Amer. J. Bot. 75(12): 1857-1870. 1988. POLLEN MORPHOLOGY OF THE EUCHARIDEAE (AMARYLLIDACEAE)' University of Florida-IFAS, Research and Education Center, 3205 College Avenue, Ft. Lauderdale, Florida 333 14; and Horticultural Systematics Laboratory, Department of Ornamental Horticulture, Gainesville, Florida 326 1 1 ABSTRACT Eucharis, Caliphruria, and Urceolina form a monophyletic group of petiolate-leaved, Neotropical Amaryllidaceae ecologically specialized to the understory of primary tropical rain forest below 2,000 m elevation. Pollen morphology of the three genera is surveyed. Pollen grains of all species of Eucharis, Caliphruria, and Urceolina are boat-shaped elliptic, monosulcate, heteropolar, and bilateral in symmetry. Exine sculpturing is semitectate-columellate and reticulate in all species examined. A transformation series in reticulum coarseness and pollen grain size is described. The large pollen grain with coarse reticulum of most Eucharis species is considered ancestral. The fine reticulation of Caliphruria is considered derived and the exine morphology of Urceolina is intermediate. Both of these genera have medium-sized pollen grains. Exine dimorphism common to all Urceolina, but rare in Eucharis and Caliphruria, may be symplesiomorphous among those taxa exhibiting this morphology. The three genera are largely uniform in pollen grain ultrastructure, with completely ektexinous exines. Pollen grain size in Eucharis is not closely correlated with style length. Several wide-ranging species show considerable intraspecific variation in pollen size. Parallelisms in pollen grain evolution among related tribes of Neotropical Amaryllidaceae are discussed. EUCHARISPlanchon and Linden, Caliphruria Herbert, and Urceolina Reichenbach form a monophyletic group of Neotropical Amaryllidaceae that is ecologically specialized to the understory of primary, rarely secondary, tropical rain forest, usually below 2,000 m elevation (Meerow, in press; Traub, 1963). All three genera have petiolate leaves, turgid seeds with copious, oily endosperm, and occur in edaphic situations of high fertility (Meerow, in press). As a group they represent a distinct tribal assemblage within "infrafamily" Pancratioidinae Traub (1957, 1963), a group delimited by various patterns of stamina1 connation, 2n = 46 as the most common somatic chromosome number, and an Andean center of diversity (Meerow, 1985, 1987a, in press). The species are rare in nature, endemism is high, and several species in each genus are undoubtedly close to extinction. Eucharis is the largest and most I Received for publication 30 March 1987; revision accepted 23 March 1988. We thank Nancy B. Dehgan for technical assistance, the curators of COL, F, FLAS, HUNT, MO, SEL, UC, and US for permission to remove pollen from herbarium sheets or for housing vouchers, and Donald E. Stone and two anonymous reviewers for their careful reading of this paper. This study was supported in part by National Science Foundation Dissertation Improvement G r a n t BSR8401208, and a Garden Club of America/World Wildlife Fund Fellowship in Tropical Botany. Florida Agricultural Experiment Station Journal Series No. 7727. widely distributed of the three genera, found from Guatemala to Bolivia. The center of diversity for the genus is in western Amazonas and the adjacent eastern Andes. In a recent monograph (Meerow, in press), seventeen species in two subgenera are recognized. Caliphruria consists of four species, three of which are endemic to Colombia (Meerow, in press). Urceolina (ca. 6-8 species) is so far known only from central and southern Peru on the eastern Andean slopes. Traub (197 1)combined all three genera into Urceolina without any supporting data, a taxonomic decision few workers have followed. The most recent treatments of the Amaryllidaceae (Traub, 1963; Dahlgren, Clifford, and Yeo, 1985) have placed Eucharis, Caliphruria, and Urceolina into a single tribe [Euchareae (Traub, 1963) or Eucharideae (Dahlgren et al., 1985)]together with Hymenocallis Salisb., Eurycles Salisb., and Calostemma Brown. The latter two genera are Australasian in distribution and Hymenocallis is entirely Neotropical. This concept of the tribe, however, appears to be polyphyletic (Meerow, in press). Fleshy seeds has been used as the unifying character of this tribe. On the basis of anatomical study, the globose or narrowly oblong, turgid seeds of Eucharis, Caliphruria, and Urceolina with their characteristic phytomelanous testa (Huber, 1969) and oily endosperm can not be 1858 AMERICAN JOURNAL OF BOTANY considered homologous with the fleshy, bulbiform (and sometimes viviparous) seeds of Hymenocallis, Eurycles, and Calostemma which lack phytomelan in the seed coat (Meerow, in press). The "pseudoseed" of Eurycles and Calostemma is actually an adventitious bulbil (Rendle, 190 1). The bulbiform propagules of Hymenocallis, while true seeds, have thick, fleshy chlorenchymous integuments with a well developed vascular system, and a starchstoring embryo (Rendle, 1901;Whitehead and Brown, 1940). These important seed differences have been overlooked by most phylogenists of the Amaryllidaceae. The narrower concept of the Eucharideae followed in this paper includes only Eucharis, Caliphruria, and Urceolina. Despite scattered reports in the literature (Erdtman, 1952), pollen morphology of the Amaryllidaceae sensu strict0 has not received a great deal of attention in a systematic context, despite proven value (Meerow and Dehgan, 1985; Meerow, Dehgan, and Dehgan, 1986; Meerow, 1987a) Brief notations on the pollen morphology of Eucharis were previously provided by Fischer (1890) and Meerow (1984a). In this paper, comparative pollen morphology of Eucharis, Caliphruria and Urceolina is reviewed and discussed in the context of generic, tribal and subfamilial evolution of the pancratioid Amaryllidaceae. MATERIALS AND METHODS -Nonacetolyzed pollen was prepared for SEM and TEM as described by Meerow and Dehgan (1985). Pollen size measurements were averaged for twenty grains examined with a Nikon Lapophot photomicroscope. Correlation of pollen size with style length was tested with SAS release 5.08 on the Northeast Regional Data Center (NERDC) of the University of Florida. Terminology follows Walker and Doyle (1975) and Erdtman (1969). RESULTS-Pollengrains of all species of Eucharis, Caliphruria, and Urceolina examined are boat-shaped elliptic, monosulcate, heteropolar, and bilateral in symmetry (Fig. 1, 3-9, 1 1-1 7, 19, 32, 33). The sulcus runs the length of the distal face of the grain (Fig. 12, 15). The pollen grains are semitectate-columellate with reticulate sculpturing composed of a network of muri and lumina (Fig. 1-19, 32-34). The reticulum becomes obsolete within the aperturate region of the grain where the exine is more or less finely baculate (Fig. 24). Pollen grain size-Pollen grain size is considerably variable in Eucharis (Table l), and [Vol. 75 a notable size class (sensu Walker and Doyle, 1975) difference occurs between Eucharis and both Caliphruria and Urceolina. Pollen of Eucharis falls mainly into the large size class (longest equatorial diameter 50- 100 pm), with the average longest equatorial diameters greater than 60 pm. The two known exceptions are E. castelnaeana and E. plicata subsp. brevidentata (55.8 and 59.9 pm, respectively). Pollen of Caliphruria and Urceolina falls into the medium size class with longest equatorial diameters averaging 50 pm. The majority ofEucharis species have pollen grains with longest equatorial diameters between 65 and 75 pm. Eucharis caucana (subg. Eucharis) has the largest pollen grains in the genus (98-1 10 pm), and E . astrophiala (subg. Eucharis) the second largest (83-86 pm). The former species also has the highest chromosome number in Eucharis, 2n = 138 (Meerow, in press), but E. astrophiala is diploid (2n = 46; Meerow, 1987d). Polar diameter of Eucharis pollen ranges from (39-) 45-60.6 pm, while that of Caliphruria and Urceolina is always less than 40 pm. A degree of infraspecific variation in pollen diameter is evident in some species of Eucharis (Table I), but must be interpreted with caution due to the small number of populations sampled. Eucharis formosa is a wide-ranging and morphologically variable species (Meerow, 1987b). Longest average equatorial diameter among the populations sampled of this species shows a 12.7% difference between the smallest and largest value. The two subspecies of E. plicata show a 15% differential in pollen diameter. Other species are much more uniform in pollen grain size. Eucharis astrophiala is a narrow endemic restricted to western Ecuador with distinctive leaf and androecial morphology that is consistent among all populations. Three populations of this species sampled show only a 3.8% difference in pollen diameter. Eucharis bouchei is a highly polymorphic, tetraploid species complex with limited distribution (Meerow, 1987c), but it shows only a 2.4% difference between the largest and smallest diameter pollen. The smallest pollen grains in Eucharis and Caliphruria are found in species with the smallest flowers (Table l), namely all species of Caliphruria and, in E. subg. Eucharis, E. castelnaeana. Nonetheless, one ofthe largest flowered species, E. sanderi (subg. Heterocharis) has small pollen grains relative to other large-flowered species. Large pollen grains characterize E. astrophiala, a species at the smaller end of flower size range in the genus. Since style length is directly correlated with perianth size in Eu- December 19881 MEEROW AND DEHGAN -POLLEN OF EUCHARIDEAE 1859 Fig. 1-6. SEM photomicrographs of Eucharis pollen grains. 1 , 2. E. astrophiala (Madison 3792, SEL). 1 . Whole grain, proximal polar view. 2. Exine sculpturing. 3-6. Whole grains, proximal polar views. 3. E. bonplandii (Bauml 686. HUNT). 4. E. bouchei var. dressleri (Meerow 1107, FLAS). 5 . E. candida (Asplund 19120. S). 6. E. castelnaeana (Schunk 14156, FLAS). All scales = ca. 5 am. 1860 AMERICAN JOURNAL OF BOTANY charis and Caliphruria, style length was tested for correlation with longest equatorial diameter of pollen. The Pearson correlation coefficient (Snedecor and Cochran, 1980) for style length with pollen size of 28 Eucharis and Caliphruria collections representing 16 species was only 0.379 (Table 1). Exine sculpturing -The reticulate exine sculpturing pattern of Eucharis, Caliphruria, and Urceolina (Fig. 1-19, 32-34), may be subdivided into three types on the basis of lumina width (Table 1).Type 1 is characteristic ofmost species ofEucharis (Fig. 1-6,7-11, 13, 15, 16). The reticulum of Type 1 exine is coarse, with the largest lumina widths equal to or greater than 5 pm. Type 1 exine can be further subdivided on the basis of muri width. In Type 1-A (Fig. 3- 1 1, 13, 15, 16), the muri are equal to or greater than 1 pm wide. This is the most common exine morphology of Eucharis. In Type 1-B exine, characteristic of only E . astrophiala, the muri are less than 1 pm wide (Fig. 1, 2). In Type 2 exine, lumina are 2-3 pm wide, and a marked reduction in reticulum coarseness occurs at the meridional faces of the grain. Only two species ofEucharis have Type 2 morphology, E. oxyandra (subg. Eucharis, Fig. 12), E. sanderi (subg. Heterocharis, Fig. 14), and one species of Caliphruria, C. korsakofii (Fig. 19). This reticulum dimorphism is also apparent on the pollen grains of E . x grandiflora (Fig. 16), a putative hybrid of E. moorei and E . sanderi (Meerow, in press). Type 2 exine is characteristic of all Urceolina species examined (Fig. 32-34). Width of the muri, however, is variable among the species with Type 2 sculpturing, ranging from less than 0.4 pm in E . oxyandra, to ca. 0.6 pm in Urceolina, ca. 0.75 pm wide in E . sanderi, and ca. 1 pm wide in C. korsakofii. Type 3 exine sculpturing, characteristic of the three Colombian species of Caliphruria (Fig. 17, 18), is finely reticulate with lumina only 12 pm wide, and the muri 0.5-0.6 pm wide. As in Type 1 the reticulum is predominantly consistent in coarseness throughout the grain surface. [VO~. 75 interposed between the ektexinous footlayer and the intine was observed in most sections (Fig. 21, 25-27, 30, 36), and probably represents an artifact of using unacetolyzed pollen rather than the presence ofendexine. Very small vertical perforations through the footlayer were observed in the ektexine of E . amazonica (Fig. 30), E. moorei (Fig. 27), E . x grandiyora (Fig. 3 l), and U. microcrater (Fig. 36). DISCUSSION-Large, boat-shaped-elliptic, monosulcate pollen grains with reticulate exine morphology are the most common type of pollen found in the Amaryllidaceae (Erdtman, 1952; Walker and Doyle, 1975; Meerow and Dehgan, 1985; Meerow et al., 1986). Similar morphology has been reported for many Liliaceae sensu lato (Erdtman, 1952; Walker and Doyle, 1975; Zavada, 1983), and conforms to the fossil form genus Liliacidites Couper, one of the major angiosperm pollen types described from early Cretaceous deposits (Doyle, 1973; Walker and Walker, 1984). This type of pollen morphology appears to be basic to the monocotyledonous orders in general (Doyle, 1973). Among the Amaryllidaceae, only one group of genera show a radical departure from this pollen morphology. Crinum L. and its allies (tribe Amaryllideae sensu Dahlgren et al., 1985) all have bisulculate pollen and spinulose exine sculpturing (Erdtman, 1952; Nordal, Rorslett, and Laane, 1977; Dahlgren and Clifford, 1982; Meerow, unpublished data). With the exception of Crinum, these genera are restricted to Africa, many ofthem endemic to South Africa. In an interesting example of convergence, Donoghue (1985) reported a similar divergence in the exine sculpturing patterns of the Caprifoliaceae. The Type I exine morphology that is characteristic of most Eucharis pollen seems to have phylogenetic significance within "infrafamily" Pancratioidinae (Meerow, 1985; Meerow and Dehgan, 1985). A large, white or yellow, fragrant, crateriform flower with a conspicuous stamina1 cup ("pancratioid," cf. Pancratium L.), involved with sphingid moth pollination (Morton, 1965; Bauml, 1979; Grant, 1983), is characteristic of at least one genus in each of the tribes recognized in the "infrafamily." These Pollen wall ultrastructure-Eucharis, Cali- genera are Eucharis, Hymenocallis Salisb. senphruria, and Urceolina pollen grains have uni- su stricto, Pancratium, Pamianthe Stapf, and form wall stratification patterns (Fig. 20-3 1, Paramongaia Velarde. All but Pancratium are 35, 36). The columellae arise from a thin foot- entirely Neotropical in distribution. It is as yet layer (usually ca. 0.2 pm thick), and the intine inconclusive if the "infrafamily" is monophyis as thick or thicker than the exine in most letic or if the pancratioid flower has evolved species. The tectum is quite fragile, and usually several times within the family. That the Neoca. 0.5 pm thick. No channeling is apparent in tropical tribes of the Pancratioidinae do repthe intine. A very thin, electron-dense layer resent a monophyletic group is a much more December 19881 MEEROW AND DEHGAN-POLLEN OF EUCHARIDEAE 1861 Fig. 7-12. SEM photomicrographs of Eucharis pollen grains. 7 , 8. Whole grains, proximal polar view. 7. E. corynandra (Ravenna 2090, K). 8. E. cyaneosperma (Seibert 2145, US). 9, 10. E. formosa (Meerow, 1103, U S ) . 9. Whole grain, proximal polar view. 10. Exine sculpturing. 1 1 , 12. Whole grains, proximal polar view. 11. E. plicata (Meerow 1025, FLAS). 12. E. oxyandra (Hutchison et al. 5983, UC),oblique distal polar view. All scales = ca. 5 pm. = s a p I[V .map ~ q o d IeIUlXOJd 'u!& aloqM '(SVM '9601 MOW) !z&oyus~oy 'urtl s '3-61 -8uun1dlnx au!x3 -gt .ma!A mlod 14s!a -LI '(x'.u.s iroy xa) vlvluapaqns .3 '8 1 ' L I 'Ma!n leu!pnipuol IEJ~IEI '(SW ' L Z [ [ MoAaaN) u ~ o , y j p u ~.g~-91 x .Map m[od 14~1 anb11qo 9 '(SW J '0111 ~ o . ~ d a!layJtnq ~ q ) S!AU~J!~VD x 'SI 'Map [eupn@uol pJaIq anbgqo '(sn '~OPSE~ ' N ! x !rapuvs ) . g -PI 'Ma!A J E I O ~~ E U I ! X O J ~'(s 'P[z~[ punldsv) V ~ U O Z - m v '3'$1 'sup~8a l o m ' LI* I -su!er%uanod u1.1n.iydfp~~ pue stlvymg jo sydeJ%o~~~uro~oyd ~ 3 -6 s 1-E I '81d December 19 8 81 MEEROW AND DEHGAN-POLLEN OF EUCHARIDEAE 1863 Fig. 20-3 1. TEM photomicrographs of Eucharis and Caliphruria pollen grain sections. 20, 22, 24, and 28. Whole grain sections. 2 1, 23, 25-27, 29-3 1. Sections through pollen grain wall. 20,21. E. astrophiala (Meerow 1111, FLAS). 22,23. E. bouchei var. bouchei (Meerow 1157, FLAS). 24,25. E. plicata var. plicata (Meerow 1025). 24. Note reduction of tectum near aprture. 25. Aperture/interaperture transition region. 26. C. subedentata (Meerow 1152, FLAS). 27. E. moorei (Meerow 1141, FLAS). 28, 29. E. sanderi (Cuatrecasas 16380. F). Black globules are extruded lipids. 30. E. amazonica (Meerow 1105, FLAS). 31. E. xgrandifora (Meerow 1127, FLAS). Lamellations and cytoplasmic breakdown visible in Fig. 3 1 are probably aberrations caused by pollen inviability. c = columellum, ek = ektexine, fl = foot layer, i = intine, Ig = lipid globule, p = perforation, s = sulcus, t = tectum. Scales = 10 pm (20, 22, 24, 28); 1 pm (21, 23, 25-27, 29-3 1). 1864 AMERICAN JOURNAL OF BOTANY p o l . 75 Fig. 32-36. SEM and TEM photomicrographs of Urceolina pollen. 32. U. urceolata (Allard 21516. US). 33, 34. U. ayacuchensis (Weberbauer 6652, F). 35, 36. U.microcrater (Schunke 13633, FZAS). 32. Whole grain SEM, proximal polar view. 33. Whole grain SEM, oblique lateral longitudinal view. 34. Detail of exine surface, SEM. 35. Whole grain section, TEM. 36. Section through pollen grain wall, TEM. c = columellum, ek = ektexine, fl = foot layer, i = intine, p = perforation, t = tectum. Scales = 5 pm (32, 33, 35); 1 pm (34, 36). robust hypothesis. All Neotropical Pancratioidinae are characterized by a chromosome number of 2n = 46 or derivations thereof (Di Fulvio, 1973; Flory, 1977; Meerow, 1984b, 1985, 1987a, b, d). In three Neotropical tribes of the Pancratioidinae, parallel trends in the evolution of floral morphology seem to have occurred. In each case, taxa with smaller, tubular or ventricose, brightly colored flowers with reduced stamina1 connation, and without noticeable fragrance appear to have diverged from taxa with typical pancratioid flowers (Meerow, 1985, 1987a, in press). A similar pattern occurs in all three lineages: 1) floral morphology of "derived" taxa suggests an ornithophilous pollination syndrome and 2) "derived" taxa are found, entirely or in part, at higher elevations than presumed ancestral taxa. All or some of the species of each of the genera with putatively December 19881 MEEROW AND DEHGAN-POLLEN OF EUCHARIDEAE 1865 Fig. 37-40. SEM photomicrographs of pollen grains of other genera of pancratioid Amaryllidaceae with putatively ancestral flower morphology. 37. Hyrnenocallis latifolia (Mill.) Roem. (Meerow 1121, FLAS), lateral longitudinal view. 38. Pararnongaia weberbaueri Velarde (Ferryra8638, MO), lateral longitudinal view. 39. Pancratiurn tenuifoliurnHoch. St. ex Rich (De Wilde 6859, MO), proximal polar view. 40. ParniantheparvifloraMeerow (Dodson & Thien 690, MO), proximal polar view. Scales = ca. 5 pm. primitive pancratioid floral morphology (i.e., Eucharis, Hymenocallis. Pamianthe, Pancratium, and Paramongaia) have large to very large, coarsely reticulate pollen (Fig. 3740). The pollen of related genera with divergent floral morphology shows reduction trends in both size and reticulum coarseness (Meerow, 1985; Meerow and Dehgan, 1985), both considered evolutionary trends for angiosperm pollen in general (Walker and Doyle, 1975). Colombian species of Caliphruria (Fig. 17, 18) show the greatest degree of divergence for these pollen characters in comparison with Eucharis. The differentiation of the reticulum into coarse and fine areas, characteristic of species with Type 2 exine, is largely restricted to monocot pollen (Doyle, 1973; Walker and Walker, 1984), and has been observed in some Liliaciditespollen from the early Cretaceous (Walker and Walker, 1984). The evolutionary polarity of this character is unclear, however. Meerow and Dehgan (1 985) described a transformation series fiom auriculate pollen through dimorphic reticulum to homogeneous reticulum among the subgenera of Hymenocallis (sensu Traub, 1962), suggesting that the ho- TABLEI. tl Continued $, B CI) Taxon Voucher Polar" diameter ( 4 72.5 (k2.80) 71.5 (k2.66) 68.4 (k3.53) 68.9 ( k 1.84) 59.9 (k3.01) 69.9 ( k 3.04) Peru, Loreto, Alto Amazonas, Yurimaguas, Schunke 14157 Peru, San Martin, Lamas, Schunke 1417Ia E. oxyandra (Ravenna) Ravenna Peru, Huanuco, Huanuco, Carpish, Hutchison et al. 5983 (UC) E. plicata Meerow subsp. plicata Peru, San Martin, Mariscal Caceres, Tocache Nuevo, Plowman 13951 E plicata subsp. brevidentata Meerow Bolivia, ex hort F. Fuchs, Meerow I143 E. ulei Kranzlin Peru, San Martin, Mariscal Caceres, Tocache Nuevo, Meerow 1024 Eucharis subg. Heterocharis E. amazonica Linden ex Planchon Peru, Huanuco, Leonicio Prado, Rupa Rupa, Schunke I4179 E. sanderi Baker Ecuador, Cotopaxi, km 52 on rd. from Quevedo-Latacunga, Rio Pilalo, Meerow & Meerow I141 Colombia, El Valle de Cauca, Rio Calima, La Trojita, Cuatrecasas 16380 (F) E . x grandiflora Planch. & Lind. x Calicharis butcheri (Traub) Meerow Colombia, El Valle de Cauca, Rio Anchicaya, Meerow & Teets I I27 ex hort Kew, Meerow 11 10 E. moorei (Baker) Meerow Caliphruria C. hartwegiana Herbert C. korsakojii (Traub) Meerow Colombia, Huila, La Plata, Di'az et al. 534 (COL) Peru, San Martin, Hierra waterfalls, Meerow 1096 C. subedentata Baker Colombia, ex hort, Meerow I152 C. tenera Baker Colombia, Cundinamarca, Triana 1289 (COL) Urceolina U ayacucensis Ravenna U. microcrater Kranzlin U. urecolata (R. & P.) M. L. Green * Peru, Ayacucho, La Mar, Rio Masamerich, Weberbauer 6652 (US) Peru, Huanuco, Leoncio Prado, JosC Crespo y Castillo, Aucayacu, Schunke 13633 Peru, Cusco, Quispicanchi, Marcapata Valley, Weberbauer 7822 (US) Longesta equatorial diameter bm) 48.6 (k2.89) 39.8 (k3.01) - 35.4d 32.3 (k2.47) 39.3 (k3.63) 35.2d 32.3d 29.8 (k3.58) 33.9 (k3.45) Mean of twenty grains. Standard deviations are in parentheses below means. See text for explanation. No values are given for hybrids. The high percentage of deformed, inviable pollen grains causes enormous variance in size measurements. Values without standard deviations were derived from statistically insignificant quantities of pollen. Exineb type style length (mm) 52.8 @ 2 - $ 00 49.6 2 32.8 I-A 27.5 1-A 32.0 1-A 44.6 5 2 1868 AMERICAN JOURNAL OF BOTANY mogeneous reticulum is an advanced character state. Extreme dimorphism is exemplified by the auriculate pollen grains of Hymenocallis species (Fig. 37; Meerow and Dehgan, 1985). A dense area of narrow, baculate outgrowths from the tectum occurs on pollen grains of the monotypic genus Paramongaia (Fig. 38) in the same regions of the grain where the auriculae of Hymenocallis pollen are located. Pollen of Pancratium tenuifolium (Fig. 39) shows a very abrupt reduction in reticulum coarseness at the equatorial ends of the grain. The three species of Eucharis and Caliphruria with dimorphic exine sculpturing ( E . oxyandra, Fig. 12; E. sanderi, Fig. 14; and C. korsakofii, Fig. 19) each represent divergent taxa of their respective genus or subgenera (Meerow, in press). The dimorphic reticulum in these three species may thus be symplesiomorphous (sensu Hennig, 1966) with LTrceolina, all species of which have slightly dimorphic exines (Fig. 32, 33). Alternatively, each of three species differ in muri width, thus the Type 2 exine morphology may have had an independent, derived origin in each of the three. In width of the lumina, the pollen of E. oxyandra (Fig. 12) resembles that of Urceolina (Fig. 32, 33), though pollen of the latter genus fits the medium size class of Walker and Doyle, and muri of E. oxyandra pollen are narrower than those of Urceolina. Eucharis oxyandra is a problematic species morphologically as well, with androecial characters intermediate between Eucharis and Urceolina (Meerow, in press; Ravenna, 1983). Meerow (in press) has suggested that E. oxyandra may represent a relict entity related to the ancestor of LTrceolina, a possible intergeneric hybrid, or a peripheral isolate from the Amazonian center of distribution for Ezlcharis. At present, this species, known only from transient cultivation in Peru, is too poorly understood to confirm any of these hypotheses concerning its origins. Taxonomically, it has been assigned to Eucharis subg. Eucharis, but with the designation incertae sedis (Meerow, in press). Kress and Stone (1982) and Zavada (1983) conclude that the lack ofendexine in the pollen grain wall is a characteristic of most monocot pollen. Previous reports for amaryllid pollen to date have not observed an endexine layer (Afzelius, 1956; Gullvag, 1964; Larson, 1964; Zavada, 1983; Meerow and Dehgan, 1985; Meerow et al., 1986). Simpson (1985) recently observed endexine in the pollen grain walls of severalmembers of the Tecophilaeceae. Gukdks (1 982), however, strongly argues that both the chemical (see Kress and Stone, 1982, for review) and electron density criteria for identi- [Vol. 75 fying endexine in pollen are insufficiently reliable. The slight exine stratification we observed in unacetolyzed pollen of the Eucharideae probably does not represent an endexine layer. The thin footlayer and columellate structure of the exine found in Eucharis and Caliphruria is, however, common to all other genera of the Pancratioidinae that we have examined (Meerow and Dehgan, 1985; Meerow et al., 1986, Meerow, 1987a). It may be basic to the Liliflorae in general (Doyle, 1973; Walker and Walker, 1984; Simpson, 1985), and thus appears highly conserved. Within the Eucharideae, we find a similar correlation between patterns of flower and pollen morphological divergence as had been previously described for the subgenera of Hymenocallis (Meerow and Dehgan, 1985). Primary divergences from the pancratioid flower, as represented in the Eucharideae by Caliphruria and Urceolina, have smaller pollen grains with reduced coarseness of the exine reticulum. Pollen of Caliphruria, with regard to the latter character, is more highly specialized. The similarity in exine reticulum morphology of Eucharis to that of other putatively basal pancratioid genera such as Hymenocallis subg. Hymenocallis and Pancratium is striking (Fig. 37-40), and is, at present, the best evidence that "infrafamily" Pancratioidinae represents a monophyletic group. Of the approximately sixty genera of Amaryllidaceae that we have presently examined, most with reticulate exine sculpturing, only two species of Pyrolirion Herbert ("infrafamily" Amarylloidinae), P. aureum Herbert and P. cutleri (Cardenas) Ravenna, have had lumina and muri approaching the dimensions characteristic of pollen of the pancratioid base. Cruden and Lyon (1985) observed that all studies which have shown a strong correlation between pollen grain size and style length involved related species, while noncorrelating studies involved unrelated taxa. Testing correlations between both style length and stigma depth (an approximation of the distance a pollen tube must grow to reach exogenous resources in the transmission tissue of the style) and pollen grain volume among species of several genera in several families, Cruden and Lyon (1985) concluded that style length has little correlation with pollen size, while stigma depth was highly correlated with style length. Where style length and pollen grain volume do correlate, i.e., among related species, they suggested that phylogeny, rather than function, is represented. Within the pancratioid Amaryllidaceae, there appears to be some correlation between style length and pollen size. Many December 19881 MEEROW AND DEHGAN- .POLLEN OF EUCHARIDEAE species of Hymenocallis subg. Hymenocallis have styles as long as 20 cm, and very large pollen grains over 100 pm long (Meerow and Dehgan, 1985). However, in Eucharis and Caliphruria as a whole, we found little correlation between style length and pollen grain size (as represented by longest equatorial diameter in our study, rather than volume). Eucharis caucana, the species with the longest pollen grains in the genus, is intermediate in flower size. Eucharis astrophiala, the species with the second largest pollen grains in the genus, is among the small-flowered group of species in the genus, and a large flowered species, E. sanderi, has relatively small pollen grains (Table 1). In conclusion, the evolution of novel floral adaptations within the Eucharideae has been accompanied by reduction in pollen grain size, and coarseness of the exine reticulum. Pollen grains of Eucharis, the most cladistically basal of the three component genera (Meerow, in press), are the largest in size and have the coarsest reticulation pattern. Both of these states are judged to be ancestral on the basis of comparisons among other tribes in the Pancratioidinae (Meerow and Dehgan, 1985). Though pollen grains of Caliphruria and Urceolina are similar in size (medium-size class), exines of Caliphruria species are the most derived in their sculpturing pattern. Exine sculpturing of Urceolina pollen is conceived as intermediate between Eucharis and Caliphruria. Within Eucharis there is no consistent correlation between pollen grain size and style length, and size differences among the species are only rarely significant. LITERATURE CITED AFZELIUS, B. M. 1956. On the fine structure of the pollen wall in Clivia miniata. Bot. Not. 108: 141-143. BAUML,J. A. 1979. A study of the genus Hymenocallis (Amaryllidaceae) in Mexico. M.S. thesis, Cornell University, Ithaca, NY. CRUDEN,R. W., AND D. L. LYON. 1985. Correlations among stigma depth, style length, and pollen grain size: do they reflect function or phylogeny. Bot. Gaz. 146: 143-149. DAHLGREN, R. M. T., AND H. T. CLIFFORD.1982. The monocotyledons: a comparative study. Academic Press, London. -, AND P. F. YEO. 1985. The families of the monocotyledons. Springer-Verlag, Berlin. DI FULVIO,T. E. 1973. Contribution a1 conocimiento cariologico de Amaryllidaceae. Estudio cromosomico en Hieronymiella y otras genCros afines. Kurtziana 7: 117-131. DONOGHUE, M. J. 1985. Pollen diversity and exine evolution in Viburnum and the Caprifoliaceae sensu lato. J. Arnold Arb. 66: 421-469. DOYLE,J. A. 1973. Fossil evidence on early evolution of the monocotyledons. Quart. Rev. Biol. 48: 399413. 1869 G. 1952. Pollen morphology and plant taxERDTMAN, onomy, Vol. 1. Angiosperms. Almquist and Wiskell, Stockholm. . 1969. Handbook of palynology-taxonomy-ecology. Monksgaard, Copenhagen. FISCHER,H . 1890. Beitrage zur vergleichenden Morphologie der Pollenkorner. Thesis, Breslau. FLORY,W. S. 1977. Overview of chromosomal evolution in the Amaryllidaceae. Nucleus 20: 70-88. GRANT,V. 1983. The systematic and geographical distribution ofhawkmoth flowers in the temperate North American flora. Bot. Gaz. 144: 439-449. GUEDES,M. 1982. Exine stratification, ectexine structure and angiosperm evolution. Grana 2 1: 16 1-1 70. GULLVAG, B. 1964. The fine structure of the pollen of Clivia miniata. Grana Palynol. 5: 258-263. HENNIG,W. 1966. Phylogenetic systematics. University of Illinois Press, Urbana. HUBER,H. 1969. Die Sammenmerkmale und venvandtschafts verhaltnisse der Liliiflorae. Mitt. Bot. Staatssamml. Munchen 8: 219-538. KRESS,W. J., AND D. E. STONE. 1982. Nature of the sporoderm in monocotyledons with special reference to the pollen grains of Canna and Heliconia. Grana 21: 129-148. LARSON, D. A. 1964. Further electron microscope studies of exine structure and stratification. Grana Palynol. 5: 265-277. MEEROW, A. W. 1984a. Two new species of pancratioid Amaryllidaceae from Peru and Ecuador. Brittonia 36: 18-25. -. 1984b. Karyotype evolution in the Amaryllidaceae. Herbertia 40: 139-1 54. . 1985. The evolutionary significance of pancratioid floral morphology in the Amaryllidaceae. Amer. J. Bot. 72: 962. -. 1987a. A monograph of Eucrosia (Amaryllidaceae). Syst. Bot. 12: 460-492. -. 1987b. New species of Phaedranassa and Eucharis (Amaryllidaceae). Sida 12: 29-49. -. 1987c. Biosystematics of tetraploid Eucharis (Amaryllidaceae). Ann. Missouri Bot. Gard. 74: 278296. -. 1987d. Chromosome cytology of Eucharis, Caliphruria, and Urceolina. Amer. J. Bot. 74: 1559-1575. . In press. Systematics of the Amazon lilies, Eucharis and Caliphruria (Amaryllidaceae). Ann. Missouri Bot. Gard. -, AND B. DEHGAN. 1985. The auriculate pollen grain of Hymenocallis quitoensis Herb. (Amaryllidaceae) and its systematic implications. Amer. J. Bot. 72: 540-547. -, N. B. DEHGAN,AND B. DEHGAN. 1986. Pollen tetrads in Stenomesson elwesii(Amaryl1idaceae). Amer. J . Bot. 73: 1642-1644. MORTON,J. K. 1965. The experimental taxonomy of the West African species of Pancratium L. (Amaryllidaceae). Kew Bull. 19: 337-347. NORDAL,I., B. RORSLETT,AND M. M. LAANE. 1977. Species delimitation within the Crinum ornatum group (Amaryllidaceae) in East Africa. Nonveg. J. Bot. 24: 195-2 12. RAVENNA, P. F. 1983. Revisional studies in the genus Urceolina (Amaryllidaceae). 11. Wrightia 7: 25 1-253. RENDLE,A. B. 1901. The bulbiform seeds of certain Amaryllidaceae. J. Roy. Hort. Soc. 26: 89-96. SNEDECOR, G. W., AND W. G. COCHRAN.1980. Statistical methods, 7th ed. Iowa State University Press, Ames. 1870 AMERICAN JOURNAL OF BOTANY SIMPSON,M. G. 1985. Pollen ultrastructure of the Tecophilaeaceae. Grana 24: 77-92. TRAUB,H. P. 1957. Classification of the Amaryllidaceae: subfamilies, tribes, and genera. PI. Life 13: 76-8 1. -. 1962. Key to the subgenera, alliances and species of Hymmocallis. PI. Life 18: 55-72. . 1963. Genera of the Amaryllidaceae. American Plant Life Society, La Jolla. . 197 1. Amaryllid notes. PI. Life 27: 57-59. WALKER,J. W., AND J. A. DOYLE. 1975. The bases of angiosperm phylogeny: palynology. Ann. Missouri Bot. Gard. 62: 664-723. [VO~. 75 ,ANDA.G. WALKER. 1984. Ultrastructureoflower Cretaceousangiosperm pollen and the origin and early evolution of floweringplants. Ann. Missouri Bot. Gard. 71: 464-521. WHITEHEAD, M. R., AND C. A. BROWN. 1940. The seed of the spider lily, Hymenocallis occidentalis. Amer. J . Bot. 27: 199-203. ZAVADA, M. S. 1983. Comparative morphology of monocot pollen and evolutionary trends of apertures and wall structures. Bot. Rev. 49: 331-379.