neff, N. A. 1986. A rational basis for a priori character weighting. Syst. Zool. 35:110-23. Nelson, G. 1978. Ontogeny, phylogeny, paleontology, and the biogenetic law. Syst. Zool. 27:324-45. Philipson, W. R. 1974. Ovular morphology and the major classification of the dicotyledons. J. Linn. Soc., Bot. 68:89-108. 1977. Ovular morphology and the classification of dicotyledons. Pl. Syst. Evol., Suppl. 1:123-40. Poethig, R. S. 1987. Clonal analysis of cell lineage patterns in plant development. Am. J. Bot. 74:581-94. Puri, V. 1952. Placentation in angiosperms. Bot. Rev. (Lancaster) 18:603-51. 1961. The classical concept of angiosperm carpel: a reassessment. J. Indian Bot. Soc. 40:511-24. Queiroz, K., de. 1987. Phylogenetic systematics of iguanine lizards. Univ. Calif. Publ. Zool. 118. Retallack, G. J. and D. L. Dilcher. 1981. Arguments for a glossopterid ancestry of angiosperms. Paleobiology 7:54-67. Robertson, R. D. and S. C. Tucker. 1979. Floral ontogeny of Illicium floridanum, with emphasis on stamen and carpel development. Am. J. Bot. 66:605-17. Sattler, R. 1974. A new approach to gynoecial morphology. Phytomorphology, 24:22-34. 1978. ‘Fusion’ and ‘continuity’ in floral morphology. Notes R. Bot. Gard. Edinburgh 36:397-405. Sattler, R. and L. Perlin. 1982. Floral development of Bougainvillea spectabilis Willd., Boerhaavia diffusa L. and Mirabilis jalapa L. (Nyctaginaceae). J. Linn. Soc. Bot. 84:161-82. 23 Angiosperm Ovules and Carpels: Postilla 208 Their Characters and Polarities Sattler, R. and C. Lacroix. 1988. Development and evolution of basal cauline placentation: Basella rubra. Am. J. Bot. 75:918-27. Siegel, B. A. and J. A. Verbeke. 1989. Diffusible factors essential for epidermal cell redifferentiation in Catharanthus roseus. Science 244:580-82. Snedecor, G. W. and W. G. Cochran. 1967. Statistical Methods. Ames, lowa State Univ. Press. Sporne, K.R. 1977. Some problems associated with character correlations. Pl. Syst. Evol., Suppl. 1: 33-51. Stebbins, G. L. 1974. Flowering Plants. Cambridge, Belknap Press. Swamy, B. G. L. 1949. Further contributions to the morphology of the Degeneriaceae. J. Arnold Arbor. 30: 10-38. Takaso, T. and F. Bouman. 1984. Ovule ontogeny and seed development in Potamogeton natans L. (Potamogetonaceae), with a note on the campylotropous ovule. Acta Bot. Neerl. 33:519-33. 1986. Ovule and seed ontogeny in Gnetum gnemon Linnaeus. Bot. Mag. (Tokyo) 99:241-66. Takhtajan, A. L. 1969. Flowering Plants: Origin and Dispersal. Edinburgh, Oliver and Boyd. [Transl. by C. Jeffrey] 1980. Outline of the classification of flowering plants (Magnoliophyta). Bot. Rev. (Lancaster) 46: 225-359. Taylor, D. W. 1988. Reexamination of ovule characters in angiosperms and their outgroups: implications for origins of the angiosperms. Am. J. Bot. Suppl. 75:211. Taylor, D. W. and L. J. Hickey. 1990. An Aptian plant with attached leaves and flowers: implications for angiosperm origin. Science 228:1087-89. Thomas, H. H. 1933. On some Pteridospermous plants from the Mesozoic rocks of South Africa. R. Soc. Lond. Philos. Trans. 222:193-265. Tobe, H. and P. H. Raven. 1986. A comparative study of the embryology of Ludwigia (Onagraceae): characteristics, variation, and relationships. Ann. Missouri Bot. Gard. 73:768-87. 1987. Systematic embryology of the Anisophylleaceae. Ann. Missouri Bot. Gard. 74:1-26. Tucker, S.C. 1959. Ontogeny of the inflorescence and the flower in Drimys winteri var. chilensis. Univ. Calif. Publ. Bot. 30:257-336. 1981. Inflorescence and floral development in Houttuynia cordata (Saururaceae). Am. J. Bot. 68: 1017-32. Tucker, B. U. and E. M. Gifford, Jr. 1966. Carpel development in Drimys lanceolata. Am. J. Bot. 53: 671-78. Vijayaraghavan, M.R. 1964. Morphology and embryology of a vesselless dicotyledon—Sarcanara irvingbaileyi Swamy, and systematic position of the Chloranthaceae. Phytomorphology 14:429-41. Walker, J. W. and A. G. Walker. 1984. Ultrastructure of Lower Cretaceous angiosperm pollen and the origin and early evolution of flowering plants. Ann. Missouri Bot. Gard. 71:464-521. Watrous, L. E. and Q. D. Wheeler. 1981. The outgroup comparison method of character analysis. Syst. Zool. 30:1-11. Weberling, F. 1981. Morphologie der Bluten und der Blutenstande. Stuttgart, Eugen Ulmer. The Author David Winship Taylor. Department of Biology and the Peabody Museum of Natural History, Yale University, New Haven, CT 06511. Current address: Indiana University Southeast, Department of Biology, Division of Natural Sciences, 4210 Grant Line Road, New Albany, IN 47150. Postilla 208 Angiosperm Ovules and Carpels: Their Characters and Polarities 24 a6ed }xeu uO penuljuoD 6 iun Nu Jauul iquAs ddo Jauul wAse eue xoud suo0joe7 sselo —_— Bajiq snodoujeue aeaoepi0}Ie7) 8 auou niNuw Z ujoq iquAs ddo J@uul wAse eue xold eyjoubeyy (n}qo) Z auou Nu Z yjoq iquAs ddo J@uUul wAse eue xold UOJPUBPOUIT Ww} sseso Bayiq snodoujeue eeaoeloubew Jauul 9 jun niNuu y}oq iquAs ddo J@uUl wiAse eue xold eAayleqousny lL SsBld J@uU! Bag snodoujeue aesadeAajieqoujsny S _— nN Jauul iquAs ddo Jouul wAse eue xoud eiewodnyg sselo — Bajiq snodoijeue eeaoenewodny (inyqo) v @uou INU Jauul iquAs ddo Jauul wAse eue xoud Bunge (L) snodouyeue aeaoespuejuewiY € auou Nu Jouul iqudAs ddo J@uUul wAse eue xoJd euauebaqg | sseso J@uul Belg snodoujeue aeaoeiauebag zZ un njnuu J@uul iquAs ddo Jouul wAse eue xoud Bia]UIMOpNasd lL iun NW J9UUI iquAs ddo J@UUI wAse eue xoud sAWUG (9uou) (qjNwW) I sseso J@uul Baliq snodoujeue DCSICIIIUIN 43Y Ol 6 8 L'9 G v € c IL SAN3E) ASO TOHdYOW ADO TOHdHOW ADOTOHdHOW AINAO WHAN3S) YVTISONN LNAWNDALN| ‘jewuepideiun = jun ‘auou = auou ‘jewapidaninw = yjnw ‘sieXe| Auew = W :siwuepide se\}9ONU JO UOISIAID ‘QE JO}OeIEYO ‘eyouedIUN = JUN ‘Ve\]aONUINUa, = Nua} ‘Je||eONUISseJOOpNesd = JOd JO sesod ‘auoU = auOU ‘jeJeUedHINWU = WWNW ‘sedA} pexiW = PaxiW ‘Je||eONUIsse1O = SSBIO :|]90 jeUOdSEUOJe JO UOISIAIP ‘6 4aJOBIBYO ‘ouU! = JauU/ ‘Bez6iIz ylog = Z Yjog ‘yjog = Yjog :a\AdouollW ‘g sayoeseYd ‘o1WHeyuN = Bayun ‘o1w6eu, = Baz] ‘onuBeyiqués = iquds ‘oiw6Geyuniwas = juniwas ‘o1uBayigiues = iqiwes ‘soyeIN\GO = sNIGo ‘o\wWHeyUuNIWeY = /UNIWeY ‘o1WBeyIqueY = quay ‘o1WBa}Iq = Bayiq ‘owBeyunode = junode ‘o1wbeligode = igode :sequinu juawn6e4u! pue ‘sninojuny pue juewunBeju! jo Ayinujuoo ‘z'g9 sayBseYO ‘jewixoJd = xosd ‘a}Isooddo = Ado ‘je\sIp = JSID :snjjeonu Jo JUsWYORNe ‘¢ sayoeseyo ‘AyawwAs oes = oes ‘AjawwAs 19}NO = 4ajnNo ‘AjeWWAS Je|}eonU = /aonuU ‘AjaWWAS Buryoe) = YOR) ‘AewWAS JauU! = JQUU! :ALJQLULUAS JO S|QA2Q} ‘fp sayoeIeyo ‘/eou -yoluwAS = WAS ‘jeoujewuuAse = WAse :AljewWwAs aja|dwod ‘g sayoeseYo ‘ajGueIWwes = wes ‘a|GueouO = oyyo ‘a|bueiway = way ‘ajbueWNo.Io = ou ‘a|Bue -eue = eue ‘ajGue ‘% seyOeseYO ‘jewixoid = xoud ‘ayisoddo = ddo ‘eysip = JsIp :ejezZeUO Jo JUaWYOEIe ‘| JaJOeIeYD «JaJOeIEYO YOR JO} uolnelaiqqy ‘ayes uleaoUN Ing ajqeqoid = é PUR -UOXe} AU} JO} a}e}s o1e1 = () ‘eyep yo yOR| = — ‘paullwexe esaUaB ay) Aq pajuasaides jou saje}s 4! Alle} auJUe OU) UO (QQ6I) SINeq JO ABAUNS 0} 19}91 OL pue 6 Si9}OeBJeYO JO} SasayjUaed UI! Seye}s Ja}OeIeYD ‘peonpold ae yey} S|j9O ;eIOdseuUoJe JO JequUNU Ay} 0} BuIWeJes oul) AjlWe} UO Q| Ja}0eJeYO JapUN equuNU UNM (9961) SIAeq pue (1 g6}) JsINbUOID Woy SalIWey JO} eye “salle; WUedsoiBuUe ese Jo JaquUNU jOaI9S B JO} BBe}S eS-OAIGWA au] Je Sa}e}]s 18}0e1eYO aINAO L xipueddy aed }xau uo penuljuog 02 auou nyu Z yjoq iquAs ddo J@uul wAse ould xoud eyjAsseg 4yyoq WwW ‘| sselo Jouul Ba1q snodoujeue eeeoerine — — Bayiq snodouyeue aeadeuadsoip| gL nin — ujoq Iqiwway ddo Jouul wAse éeue xoud snyjueuOWwiYD 6k NW @uoU ujoq Iqiway ddo JOuU! wAse geue xold snyjueosjeg uw seiod Jauul Baliq snodoujeue aesoeujuedAen — — — snodo.j}eue eeadeBayowoy ore jun Nu Jauul iquAs ddo Jauul wAse Way xoud eioydhiog eeapiouuedsoisuly lz a@uou INU Jouul iquAs ddo Jouul wAse eue xold BIUOLIOH @eaplOluOLOH QL ‘ZL nNW niNuw Jouul luNnIWay ddo jaonu wAse Way xoud Blunedig co eeepioiunsedis & Ol — — J@uU! iquAs ddo Jouul wAse eue xoud BOI we Ol — = J@uUul iquAs ddo Jouul wAse eue xold e/reok{peyH = SPSPIOIWIUO/\ 3 Bayun oO. Ww ‘| sseso J@uUl Baiiq snodoujeue SeSIPIWIUOW Gl auoU, 1NUWdg, Jauul iquAs ddo Jeuul wAse eue xoud BIUBWH Ww sseio Z yjoq Bajiq snodoujeue eesoeluaWw | tl — — Jouul iqode ddo 49jno WAS OULO ddo B/|AJOQUYY a sseso — Bajiq snodouyeue aeaoe|jai0oquiy On € @uoUu Nu Z ujoq Iqiuuas ddo jaonu wAse IW9U }SID eIBINque/| c £ sselo Z yjoq Belg snodoujeueiway aesoe|jaueg Oa €l @uoUu nin Jouul Iqiwey ddo Jouul uwAse eue xoJd BONsUAW no} | sselo — Baxiq snodouyeue += aesoeonsuAW f= (or Bo Zk — — J@uu! iquAs ddo Jouul wAse eue xoud eOueuey) 8 § LL @uo0U INU _— — — oes wAse — — BUILUIS\/ Do Ol —_ — JOUuUul iquAs ddo JQUuU! wAse eue xoud PBUOUUY 6 9 (694111) Ss © " sseJo Jouul Belg snodoujojAdweo ‘snodowjeue eeadeuoUUYy 3. 5 AY Ol 6 g 1'9 g y € Z | SAIN3ID 4 3 ADSOTOHdHOW ADO TIOHdHOW ADOTOHdHOW SINAO WHANS3S) cco YVTISONN LNAWNDSLN| te ee ee penuljuoD L xipueddy 25 SS ——— 8 ee ee eee eee eee abed )xau uo panuluod (1un) yjoq Iqiuay (je0nu) ley 62 @uoU Nw Jouul iquAs ddo Jeuul wAse eue xoud wmesy (1un) 8z auou nN Jouul iqiuay ddo Jauul wAse eue xoid BIYDO/O}SUY/ pexiw (Ww) | sselo Jouul Be1ig (Snodosjeueulosio) snodowyeue aeadeIyIO/O]SUY Uh Kole | 92 auoU Nu JQuUU! igode ddo 4J9]No WAS OULIO ddo J0Qlg 92 @uoU INU JQuU! junode ddo Jouul WAS OULO ddo PIWwOadaYy (1un) 9z @uoU nyNuWwW ujoq iqode ddo 491NO WAS ouyO ddo PUAYIOH yjoq Beyun Sselo Jouul Bag snodoujoyyo aeaoeladid G2 @uou INU y}Oq igqode ddo J94no wAs OUO ddo snununes © (1un) ro) ! SU tc a@uou @uOU y}oq iqode ddo 4J9jno wAS OULO ddo BiuAnyjnoH fe €z @uoU NW vitefe| iqode ddo 494NO wAS OUyO ddo sisdoweuy = nua} | sselo ujoq Beg snodoujiwey ‘snodosjouyo aeaoeinines 22 auou nynu u}Oq igode ddo 491NO WAS ouyO ddo BpUueaIeS 22 e@uou nynuu Jouul iqode ddo 49{No WAS OuyO ddo WwnwsofspeH 1un fe Zc auou NW J@uUul iqode ddo 4J91no WAS OUyO ddo snyjUuesojyD # lung a. rd) Zc @uoUu njnu u}Oq iqode ddo 494nNO wAS Oyo ddo eBuuleosy S = yjoq O ® = © | sseso Jauul Baiiq snodosjouyo aeadeyUeIOIyUD co AN}Qog w 2 iz — — Z ujoq iqiwey ddo Jouul wAse eue xojd sndies01A5) oO sselo — Belg snodoujeue aeadeipueUusaH > 5 gl nin nynu Z yjog iquAs ddo JOUU! wiAse Ould xoud smune7q 2 > gl nin nynuu Z ujog iquAs ddo JOUUI wAse Ould xold WINWOWeUUID © a £ BY OL 6 8 L'9 g b € Zz | SAINI) a a ae ol 2 = ADOTOHdDHOW ADOTOHdHOW ADOTOHdHOW JINAO WWHANSS) (= SZ YVTISONN INAWNDALN| te penuluoy L xipueddy 26 Postilla 208 Angiosperm Ovules and Carpels: Their Characters and Polarities 27 L8 ‘Ze 9€ ce 0€ EE} S| a a eee penuljuoD L xipueddy SS a se @uoU @uoU J@uul junode " nua} — Beyiun (1uné) un Jouu! iquAs NW @uou J9@UU! JUNILUAY — = J@UU! iquAs = = Z yyog IQiWay JUNILUaY IUNILWAS — — Jeuul junode (nw) (1un) seiod Beyun (Ww) | sselo — Be1q auoU nynu Juul junode L sseld J@uUU! Beayun INU nyNuWW Jouul IquAs a@uoU un Jouul iquAs fi sseso Jouul Baxiq @uoUu NW JOUULs, iqode | sselo — Bajiq — nu Jouul iquAs 9u0u Nyjnu } Kole} Iqiwey ()Ntwu) (nw) sselo Jeuul Bajiq NW nynud J@uul 1quAs | sseso Jauul BayIq (1un) 9u0u yuu yjoq IQiuay Ww SSBlo ujoq Belg 9uou Jud yjoq Iqiwey Ww sseso — Bayiq Ol 6 8 19 ADOTOHdHOW ADOTOHdHOW YWTISONN LNAWNDSLN| abed jxau uo penuljuog ddo ddo ddo ddo ddo ddo ddo ddo ddo ddo ddo ddo ddo Juul Juul jgonu J@uUuI JQuUu! jaonu Juul WAS wAse wAse wAse wAse wAse WAS wAs wAse wAse wAs wAse wAse wAse wAse snodoujeue 0} snodoojAdweo wAse € OuULO ddo snodojoyuyo fe] 0) xoid IW3Y }SIp eue xoud eue xoud O40 eue xold snodouj}eue snodouyeueiway OUVO ddo snodojouyo eue xold eue xoud snodoujeue ILS OuLIO ddo snodojouyo eue xoud eue xoid snodojeue eue xoud snodo}eue eue xoud eue xold snodoujeue c L ADO TOHdYHOW JINAO TWHAN3S) JOJSCICDND aeaoeRsa}SeaeIiD WNAOIEY | sninounuey winiulydjaq eibajinby QUOWOUY aeaoerjnounuey wini|Adoje1a9 aeaoel|Aydojyesan equwoqey eBluaselg aeaoequoqgey eAejoleg aeaoeAejoieg seyann eaeyduisy aeaoevaeydwAn OquIn|aN aeaoeuoquNnjaN BIPUBSIYIS aeaoespuesiyos WINIIII] aeadeidili| SMN35) Postilla 208 Angiosperm Ovules and Carpels: Their Characters and Polarities 28 jay 9U0U NNW yjoq IQiWay (Ww) | sselo _ Belg = = Z yjoq Iqiwey Nw NW Z yyog IqiWiay (L) w sseBso = Baiiq un nN J9@uu! junode (/Nw) 1un nynud Juul juNIWas | sselo = Beyiun nN Nw J@uUl iquAs | sselo — Baliq nN njnw J@uul JUNILU@S — — J@uU! igode NW NW JOUUl Iqiway Beayiun l sseso JOUU! Bajiq @uou nynw J@uUl iqiway _ — JOUuUlé iqode (Nw) (Nw) I sselo J@uUul Bajiq = = J@uul Iqiwas = = Z yjog IQIWay = = Z yjog IQiuay = = Z yyog IQ|wWsy = = Z yyoq Iqiwey (q/NwW) (9u0u) = = Z yyog IQ!WSy — — Z yo IquAs (11/NUu) (Nu) (4od) WwW sselo ujoq Baliq Ol 6 8 L'9 ADOTIOHdYHOW ADOTOHdDHOW YVTISONN LNAWNDSALNI abed }xau UO panuljuo) 1SIP ddo ddo ddo ddo ddo \sIp ddo 4SIp ddo ddo ddo ddo ddo ddo ddo ddo ddo yoe| Juul Juul jeonu jeonu Juul oes Juul oes Juul J@\nNo J@UU! Juul Juul Juul! J@UU! Juul J@UU! wAse IWaY }SID snodo.jo|Aduueo ‘snodojeue wAse eue xoud wAse eue xoud snodouyiydwe ‘snodosjo|Adweo wAse IW9S 1SIp wAse IWS }SIp snodoJjyeueiway wAse ould xold snodoujeue wAse IWW8U 1SID wAse IWS 1SIP wAse IW9Y 1SIp snodoujiydwe 0} snodojiway wAse IW9Y xoid IWS wAS OULO ddo snodoujeue snodosjojAdweo ‘snodosjouyo wAse Way 1SID snodouyeue }souje 0} snodoujeueiwey wAse Way xoud wAse Way xoud wAse eue xoud wAse Way xold wAse eue xoid wAse IW9y xoud snodouyeue ‘snodojjeueiway € Cc L ADOTIOHdHOW FINAO WWHANAS) eueWN aesoevewn4 eueunbues BIZ]JOYOSYISF aeaoesaaeded BIges BUISOI/A/\V aeaoeriqes BUeLIOD aeadewelog eiodsoul | Wwinuwwedsiueyy snino909 aeadeWwedsiua eausie0egq eIGaIy aeaorjeqezipiey exopojuabies aeaoexopojuabies BlueZUueY winjjAydopog BUIDUBN winipawidz suaqueg sXyoy aeadepueqieg SMNad) penuiuoD L xipuaddy 19 99 9 v9 9 €9 z9 fo) 19 (jo) N & 09 = 6S 8 8S a. 1S wv 9S o 8 3 = GS Os 2 0 vs CH OG eG > wo 6 o Eg Eo oO DO ze 29 43y Ol ASO TOHdHOW YVTISONN Bayiq Iqiueay Iqiwuay IqIwas Bayiq Iqiuas Iqiway Iqiluas Iqiuas Bayiq iquAs Bayiq IquAs¢ Iqiway Iqiweay iqode Bayiq igode Bayiq Iqiway Bayiq Iqiwey Bayiq Iqiwey Bayiq IQILUSS Bayig 19 ADOIOHAHOW LNAWNDALN| a6ed }xeu uo panuljuoD snodo.jo|Aduued 0} snodoujeueiwey oes wAse eue \sIp oes wAse eue 1S! oes wAse eue 1SID (snodosjeue jsouyje 01) snodoujo|Adweo oes wAse eue }SIp oes wAse eue xold oes wAse eue xoid oes wAse IWOY }SIp snodo.jo|Aduueo JQUU! wAse eue xoud snodo.jeue JOUULé, wAse IWS, xoud oes wAse eue xoud J@UUl wiAse eue xoud J91NO wAs OuWO ddo snodoujoyyo ‘snodoujeue J@UUI wiAse IWW9S 1SIp J9\No wAs Ou ddo snodoujeueiwey 1UuBIs 0} snodosouyO J@UU! wAse IwWay xoud snodouj}eue J9uUU! wAse eue xoud snodoujeue J@UU! wAse eue xold snodou}eue J@uU! wAse Way xold snodoujeue v € c I ADOTOHdHOW SINAO WWHANAE) aeade||Aydofueg ePIBLIC oBnyjoyy WNaWIT aeaoeulbnyi OW BooR|Oo}AUd BLUAI/IH BIYOSIE) BOaYyjOWApIG aeaoeoorjo}Aud SNUWWeYjOIAW aeaoeuWeUOJAW Jequepinbi7 SILLY wNIAISIQ eBIBul|\/ aeooepljawewey SNuejeld aeaoeue}eid eajajdnz aeaoeajajdnyg wnAYalpiosag aeaoeAydipiniag UOJPUBPOYIOI aeadesJpuspoyool | UOUBIEMNA | aeaoesUsNeIja | SNNad) penunuoD L xipueddy Postilla 208 Angiosperm Ovules and Carpels: 30 Their Characters and Polarities == —_ Bayiq snodoujeue aeaoeyjuelejedijeig = — Belg snodouyeue aeaoeiuospiAeg 8/ @uou NW Z yjoq IQIWUaS ddo Jouul wAse IWay \sIp PUAWOZIYIS o¥ @uou INU J@uUU! IQILUaS ddo J@uul wAse IWW9U 1SIp winjejadoje1a9 ro¥ auoU Nu Z yjOq iqiwey ddo jeonu wiAse eue xoud winjejadoueydy (jw) (1NwW) Ww sseJo Z yjoq Baiiq snodoujeue ‘snodoujojAdweo ‘snodoijeueiwey eeaoeiuounyD — — Bayiq snodoujeue eeaoriydAiong BL a@uou HJM ujoq igqode ddo jeonu wAse IWS 1SIP sisdoaunoy BL e@uou INU u}Oq Iqiwaes ddo Jouul wAse IWS 1SID smieuuog uw sseso yjOq Beg snodoujeue ‘snodoujeueiway ‘snodosjouvo eeaoeseUUOD = = Belg snodoujeue aeaoeljjounlg LL INU nu Z ujoq iqiwey ddo Jauul wiAse IW9Y xoud BIUOBEY | sseJo — Beiq snodou}eue aeadeRiuosed pl @uoU NW Z yjog iquAs 1SIp oes wAse eue 1SIp BILUIOM QL ddo eue 1SIDZ Gl un yn Z yjog IqiwayYy ddo oes wAse way xoud BILIO@QQIH Gl J@UU! ddo lwau pl @uou yn Z yjog iqiuwau 1S! oes wAse eue xoud B1UayIG el nN nynu Z yjog IQIWUaUg, 1SIp oes wAse eue xoud BWAOIOV (uu) | sseso — Bajiq snodoujeueiwey ‘snodosjiydwe ‘snodojeue ‘snodoiye eeaoeiualliq Zl @uoU INU Jouul iquAs ddo Jouul wiAse ould xold eabo) LZ @u0U nyjnuu J@uU! iquAs ddo J@uU! wAse O49 xoud obequinid (WwW) | sselo _ 6a11q snodouyeue aeaoeulbequinid OL _— — Jouul iqode ddo 49]nNo WAS OuyO ddo wineyYy OL — — Jauul iqode ddo 491NO wAS eue ddo BIYOIUUMg (@uoU) (1un) (1un) (q)NW) Ww sselo Jouul Bayiq (snodosjeue) snodoujoyyo aeaoeuo0BAjogd 69 =< — J@UU! Iqiuas 1sIp oes wAse ould 1SIp Bullaqen 89 = — JQUUI Iqiwas }SID oes wAse IWaU }SIp winiydga |, g9 — — J@uul iqiwes xoud auou wAse IWS xoid snyjuesaj9S 43Y Ol 6 8 L'9 G v € c L SANA) ADOTIOHdHOW ADO TOHdHOW ADOTOHdHOW SINAO WHANAS) YVTISONN LNAWNDALNI penuljuod L xipueddy abed }xau UO panuljuoD Postilla 208 Angiosperm Ovules and Carpels: 31 Their Characters and Polarities +8 jun auou J@uU! iquAs ddo J@uul wAse eue xojd snwojng | seiod Jouul Ba11q snodouyeue aeaoewojng 8 'LL njNwW INU — IQIwWaUg, — youd wAse eue xoud BUBLINS Ww sselo _— 6ayiq snodou}eue aeaoeuUeLINS 8) — = Z Uyjog Iqiuay ddo oesé wAse eue xold PUIOSOSSOID sselo — Bayiq snodoujiyudwe ‘snodosjojAdweo aea0e}PWOSOSSOID sseso = Bayiq snodouyeue aeaoepeinaN JUNILUAY iquAs jaonu Way 28 nyNu nyNW J@uul iqiway ddo Jouul wAse eue xod eaelds 28 nN njNuu J@uul UNIS ddo jaonu wAse IWS }SID euebels 28 niNw Nu Jouul iqiwey ddo J@UUI wAse IWW9U xod snbaeje1p 28 QuoUg, Nu Jouul iqode ddo Jouul uwAse IWS }SIp snjepbhwy Z8 nN NW J@uul Iqiuay ddo J@uUl wAse IWaU xoud JOIYQUEJOWYY yjoq Beyun uw sselo J@uul Bayiq snodoujo|Adwed ‘snodoujeue ‘snodoseueiwey aeaoesoy 1g un @uou ujOq iqiway ddo Jouul wAse eue xold Basel] 08 un Nu Z ujoq iquAs ddo Jouul wAse eue xoud ebeYIXeS 08 auoU e@uoUu Z ujoq iquAs ddo Jouul wAse eue xoud PISSBUIEd nude} Beyiun (WW) | sselo = Baiq snodouyeue aeaoebeyixes sselo — Baiiq snodoujeue eeaoejojeydag JOUU! JUNILUaS jaonu O40 6L é Iung, u}Oq Iqiway ddo J@uUul wAse eue xoid winpas 61 e@uou un u}Oq iqode ddo 494no wAS IWS ddo BINsseID (4auul) L sseso ujoq Baiiq snodou}eue aeaoeinsselg Lp nynuu nynww Z ujoq iqudAs ddo Jouul wAse eue xold BIAQID) | sseso — Bajiq snodouyeue aeaoelAaly 43Y OL 6 8 L'9 S v € c L SNN3 ADOTOHHOW ADOTOHdHOW ADOTOHAHOW SINAO TWH3NSD HVTIZONN LNAWNODALN| penuljuoD | xipuaddy aed }xau uo panuljuoD Postilla 208 Angiosperm Ovules and Carpels: Their Characters and Polarities 32 a6ed jxau uo panuljuoD 1S = = _ iqode ddo J94no wAs OUYO ddo XB|ILUS sselo —_ Be1iq snodo.jo|Adweo ‘snodoujeueiwey ‘snodosjouyo aeaoeoRIIWS (9uou) (inw) 06 yn jun J@uul Iqiwwey ddo Jouul wAse eue xoud WINNIE lun (an}qo) 06 NW @uOU Jouul iquAs ddo Jouul wAse IW9Yy xoud sndojoosg 06 nJNW iun Juul! iqiwey ddo J@uul wAse IlwaU xoid SUC 06 a@uou @uoU Jeuul Iqiuiay ddo J@uul wAse eue xoud PjOapayy (uu) | sselo — Ba}1q snodoujeue aeooel||idL 26 @uoU jun Juul \quAs ddo Jeuul wAse eue xoid 98 — = Jouul iqode _— — WAS OuyO ddo BUOWA}S 98 — — JOUU! iqiweay ddo JOUULs wAse IWS xoud BIWOOID l sselo J@uUl Bayiq snodoujeue ‘snodosjoyyo aesdeuOW3}S 68 @uoU un Jouul iquAs ddo J9@uul wAse eue xoud eooe | 88 @uoU un Jouul iquAs ddo J@uUl wAse O4Id xoud esdeooziyos | sselo Jouul Bajiq snodouyeue ‘snodoujo\Adweo eeeorooe | 18 @uoU NW ujoq Iqiwiay ddo Jouul wAse eue xold PaIODSOIG | sselo ujoq Baiiq snodoujeue aveoeral00soig (inyqo) 98 @uoU @uoU Z ujoq iquAs ddo J@uul wAse eue xoud sndoyou | l nua} = Baiiq snodouyeue aeadepodoyou (1nyqo) 6E jun 9@u0u Jeuul iquAs ddo J@uul wAse eue xoud BUISI/\Y | seiod Jouul Beg snodoujiyudwe ‘snodoujeue aevooerjewsily ((anyqo) fore) jun @uou J@uul iquAs ddo Jouul wAse eue xold SUPCYIOUWIT (Wu) 4 seiod —_ Bayiq snodouyeue ‘snodosjo|Adweo aesoe Wey IoUWIT 43Y Ol 6 8 L9 G v € Cc L SNN3) ADOTOHdHOW ADOTOHHOW ADOTOHAHOW SINAO TWH3N35) YHYTISONN INAWNDALN| penuljuod L xipueddy Postilla 208 Angiosperm Ovules and Carpels: 33 Their Characters and Polarities ‘EGP:71. ABojoudiowoyAUd “7961 AWEMS 'Z6 ‘€1:8 108 Tf ‘PION ‘B86 UBUD ‘16 ‘1:7 (SN) UNIEN ‘Je ‘CISCO ‘Pex ASUEPIA JON JAS ‘2961 519g ‘06 ‘LHh'7L BIUld “LE6L MO}eed ‘68 ‘JSZ ‘6BL:LZEL “JON JOG LZ6L UOSsUeYeH gg ‘LZ1:¢ ABojoydiowo}AUd “EG6L COBY ‘2g ‘suOpa|A}ooOUOW 94} JO salle OY] “GEGL ‘Ie 19 UaH|YeQ 98 ‘612:ZE |OMUd MEN ‘BEG! Yor “Gg ‘9:2 ABojoyduowoyAYd ‘ZG6} 1edoYy “PB ‘4:18 ISAS “QUYeL JOG “LOS JAIIMZIND “EB ‘1:91 (8) 10g JEN 10S ‘UU ZOGL BINOYded ‘ZB ‘EEE JOG T ‘WY PGI J98H “18 :90E'rS ZED OG ‘ZiGL 80ed ‘08 ‘ZE:2 (8) JOG JEN 19S ‘UY ‘9681 ‘WEQNH,P “6Z ‘(z)GE wASSUY ‘AIUN SPUN] ‘GEE UOZWNeYW 'B/ “yZL:0G 108 T ‘Wy ‘E9EL PYeqqnH pue dwed ‘// ‘SJeMol JO SIS -aueBoueBiO ‘€/61 JANES ‘OZ 'S6P'LLL JON 108 ‘BSEL MISES “SZ ‘ZZL:79 IOS Pedy eMO} ‘OO1q LGB] OBY ‘vd ‘10€:g ABojoydsowo}AYd ‘Gg6} AWemseliad pue AWEMS '€/ ‘O9E:EL G 19S ‘peOW UPIPU] Old “LPG UBYY PUR INYeW\ ‘ZZ ‘6P9:S6 ‘ZED 10g ‘pESL IdNeH ‘LZ ‘6ze~e 10a fT ‘WY ‘ZE61 seAeBUEne) 'O/ '9Or'E8 ISAS “quUer JO "G9BI JePEMYOH ‘69 ‘102-06 ISAS ‘quer 10g ‘0/61 JEPEMYOH 'gg ‘pG:E1 ABojoydiowoyAUd E96} YPC] Puke eUeABIEN ‘79 ‘66E:LZ ING 19S ‘suey ‘AU EEG) OUAEA ‘99 ‘O0L:Sz eIPAO!] ‘Z96I er pue eURABIEN ‘GQ ‘E11 :19 ‘885 JOG ‘YOSIG JOG “PGEL IPYEYUDZ 79 ‘LZ:E A 10S ‘peo UeIPU| ‘OOlq “9EG OBY PUR IYsOP ‘Eg ‘Eg YOoUUSZUeY SC ‘606L JONEM ‘29 ‘Lbe:%p “ZUeYd ‘101g wWEg 996} WNZ-seBer ‘19 ‘ZLOL:69 10a fT ‘WY ‘ze6l el60g pue I4SsMalUSIN “09 ‘BPZ:6E ‘ZED 0G ‘GOGI JOEYBWOOUS ‘6S ‘ELL :pZ ABQ|OYdJowO}AUd “p/6} COBY ‘8G ‘S}UB|q BuLamol{ Jo UOHBOYISSE|D JO WE}SAS payes6ayu| “| g61 JsINDUOID “JG ‘EBE:Z ‘UY 0G “PZEl J9IZIAIG “9G ‘6ZZ:6Z ‘SPD 10g ‘ZIeMYOS Jeg “6961 SSOJPUF ‘GG ‘/8L:0E ‘AUN “eH JOM ploy T Er6) Aaileg pue AWems “‘yG (19292 ‘AlUN ‘eH JOquY Plo ‘f ‘SpE Aaleg pue ISeN ‘ES ‘60: ABojoydiowoyAUd “pG6l BUESHES ‘ZG 'S/1:0G JOG ‘90S ‘UUIT fT ‘GEEL UBSHOIG ‘1G “vLL:g ABojoydiowoyAUd "GGG. UY UBUOWY PUe JeYOeS ‘0S ‘EEh:92 0G f WY GEG! IUSOP ‘Gr ‘P6b:0Z EIPU| 19S ISU] ‘JEN “DOJ “PGEL UISES “Bh ‘1OL:QE6L JON 108 “QEGL UOZUNeW ‘Zp :2G:G g 19S ‘peoy UBIPU| ‘90d “JEGL 'USOP “Qh ‘ZOE:61 BIPUI ‘IDS YSU] JEN ‘901d ‘ESL AWEMS ‘Gp ‘G/G:GG ‘10g ‘fT ‘WY ‘B96 CBees pue oUAed ‘pp ‘ZLI6 ‘LLL6:1S1 Bew 30g snd ‘g9z6l Jdels ‘ep ‘Bez 'GZ JASPIL ‘JOG UAS ‘OEE IIEUNP ‘Zr ‘OSP'6SL ROLY O26} UepUeUg pUe URAeYBeLeAe!IA ‘Ly ‘OL: JOG Tf Udr ~BEEL EmeZeLUNY ‘Op :/6Z:92 ZED 10g ‘BEBl ‘AESSEQ ‘GE ‘Zbh:12 BIULld ‘PEGI NOL ‘BE ‘€09'9p 0G ‘UU ‘OBEL UOSUIIIOD “JE ‘6ZEE JOG Tf ‘ISMY ‘GO6L eUUEYH ‘OE ‘Ezz:6EL ‘ZED JO B/El Aepleuyos “GE ‘EGz:1z ABojoydiowioyAud 1/6} Aelesow| ‘VE ‘SOE 0g ‘Bey 0g “2961 eUUeYM ‘EE ‘PB:ZL EEN JOG BOW ‘EQEL UBMNGE7 UPA ‘ZE 'GBZ!Z11 JON 10 PO6L UBIEr pue jidey “LE :2Z:6z Al “AUN NYOYOL ys] ‘Sey “dey 19S “~EgE) !yseARH ‘OE ‘yQ:LZ “90S ‘19S }eYOW BUSIIZ ‘T “SG6L WeAM ‘6z ‘EZ1:S ABojoydsowoyAUd “Gg6} seBeuyeYyg Pur UYOL “GZ ‘66L:EEL IOAZ ISAS JURld ‘OBE! SSEsPUJ ‘LZ ‘1ZE-vE “ZED 10g 206} UOSUYOr ‘OZ ‘ZOL:Br ‘PAeD JOG UNOsSsI ‘UY “1961 Ney ‘Sz ‘6zE:01 ABojoudsowoyAUd ‘0961 AUNW “bz ‘6PZ:ZOl ‘ZED 108 LPL AGIND ‘Ez ‘ESL:601 “ISAS ‘quUeL ‘Og “2BEl SSE4PUF ‘Zz ‘ECL'SE JOG f JOBS] ‘QBEL OPH “LZ ‘ZEL:EZL ‘ZED 10g ‘ZOEL UISES OZ ‘EZE-E g ‘BIPUI 19S }SU| JEN “Old “QOGL INUJEW “GI ‘LEE:26 ISAS ‘Guyer Og “2/61 SSEsPU ‘Bl ‘202:G2 “JHSPIL "JOG “UEAS “LEGL WOIIOH “ZL ‘6Z:~EL ‘IOAZ ISAS qURld ‘OBL SSE‘PUZ 9} ‘/pb:p9 ‘AUN MURH UOqUY Ploy ‘fT ‘Eg6l UOSdWeS puUe ssoipUZ ‘G| ‘eep]EWeWeH a4} Jo Ai0jsIH |IssO4 PUB SOBWE}SAS ‘UONNIOAZ ‘BBL a|Aoq pue enyBouog ‘pI ‘79g ‘d ‘ABojoydsoy jue|q Ul SOUBAPY ‘Z/6L JIEN EL ‘QOE:ZE ‘AUT ‘eH wOquY POU f ‘9G6} Awems pue Aweseled ‘Z| ‘suuedsoibuy au} Jo ABojoAuquiy oeWa}SAS ‘GOG) SIAE ‘LI ‘ZEE:BP IOMUd MON ‘6P6L JOEUIOD ‘O} ‘G9G:02 ‘WEEN J0g B1OY LZ6} UBWWNOG 6 ‘SPZ:0E g ‘BIPU ‘IDS }SU] ‘JEN “OOld 9G] Wepueyg pue dey °g ‘Sol:z1 ABojoudiowoyAud '296} 20A0g pur Jasiaey ‘7 ‘E6E:L01 ‘SAS ‘quUer 10g ‘OBEL SSesPUF '9 ‘€8:06 ‘S8D }Og ‘YOSIG 48g ‘2/6 SSeIPUZ 'G ‘GO9:2E 10g T “ASNY ‘PEL Ie 18 YSPEHCd “b ‘GBZ 'L:ZG “AIUD “eH JOquYy plow ‘fT | Z61 "epueYg “€ ‘E0E:e1 ABojoydiowoyAUY “E96! WEPULYG 'Z ‘60S:6h AUN eH Joquy plow ‘f ggg) UBWeIe}eyUe, PUe WepUeYg ‘| "sa0uaJajoY 16 ninwg WNwg Jouul iquAs ddo JQUU! wAse eue xoud BIUUBLU9}9q — = Bag @PIOPIUULCUIA}8d day Ol 6 8 L9 S v € @ L SNNaD) ADO IOHdHOW ADO IOHdHOW ADOTOHdHOW SINAO WWHANAS) HV TISONN INANNDALN| ‘penuluod L xipueddy 34 Angiosperm Ovules and Carpels: Postilla 208 Their Characters and Polarities Appendix 2 Carpel character states for a select number of basal angiosperm families. The number of ovules in the family row is from Cronquist (1981). — = lack of data, () indicates that the state is rarely found in the taxon, and ? indicates that state is probable but not certain. Abbreviation for characters: character 11, morphology; character 12, number of traces: m = multiple trace; r = residual trace; character 13, position of placentae: /at. admedial = lateral, admedially; lat. admedial/marginal = lateral, admedially placed proximally and lateral, marginally placed distally; lat. exmedial = lateral, exmedially; /at. marginal = lateral, marginally; /at. radial = lateral, radially; character 14, orientation; character 15, number and arrangement: alt. = alternate; opp. = opposite. CARPEL POSITION OF NUMBER AND MORPHOLOGY PLACENTAE ORIENTATION ARRANGEMENT FAMILY Lee ee eee ee SS eee (GENUS) lat 12 13 14 15 REF. Winteraceae 1 to many Drimys ascoplicate 2 lat. admedial/ nonplanar many, opp. 1 marginal plicate A lat. marginal Pseudowintera ascoplicate 2 lat. admedial nonplanar many, opp. to apical Degeneriaceae many Degeneria plicate A 2 lat. marginal nonplanar many, alt. 2 Himantandraceae 1 (2) Galbulimima ascidate 2 (1) lat. admedial nonplanar single 3 (plicate A 2) (lat. marginal) (double, alt.) Eupomatiaceae 2 to ca. 11 Eupomatia ascidate 2 lat. admedial nonplanar 2 to 3, alt. 4 Austrobaileyaceae 8 to 14 Austrobaileya ascidate 2 lat. admedial nonplanar many, alt. 5 Magnoliaceae 2 to 6 (many) Liriodendron ascidate 1 lat. admedial nonplanar single to dou- 6 ble, opp. Magnolia ascidate 1 lat. admedial nonplanar double, opp. 7 split Lactoridaceae 4to8 Lactoris plicate — 2 lat. marginal nonplanar few-many, 8 alt. Annonaceae 1 to many Cananga plicate — 2 lat. marginal nonplanar many, alt. 9 Polyalthia ascidate = basal planar single 10 Myristicaceae 1 Myristica ascidate m basal planar single 11 Canellaceae 2 to many Warburgia plicate B 2 lat. marginal nonplanar many — 12 Amborellaceae 1 Amborella ascidate 1 lat. admedial planar single 13 Trimeniaceae 1 Trimenia ascidate 1 apical planar single 14 Monimiaceae 1 Monimioideae Hedycarya ascidate 1 apical planar single 15 Wilkiea ascidate 1 apical planar single 16 Siparunioideae Siparunia ascidate 1 basal planar single 17 Hortonioideae Hortonia ascidate 1 apical (lat. planar single 18 admedial) (nonplanar) (double, alt.) Atherospermoideae Laurelia ascidate 1 basal to lat. planar single 19 admedial Continued on next page 35 Angiosperm Ovules and Carpels: Their Characters and Polarities Appendix 2 Continued FAMILY (GENUS) Gomortegaceae Gomortega Calycanthaceae Calycanthus Idiospermaceae Idiospermum Lauraceae Cassytha Cinnamomum Laurus Hernandiaceae Gyrocarpus Chloranthaceae Ascarina Chloranthus Hedyosmum Sarcandra Saururaceae Anemopsis Houttuynia Saururus Piperaceae Heckeria Peperomia Piper Aristolochiaceae Aristolochia Asarum Illiciaceae Ilicium Schisandraceae Schisandra Nelumbonaceae Nelumbo Nymphaeaceae Nymphaea Nuphar Barclayaceae CARPEL MORPHOLOGY 11 ?ascidate plicate — ascidate ascidate ascidate ascidate ascidate ascidate ascidate ascidate ascidate plicate B plicate B ascidate to split ascidate ascidate ascidate plicate A ascidate ascoplicate ascidate ?ascidate ascidate ascidate (ascopli- cate) both dorsal and ventral ascidate both dorsal and ventral Postilla 208 POSITION OF PLACENTAE ORIENTATION 12 13 14 apical ?planar lat. marginal nonplanar lat. admedial nonplanar apical planar apical planar apical planar apical planar apical planar apical planar apical planar apical planar lat. marginal nonplanar lat. marginal nonplanar lat. admedial nonplanar basal aplanar basal aplanar to planar basal aplanar lat. marginal nonplanar lat. admedial nonplanar lat. admedial/ nonplanar marginal basal planar basal to lat. nonplanar admedial apical planar lat. radial nonplanar lat. radial nonplanar to chaotic Continued on next page NUMBER AND ARRANGEMENT 15 { single 1 (2) double, alt. (one aborts) 1 (2) single, dou- ble, alt. 1 single single single 1 single 1 single single single single (1) 2 to 10 few to many, alt. many, alt. double, alt., opp. 1 single single single many many, alt. few to many, alt. many, opp 1 single 2 to 5 (to 11) double, ?alt. 1 (2) single many many, alt. many, alt. many REF. 20 21 22 38 36 Angiosperm Ovules and Carpels: Their Characters and Polarities Appendix 2 Continued FAMILY (GENUS) Barclaya Cabombaceae Brasenia Cabomba Ceratophyllaceae Ceratopyllum Ranunculaceae Anemone Aquilega Delphinium Ranunculus Thalictrum Circaeasteraceae Circaeaster Berberidaceae Achlys Berberis Caulophyllum Epimedium Nandina Sargentodoxaceae Sargentodoxa Lardizabalaceae Akebia Decaisnea Menispermaceae Cocculus Menispermum Tinospora Coriariaceae Coriaria Sabiaceae Meliosma CARPEL MORPHOLOGY 11 ascoplicate both dorsal and ventral ascidate both dorsal and ventral ascidate both dorsal and ventral ascidate ascoplicate ascoplicate plicate B ascidate ascidate ascidate ascidate ascidate ascidate ascoplicate ascidate ascidate plicate B both dorsal and ventral plicate — ascidate split ascidate split ?ascidate ascidate ascidate Postilla 208 POSITION OF PLACENTAE ORIENTATION 12 13 14 m lat. radial nonplanar m lat. exmedial nonplanar to planar m chaotic nonplanar = apical to lat. ?planar exmedial 1 lat. admedial/ planar/non- marginal planar 2 lat. admedial/ nonplanar marginal 2 lat. marginal nonplanar 1 basal to lat. planar admedial 1 lat. admedial planar 1 apical to lat. ?planar- admedial nonplanar 1 basal to lat. planar admedial 2 basal nonplanar 2 basal nonplanar 2 lat. admedial/ nonplanar marginal 2 lat. admedial nonplanar = apical to lat. ?planar admedial m chaotic nonplanar = lat. marginal nonplanar = lat. admedial nonplanar = lat. admedial nonplanar = lat. admedial nonplanar 1 apical to lat. planar admedial = lat. admedial nonplanar Continued on next page NUMBER AND ARRANGEMENT 15 many, alt. (1)2 to3 single to few, alt. few, alt. 1 single 1 to many single/few, opp. many, alt. ?few, Opp. single single to dou- ble, opp. 1 to2 single to dou- ble, opp. 1, 2 to many single double, opp. double, opp. many. alt. single 1 single (few) many many, alt. many, alt. 2 double, alt. (1 aborts) double, opp. double, ?opp. (1 aborts) 1 single 1 to2 double, opp. REF. 39 40 37 55 SY/ Angiosperm Ovules and Carpels: Postilla 208 Their Characters and Polarities Appendix 2 Continued CARPEL POSITION OF NUMBER AND MORPHOLOGY PLACENTAE ORIENTATION ARRANGEMENT FAMILY (GENUS) 11 12 13 14 15 REF. Papaveraceae (1) many Eschscholtzia plicate A 2 lat. marginal nonplanar many, alt. 57 Sanguinaria plicate A 2 lat. marginal nonplanar many, alt. 57 Fumariaceae 2 to many Corydalis plicate B 2 lat. marginal nonplanar 2, alt. 58 Fumaria plicate — 2 lat. marginal nonplanar single 58 Tetracentraceae 5 to 6 Tetracentron plicate A 2 lat. marginal nonplanar few, alt. 59 Trochodendraceae many Trochodenaron plicate A 2 lat. marginal nonplanar many, alt. 59 Cercidiphyllaceae many Cercidiphyllum plicate A 2 lat. marginal nonplanar many, alt. 59 Eupteleaceae 1 to 3 (4) Euptelea ascidate 1,2 apical nonplanar single to few, 59 clustered Platanaceae 1 (2) Platanus ascidate 2 apical-lat. ad- nonplanar single to dou- 60 medial ble, opp. Hamamelidaceae 2 to several Altingia plicate A 2 lat. marginal nonplanar many, alt. 61 Distylium ascidate split 2 lat. admedial nonplanar single 62 Hamamelis plicate A 2 lat. marginal nonplanar single (2) 63 Liquidambar ascoplicate — basal to lat. nonplanar many, alt. 61 admedial/ marg. plicate A 2 lat. marginal nonplanar many, alt. Myrothamnaceae many Myrothamnus ascidate 2 lat. admedial nonplanar few, alt. 64 (ascoplicate) (lat. admedi- al/marginal) Phytolaccaceae 1 Didymotheca ascidate 21 lat. admedial planar single 111 Gisekia ascidate = basal planar single 65 Hilleria ascidate r basal planar single 66 Phytolacca ascidate 1 basal to lat. planar single 66 admedial Molluginaceae 1 to many Mollugo ascoplicate 2 lat. admedial/ nonplanar many, alt. 67 marginal Limeum ascidate = basal ?planar single 112 Caryophyllaceae 1 to many Melandrium ascidate lat. admedial nonplanar few, opp. 68 Scleranthus ascidate 1, ?r lat. admedial ?aplanar single 69 to basal Telephium ascidate 2 lat. admedial nonplanar many, opp. 69 Uebelinia ascidate r lat. admediai ?aplanar single 70 Polygonaceae 1 Brunnichia ascidate r basal = single 71 Rheum ascidate r basal aplanar single 71 Plumbaginaceae 1 Plumbago ascidate r basal = single 72 Vogelia ascidate = basal = single 73 Dilleniaceae 1 to many Continued on next page Continued on next page 38 Angiosperm Ovules and Carpels: Postilla 208 Their Characters and Polarities Appendix 2 Continued CARPEL POSITION OF NUMBER AND MORPHOLOGY PLACENTAE ORIENTATION ARRANGEMENT FAMILY (GENUS) 11 12 13 14 15 REF. Acrotrema ascoplicate 2 lat. admedial/ nonplanar double to 74, 75 split marginal many, alt. Didesmanara ascidate r basal planar single 74, 75 Dillenia ascoplicate m lat. admedial/ nonplanar few to many, 74, 75 (two pairs) marginal alt. Hibbertia ascidate split 2 lat. admedial nonplanar double, opp. 75, 76 ascoplicate 2 lat. admedial/ nonplanar alt. THA 93 marginal plicate A 2 lat. marginal nonplanar many, alt. 77 Pachynema ascidate split 2 lat. admedial nonplanar double, opp. 75, 76 Paeoniaceae several to many Paeonia plicate B = lat. marginal nonplanar few to many, 78 alt. Brunelliaceae 2 Brunellia ascidate 2 basal-lat. ad- nonplanar double, opp. 79 (notched) medial Connaraceae 2 Connarus ascidate (to 2 lat. admedial nonplanar double (1), 80 split) opp. notched Roreopsis ascidate split — lat. admedial nonplanar double, opp. 80 Eucryphiaceae several to many Eucryphia ascidate 2 lat. admedial nonplanar few to many, 81 alt. Cunoniaceae (1), 2 to many Aphanopetalum ascidate 1 apical-lat. ad- planar single 113 medial Ceratopetalum ascidate = lat. admedial nonplanar few, opp. 82 Cunonia ascoplicate = lat. admedial/ nonplanar many, opp. 82 marginal Schizomeria ascidate split — lat. admedial nonplanar double to 113 few, opp. Davidsoniaceae 5 to7 Davidsonia ascidate 2 lat. admedial nonplanar few to many, 113 clustered Dialypetalanthaceae many Dialypetalanthus ascidate = lat. admedial nonplanar many, — 84 Greyiaceae many Greyia ?ascidate = lat. admedial nonplanar many, alt. 55 Crassulaceae (1) few Bulliarda plicate — — lat. marginal nonplanar many, opp. 52 Kalanchoe ascoplicate 2 lat. admedial/ nonplanar many. alt. 85 marginal Sedum ascoplicate = lat. admedial/ nonplanar many, alt. 52 marginal Cephalotaceae 1 (2) Cephalotus ascidate = basal planar single 86 Saxifragaceae (1) several 39 Angiosperm Ovules and Carpels: Their Characters and Polarities Appendix 2 Continued FAMILY (GENUS) Parnassia Penthorum Peltiphyllum Tiarella Rosaceae Amelanchier Crataegus Fragaria Spiraea Neuradaceae Crossosomataceae Crossosoma Surianaceae Suriana Butomaceae Butomus Limnocharitaceae Hydrocleys Limnocharis Alismataceae Alisma Trichopodaceae Trichopus Dioscoreaceae Dioscorea Stemomeris Taccaceae Schizocapsa Tacca Stemonaceae Croomia Stemona Trilliaceae Medeola Paris Scoliopus Trillium Smilacaceae Rhipogonum CARPEL MORPHOLOGY 11 ascoplicate ascidate split plicate A plicate — ascidate split ascidate (split) plicate A ascoplicate split ?plicate — ascidate split plicate A dor- sal and ventral ascoplicate ascidate ascidate ascidate ?ascidate ascidate plicate B plicate B ascidate ascidate plicate B ascidate ascidate split ascidate ascoplicate ascidate 12 Continued on next page Postilla 208 POSITION OF PLACENTAE ORIENTATION 13 14 lat. admedial/ nonplanar marginal apical nonplanar lat. marginal nonplanar lat. marginal nonplanar lat. admedial nonplanar basal to lat. nonplanar admedial lat. marginal nonplanar lat. admedial/ nonplanar marginal lat. marginal nonplanar basal to lat. nonplanar admedial lat. radial nonplanar lat. radial nonplanar lat. radial nonplanar basal planar lat. admedial planar lat. admedial nonplanar lat. admedial nonplanar lat. marginal nonplanar lat. marginal nonplanar apical nonplanar basal nonplanar lat. marginal nonplanar lat. admedial nonplanar lat. admedial nonplanar lat. admedial lat. admedial/ marginal lat. admedial ?planar NUMBER AND ARRANGEMENT 15 many, alt. many, clus- tered many, alt. many, alt. 1, 2 to many double opp. double, opp. 1 (2), alt. few, opp. { (1), 2 to many many, — 2 (to 5) double, opp. many many, alt. many many, alt. many, alt. 1 (Several) single 1,2 single, dou- ble, opp. 2 to many double (1), — many, opp. many many, clus- tered many, alt. 2 (many) few, opp. few, cluster few to many, opp. many, — double, opp. many, alt. 1,2 single REF. 97 98 99 100 101 109 40 Angiosperm Ovules and Carpels: Postilla 208 Their Characters and Polarities Appendix 2 Continued CARPEL POSITION OF NUMBER AND MORPHOLOGY PLACENTAE ORIENTATION ARRANGEMENT FAMILY ee ee ee ee eee eee (GENUS) 11 12 13 14 15 REF. Smilax ascidate = apical ?planar-non- — single-dou- 110 planar ble, ?0pp. Petermanniaceae many Petermannia plicate- — lat. marginal nonplanar many, — 98 References. 1. Bailey and Nast 1943. J. Arnold Arbor. Harv. Univ. 24:472; 2. Bailey and Smith 1942. J. Arnold Arbor. Harv. Univ. 23:356; 3. Bailey, Nast and Smith 1943. J. Arnold Arbor. Harv. Univ. 24:190; 4. Endress 1977. Ber. Dtsch. Bot. Ges. 90:83; 5. Endress 1980. Bot. Jarhb. Syst. 101:393; 6. Canright 1960. Am. J. Bot. 47:145; 7. Boer and Bouman 1972. Acta Bot. Neerl. 21:617; 8. Carlquist 1964. Aliso 5:421; 9. Periasamy and Swamy 1956. J. Arnold Arbor. Harv. Univ. 37:366; 10. Corner 1949. New Phytol. 48:332; 11. Wilson and Maculans 1967. Am. J. Bot. 54:214; 12. Wilson 1966. Am. J. Bot. 53:336; 13. Bailey and Swamy 1948. J. Arnold Arbor. Harv. Univ. 29:245; 14. Endress and Sampson 1983. J. Arnold Arbor. Harv. Univ. 64:447; 15. Sampson 1969. Austr. J. Bot. 17:403; 16. Endress 1980. PI. Syst. Evol. 134:79; 17. Heilborn 1931. Sven. Bot. Tidsk. 25:202; 18. Endress 1980. Pl. Syst. Evol. 133:199; 19. Money, Bailey and Swamy 1950. J. Arnold Arbor. 31:372; 20. Reiche, 1896. Ber. Dtsch. Bot. Ges. 14:225; 21. Mathur 1968. Proc. Nat. Inst. Sci. India, B 34:323; 22. Blake 1972. Contrib. Queensland Herb. 12:1; 23. Sastri 1962. Bot. Gaz. 123:197; 24. Endress 1972. Bot. Jahrb. Syst. 92:331; 25. Endress 1987. Bot. Jahrb. Syst. 1090:153; 26. Quibell 1941. Bot. Gaz. 102:749; 27. Murty 1960. Phytomorphology 10:329; 28. Rajh 1961. Ann. Missouri Bot. Gard. 48:107; 29. Johnson 1902. Bot. Gaz. 34:321; 30. Tucker 1980. Am. J. Bot. 67:686; 31. Kanta 1962. Phytomorphology 12:207; 32. Johri and Bhatnagar 1955. Phytomorphology 5:123; 33. Wyatt 1955. J. Elisha Mitchell Sci. Soc. 71:64; 34. Robertson and Tucker 1979. Am. J. Bot. 66:605; 35. Kapil and Jalan 1964. Bot Not. 117:285; 36. van Leeuwen 1963. Acta Bot. Neerl. 12:84; 37. Moseley 1961. Bot. Gaz. 122: 233. 38. Moseley 1971. Phytomorphology 21:253; 39. Schneider 1978. Bot. Gaz. 139:223; 40. Ito 1986. Bot. Mag. Tokyo 99:169; 41. Troll 1934. Planta 21:447; 42. Bessey, 1898. Bot. Gaz. 26:297; 43. Troll 1934. Planta 21:266; 44. Troll 1932. Planta 17:153; 45. Foster 1963. J. Arnold Arbor. Harv. Univ. 44:299; 46. Chapman 1936. Am. J. Bot. 23:340; 47. Saunders 1928. New Phytol. 27:175; 48. Stapf 1926. Curtis Bot. Mag. 151:9111, 9112; 49. Payne and Seago 1968. Am. J. Bot. 55:575; 50. Swamy 1953. Proc. Nat. Inst. Sci. India 19:307; 51. Joshi 1937. Proc. Indian Acad. Sci. B 5:57; 52. Payer, 1852. Traité 'Organogenie Comparée de la Fleur; 53. Joshi 1939. Am. J. Bot. 26:433; 54. Sharma 1968. Phytomorphology 18:143. 55. Mauritzon 1936. Bot. Not. 1936: 161; 56. Tucker 1976. Am. J. Bot. 63:289; 57. Dicksen 1935. J. Linn. Soc. Bot. 50:175; 58. Arber 1931. New Phytol. 30:317; 59. Endress 1986. Ann. Missouri Bot. Gard. 73:297; 60. Boothroyd 1930. Am. J. Bot. 17:678; 61. Bogle 1986. Ann. Missouri Bot. Gard. 73:325; 62. Rao 1974. Phytomorphology 24:113; 63. Shoemaker 1905. Bot. Gaz. 39:248; 64. Jager-Zurn 1966. Beitr. Biol. Pflanz. 42:241; 65. Joshi and Rao 1936. Proc. Indian Acad. Sci. B 3:71; 66. Echkardt 1954. Ber. Dtsch. Bot. Ges. 67:113; 67. Payne 1933. Univ. Kansas Sci. Bull. 21:399; 68. Bocquet 1959. Phytomorphology 9:217; 69. Rohweder 1970. Bot! Jahrb. Syst. 90:201; 70. Rohweder 1965. Bot. Jahrb. Syst. 83:406; 71. Laubengayer 1937. Am. J. Bot. 24:329; 72. Haupt 1934. Bot. Gaz. 95: 649; 73. Mathur and Khan 1941. Proc. Indian Acad. Sci. B 13:360; 74. Dickison 1971. J. Arnold Arbor. Harv. Univ. 52:319; 75. Wilson 1973. Phytomorphology 23:25; 76. Dickison 1968. J. Arnold Arbor. Harv. Univ. 49:317; 77. Wilson 1965. Phytomorphology 15:240; 78. Camp and Hubbard 1963. Am. J. Bot. 50:174; 79. Cuatrecasas 1970. Flora Neotropica 2; 80. Dickison 1971. J. Elisha Mitchell Sci. Soc. 87:77; 81. Dickison 1978. Am. J. Bot. 65:722; 82. Mauritzon 1939. Lunds Univ. Arsskr. 35(2); 83. Bange 1952. Blumea 7:293; 84, Rizzini and Occhioni 1949. Lilloa 17:243; 85. Tillson 1940. Am. J. Bot. 27:595; 86. Diels 1930. Die Nat. Pflaz. 18a:71; 87. Sharma 1968. Phytomorphology 18:193; 88. Subramanyan and Nair 1975. In: Form, Structure and Function of Plants, 401; 89. Nair and Narayaram 1962. Proc. Nat. Inst. Sci. India 28B:211; 90. Bensel and Palser 1975. Am. J. Bot. 62: 661; 91. Sterling 1965. Am. J. Bot. 52:418; 92. Sterling 1964. Am. J. Bot. 51:705; 93. Sattler 1973. Organogenesis of Flowers, a Photographic Text-atlas; 94. Sterling 1966. Am. J. Bot. 53:521; 95. Cronquist 1981. An Integrated System of Classification of Flowering Plants; 96. Gutzwiller 1961. Bot. Jahrb. Syst. Pflanz. 81:1; 97. Kaul 1976. Am. J. Bot. 63:175; 98. Dahlgren 1985. The Families of the Monocotyledons: Structure, Evolution, and Taxonomy; 99. Pax, 1888. Die Nat. Plfanz. T1 A5:130; 100. Knuth 1930. Die Nat. Plfanz. 15a:438; 101. Limpricht 1928. Das Pflanz. 92; 102. Pax 1888. Die Nat. Pflanz. T2 A5:127; 103. Tomlinson and Ayensu 1968. J. Arnold Arbor. Harv. Univ. 49:260; 104. Swamy 1964. Phytomorphology 14:458; 105. Berg 1962. Skr. Nor. Vidensk. Akad. Oslo, Mat. Natr. (NS) 3:1; 106. Krause 1930. Die Nat. Pflanz. 15a:374; 107. Berg. 1959. Skr. Nor. Vidensk. Akad. Oslo, Mat. Natr. 4; 108. Berg 1958. Skr. Nor. Vidensk. Akad. Oslo, Mat. Natr. 1-36; 109. Krause 1930. Die Nat. Pflanz. 15a:381; 110. Engler 1888. Die Nat. Pflanz. T1 A5:89; 111. Walter 19090. Das Pflanz. 83; 112. Bogle 1970. J. Arnold Arbor. Harv. Univ. 51:431; 113. Dickison 1975. Am. J. Bot. 62:433. } LT rie! i ee! Mia i i ay I | I mM 1 R , i i ve hi : i my, / z P et ut 7 : } ; ' f en! i ; real i my : fh 1 it 1 i i My fl a 0 ; jee it I Fes ys aah { i Lan "i F ; il ih i fii i f ; i i hn ; a i , Dood : i : i iets i ; , { i T { r oy ee f i : ele Of j at i i hi Len ia y re : er i I { i 7 a iV : i | ; : : a : ra Pa, i , i ity ‘ i i ! nares aT f f fi i 1 ; ‘ I : Hy Nai fr 1 a i i i os Tecra twill i t t , : { i i ; : na: fay i cay j : : : to un hall aay ha Lat yl f i : [ i ca : heal i { 2 a A el : fh i i ‘ i i Oo M i fin : : , ; Vi dt { { ; ; i i je at i Ppl i i iy., iy eat i { ‘ ; 1 fs an | i : i hoe , Mi hae ‘ ; ee a i Ted te i 1 : { fi i ; i ‘ ’ j 7 P ; - i : i i wy; t y af i : ice i ee af ae : en iu ily nti ie To Ve te ee See tate ts