\ W ZL > \, Ww &! Sy TIOST > Yi De /G°? ws ST, iN Mie Pas 22 GHW Za / Ui Mi \) 4, X “AK 7, Volume 52 1971 JOURNAL OF THE ARNOLD JZ — ARBORETUM HARVARD UNIVERSITY | _ i } if | MISSOUR! BOTANICAL ) i RY Dates of Issue No. 1 (pp. 1-204) issued 4 February, 1971. No. 2 (pp. 205-368) issued 16 April, 1971. No. 3 (pp. 369-522) issued 16 July, 1971. No. 4 (pp. 523-717) issued 21 October, 1971. Contents of Volume 52 Embryology of the Magnoliales and Comments on their Relationships. N. BHANDARI . The Genera of Urticaceae i in ‘the Southeastern United States. NorTON G. MILLER . Calliandra haematocephala: History, Morphology, and Taxonomy. Lorin I. NEVLING, JR., and THomas S. ELIas . The Ecology of an Elfin Forest in Puerto Rico, 14. The Algae of Pico del Oeste. JOHN W. FOERSTER . A Revision of the Boraginaceae of West Pakistan and Kashmir ee S. M. A. Kazmir . . The Genus Neillia (Rosaceae) i in Mainland Asia and in Cultivation. Bo OE > ce The Genera of Fagaceae i in the Southeastern United States. THOMAS S. ELIas . Statement of Ownershi What is the Primitive Floral Structare. of Araliaceae? RICHARD H. EypeE and CHARLES C. TSENG . Shoot Growth and Heterophylly in Acer. WILLIAM B. CRITCHFIELD . The Polygalaceae in the Southeastern United States. NorTON G. MILLER. . Embryology of the Magnoliales a and eae on their Saoaeraes Sesame . N. BHANDARI . 8 Genera of PRES in n the Southeastern United States. THOMas S. ELIAS Comparative Morphological ‘Studies i in Dilleniaceae, VII. Additional Notes on Accotrema: WILLIAM C. DICKISON . A Revision of the Boraginacea of West Pakistan and Kashmir ( ape KAZMI . . 197] 40 69 267 285 319 334 A New Species of Reynosia (Rhamnaceae) from J sat MARSHALL C. JOHN Studies of Pacific Island oe XXIII, The Genus Diospyros (Ebenaceae ) in Fiji, The Genera of Orobanchaceae i in the Southeastern United States. JOHN W. THIERET The Ecology of an Elfin Forest i in - Puerto Rico, 15. A Study of the Leafy — Flora of the Luquillo Mountai MancaKer FULFORD, BARBARA CRANDALL, and RAYMOND STOTLER . : Meee rcioaial Studies i in Cordyline (Agavaceae) I. Introduction and General Morphology B. TOMLINSON and J. B. FISHER . The Saururaceae in the Southeastern United States. CARROLL E, Woop, Jr. . A Revision of the Boraginaceae of West Pakistan and Kashmir (continued). . KAZMI . : Comparative Anatomy of Ulmaceae. DWARD M. SWEITZER . ; The Ecology of an Elfin Forest i in Puerto Rico, 16. The Flowering Cycle and an Interpretation of its Seasonality. Lorin I. NEVLING, JR. . Taxonomic and Nomenclatural Notes on Jamaican ge Rag Plants. WILLIAM T, STE : The oe irk in the a. United Sta Senna R ROBERTSON . : A Revision of the Boraginaceae of West Pakistan and Kashmir (concluded). >. AZMI Two New teers = # Leguminosae. G. P. Yakov Leonard J. Bras (1960-197 A lial cay aE : ® Appreciation, 364 369 435 459 479 486 23 586 614 Journal of the Arnold Arboretum Published quarterly in January, April, July, and October by the Arnold Arboretum, Harvard University. Subscription price $10.00 per year. Subscriptions and remittances should be sent to Miss Dulcie A. Powell, Arnold Arboretum, 22 Divinity Avenue, Cambridge, Massachusetts 02138, U.S.A. Volumes I-XLV, reprinted, are available from the Kraus Reprint Cor- poration, 16 East 46th Street, New York, New York 10017, U.S.A. EDITORIAL STAFF B. G. Schubert, Editor C. E. Wood, Jr. T. G. Hartley CIRCULATION D. A. Powell Printed at the Harvard University Printing Office, Boston, Massachusetts COVER: The Golden Larch, Pseudolarix Gordon The cover design is the work of Norman Comeau of the Art Depart- ment, Thomas Todd Company, Printers, Boston. Second-class postage paid at Boston, Massachusetts JOURNAL OF THE ARNOLD ARBORETUM VoL. 52 JANUARY 1971 NUMBER 1 EMBRYOLOGY OF THE MAGNOLIALES AND COMMENTS ON THEIR RELATIONSHIPS 2 N. N. BHANDARI EVER SINCE THE PROPOSAL of Bennettitalian origin of angiosperms, the magnolian complex and particularly the Magnoliaceae, have occupied a pivotal position in all phylogenetic considerations, since in the “Mag- nolia flower” the proponents of this theory visualized all the representative analogous reproductive parts in the Bennettitalian fructification. No doubt the various theories to interpret the angiosperm flower and its appendages, such as the telome theory (see Zimmermann, 1955; Wil- son, 1937, 1942, 1953; Wilson & Just, 1939), gonophyll theory (Melville, 1962, 1963) and interpretation of the angiosperm flower (Barnard, 1961; Sporne, 1949) are exhaustive mental exercises based directly on the ex- tinct plants. However, the fast emerging fact is that the living ranalian members should be subjected to an extensive survey from all points of view, in order to understand, if not the origin of the angiosperms, at least the basic primitive patterns and their further diversifications during the course of phylogeny. Consequently, this resulted in a concerted effort to investi- gate fully the morphology, anatomy, and embryology of this interesting group of the dicotyledons. It is rather fortunate that the group attracted the attention of two doyens of plant morphology and embryology, Professor I. W. Bailey and Professor P. Maheshwari, who established strong schools in their re- spective fields to gain an insight into the ranalian complex. By coordinat- ed effort over a span of 30 years or more Professor I. W. Bailey and his colleagues have projected explicitly the basic primitive morphological and anatomical patterns and their amplification, in ascending or descend- ing phylogenetic series, in numerous ranalian families. Their efforts have resulted in studies on the Magnoliaceae (Canright, 1952a,b; 1955; 1960; 1965), Degeneriaceae (Bailey & Smith, 1942; Swamy, 1949), Winter- aceae (Bailey, 1944b; Bailey & Nast, 1943a,b; 1944a,b; 1945; Nast, 1944), Himantandraceae (Bailey, Nast, & Smith, 1943), Schisandraceae * Dedicated to the memory of the late Professor P. Maheshwari, a renowned em- bryologist and my teacher, and the late Professor I. W. Bailey, a great plant anatomist and morphologist. 2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 and Illiciaceae (Bailey & Nast, 1948; Smith, 1947), Austrobaileyaceae (Bailey & Swamy, 1949), Trimeniaceae and Monimiaceae (Money, Bailey, & Swamy, 1950), Amborellaceae (Bailey, 1947; Bailey & Swamy, 1948), Gomortegaceae (Money, Bailey, & Swamy, 1950), Trochodendra- ceae and Tetracentraceae (Bailey & Thompson, 1918; Bailey & Nast, 1945; Nast & Bailey, 1945; Smith, 1945), Eupteleaceae (Nast & Bailey, 1946; Smith, 1946), and Cercidiphyllaceae (Bailey & Nast, 1945; Nast & Bailey, 1945; Swamy & Bailey, 1949). On the other hand, Professor P. Maheshwari interested some of his pupils in taking up a study of the embryology of some ranalian families. They have contributed substantially towards our knowledge of such families as the Magnoliaceae (Bhandari, 1967a; Kapil & Bhandari, 1964), Degeneriaceae (Bhandari, 1967b), Himantandraceae (Vijayaraghavan & Bhandari, 1970), Winteraceae (Bhandari, 1963b, 1967c; Bhandari & Ven- kataraman, 1968), Myristicaceae (Bhandari, 1967d), Canellaceae (Bhan- dari, 1967e), Schisandraceae (Bhandari, 1967f; Jalan & Bhandari, 1963; Jalan & Kapil, 1965; Kapil & Jalan, 1964), Illiciaceae (Bhandari, 1967g), Monimiaceae (Bhandari, 1967h), Calycanthaceae (Bhandari, 1967i), Lauraceae (Bhandari, 1967j), Trochodendraceae (Bhandari, 1967k), Cercidiphyllaceae (Bhandari, 19671), Chloranthaceae (Vijayaraghavan, 1964), and Ranunculaceae (Bhandari, 1962, 1963a, 1965, 1966, 1968; Bhandari & Asnani, 1968; Bhandari & Kapil, 1964; Bhandari & Vijayara- ghavan, 1970; Jalan, 1963; Jalan & Bhandari, 1963; Kapil & Jalan, 1962; Vijayaraghavan, 1962, 1963; Vijayaraghavan & Bhandari, 1970; Vija- yaraghavan & Marwaha, 1969a,b). | Erdtman (1952), Erdtman & Metcalfe (1963), and Wodehouse (1935, 1936) have contributed immensely towards the study of pollen mor- phology and its application in systematics. Pollen characters no doubt form one of the fundamental features in considering interrelationships amongst various families and must be fully utilized. The chromosome number and t meaningful in considering interrelat Unfortunately, as compared to the order, chromosome numbers for onl Furthermore, as pointed out by titiv : ; however, recent in- vestigations by Riidenberg (1967 ), Stone (1968). a man (1968) are welcome additions. en aa : ae ore realized that although it is not possible to erect a phylogenetic ci ~ assification based on embryological characters alone. it is also ba . e aN any other single series of characters, However that the Tyological characters play an important role in deciding the inter- 1971] BHANDARI, MAGNOLIALES 3 relationships in particular instances has been amply emphasized by Ma- heshwari (1950a,b, 1954, 1956, 1959, 1963, 1964), Johri (1963), Kapil (1962), Subramanyam (1962), and Cave (1959). Therefore, an attempt like the present one to tabulate and evaluate the embryological evidences in conjunction with others in phylogenetic considerations of the Mag- noliales needs no justification. The order Ranales sensu lato has undergone tremendous changes be- cause of the injunction of phylogeny in taxonomy (Cronquist, 1968; Eng- ler, 1964; Takhtajan, 1966; Thorne, 1968) and there have been many additions or deletions from one family to another and vice versa. F urther- more, a large number of taxa have been excluded from or included in this group by one or the other taxonomist. In this review I have followed the recently proposed system by Cronquist (1968) which includes 19 families in the order Magnoliales, comprising chiefly the woody members of the ranalian complex. The families Trochodendraceae and Cercidiphyllaceae have also been included although these along with Tetracentraceae and Eupteleaceae are placed in the orders Trochodendrales and Hamamelidales. This system (Cronquist, 1968) has been followed only to have a framework for the review without giving due consideration to any of its phylogenetic im- plications. Embryological information on Lactoridaceae, Eupomatiaceae, Trimeniaceae, Amborellaceae, Gomortegaceae, Hernandiaceae, Tetracen- traceae and Eupteleaceae is wanting. These have, therefore, been ex- cluded from the description of the families, but since well established re- lationships have been shown based chiefly on the morphological and anatomical evidence I have included some of these families in the dis- cussion on interrelationships of the Magnoliales in order to make the evaluation as complete as possible. EMBRYOLOGY AND CYTOLOGY Austrobaileyaceae. The family comprises a single genus, Austrobaile- ya, with two species distributed in Australia. Embryological information is meager and according to Davis (1966) is unknown; perhaps she over- looked the passing reference made by Bailey and Swamy (1949) concern- ing some points in the anther wall development and male gametophyte. The wall of the young microsporangium comprises the epidermis, three wall layers, and an irregularly two-layered glandular tapetum having binu- cleate cells. At maturity the epidermal cells become protuberant, the en- dothecial cells enlarge but it could not be made out from the illustrations whether these develop characteristic fibrous bands or not. The middle lay- ers and the tapetum become absorbed and only their remnants are per- ceptible. Cytokinesis is simultaneous. As in Degeneria the exine for- mation begins while the microspores are still in tetrads. The germinal furrow develops along the distal face where also the generative cell is. cut off. The mature pollen is spherical, monocolpate with colpus extend- ing from one pole to the other; the exine is finely pitted. The external 4 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 surface of the furrow possesses minute protuberances. Riidenberg (1967) has investigated the karyotype of Austrobaileya and found 2n = 44. The basic number in the family would be x = 22. Magnoliaceae. The Magnoliaceae are distributed primarly in eastern Asia, Malesia, southeastern North America, Central America, the West Indies, and Brazil. In India the family is represented by 22 species be- longing to 5 genera. Embryological literature concerns only Magnolia (Brandza, 1891; Earle, 1938a,b; Farr, 1918; Hayashi, 1960, 1964; Kapil & Bhandari, 1964), Michelia (Padmanabhan, 1960) and Liriodendron (Kaeiser & Boyce, 1962). The anther wall at the microspore mother cell stage comprises the epidermis, endothecium, 3 or 4 middle layers and bilayered glandular tapetum (Ficure 1 A). By the time the cytokinesis is completed in the microspore mother cells, a large number of Ubisch granules line the inner walls of the tapetum. In a fully mature anther the papillate epidermis and endothecium along with 2 or 3 middle layers persist (FicurE 1 B). In Magnolia stellata (Kapil & Bhandari, 1964) the endothecium possesses reticulate thickenings instead of the usual fibrous bands found in other members, After meiosis II in the microspore mother cells the cytokinesis takes place by furrowing (Ficurrs 1 C, D), resulting in tetrahedral or isobilateral tetrads (Ficure 1 E). The mature pollen is monocolpate (FicurE 1 G) and is shed at the 2-celled Stage (Ficure 1 F). The gen- erative cell is surrounded by a thin sheath of finely granular cytoplasm : i efly from the out. inner integument is represented by a layer of crus er integument while the In a ripe seed the testa consists of hed cells (Ficurr 1 M). an outer fleshy region comprising (1) 1971] BHANDARI, MAGNOLIALES > epidermis of the outer integument, (2) 2 or 3 layers of tangentially elon- gated cells, (3) a 10- to 12-layered fleshy zone, (4) 2 or 3 layers of tan- gentially compressed cells, and an inner stony region of 3 or 4 layers of lignified cells. eee 4 i 9 /y O ‘ iat \\ a aa. ES [7)> BER) OF GA Spo Ss 2 oS oe, OM (> 22 a SS ee Se ae SAS sSar ~ SERN FicuRE 1. Magnoliaceae. (ant, antipodal cells; ch, chalazal haustorium; eg, gg; emb, embryo; end, endosperm; ff, fleshy layer; hyp, hypostase; 1, inner id; ar zygote. aa EN aren ? ? endothecium. C-E, Stages in microsporogenesis; cytokinesis occurs by fur- rowing. F, Two-celled pollen. G, Palynogram. H, section of ovule show- ing parietal layers, sporogenous tissue and initiation of the two integuments. I, 1. Linear and T-shaped tetrads of megaspores ; anized embryo sac. L, 6 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 A number of stomata of the haplocheilic type develop on the epider- mis of the outer integument (see Paliwal & Bhandari, 1962). Cytological investigations (see Darlington & Wylie, 1955; Nanda, 1962; Raven, 1967) concern Liriodendron (2 spp.), Manglietia (2 spp.), Michelia (7 spp.), Pachylarnax phiocarpa, Talauma hodgsoni and T. phellocarpa, and Magnolia (39 spp.). The basic chromosome number is x = 19. Poly- ploidy is reported in Magnolia which includes 4 tetraploid, 11 hexaploid, 1 pentaploid and 1 triploid species. Janaki Ammal (1953) reported 2n = 95 in M. soulangeana whereas Nanda (1962) observed n = 38. If the identification of both the plants is dependable, then this is the only rec- ord of intraspecific cytological forms. Whatever the origin of basic num- ber, polyploidy seems to have played some role in speciation in Mag- nolia, whereas the chief differences in the genomes seems to have occurred at the gene level within various genera. of the generative Cc 2 H), whereas in and protrudes through 1971] BHANDARI, MAGNOLIALES 7 Hf @) Re fy. as he @9) oD I grain out of a tetrad germinating. (A, B, D, Dr & Venkataraman, 1968: C, Pseudowintera colorata, after Bhandari, 1963; E-G, Zygogynum baillonii, after Swamy, 1952; H. Pseudowintera axillaris, after Sampson, 1963.) in permanent tetrads at the 2-celled stage (Ficurr 2 I). In Zygogynum (Swamy, 1952) all the pollen grains germinate, while in Pseudowintera colorata only one germinates on the stigma; the rest degenerate (FIGURE fis The ovules are laminar, anatropous, bitegmic, and crassinucellate; the micropyle is formed by the inner integument alone (Ficure 3 A). A OL. 52 JOURNAL OF THE ARNOLD ARBORETUM [vou FIGURE 3. Winteraceae. (ant, antipodal cells: cc, chalazal chamber; ds, de- generating synergid ; €, epidermal cell; eg, egg: - D, Two-nucleate embryo sac. E, Mz Sac, note basal vacuole in the egg also, F , Two- ure em- micropyl - G-I, Stages in de celled endosperm with velopment of embryo; 1971] BHANDARI, MAGNOLIALES 9 single hypodermal archesporial cell divides periclinally to cut off a parietal cell which forms a massive parietal tissue (Ficure 3 B); the nucellar epidermis may also contribute to the parietal tissue (Ficures 3 B, C). A linear or T-shaped tetrad is produced at the end of reduction divisions in the megaspore mother cell (F1curE 3 C). The chalazal megaspore func- tions to form the Polygonum type of embryo sac (Ficures 3 C-E). In Pseudowintera the synergids have filiform apparatus; the polar nuclei fuse before fertilization and the three small uninucleate antipodal cells are ephemeral (Ficure 3 E). Triple fusion precedes syngamy and the endosperm is ad initio cellular. In Pseudowintera the first division may be transverse (FicuRE 3 F) or vertical. Both the chambers contribute towards the formation of endo- sperm tissue. The embryogeny has been worked out only in Drimys (Bhandari & Venkataraman, 1968) and it is irregular (Ficures 3 G-I), not conforming to any particular type of Johansen (1950). A suspensor is present but the embryo perhaps remains undifferentiated at the time of shedding of the seeds (Ficures 3 H, I). The seed coat is formed chiefly by the outer integument, and the inner becomes almost crushed and is represented by a thin strip of greatly stretched cells (FicurE 3 J). The cells of the epidermis of the outer in- tegument become elongated, thick walled, brittle, and extremely hard. The pericarp is about 14—16-layered and remains fleshy and parenchy- matous. In Pseudowintera meristematic plates of tissue develop in be- tween the maturing seeds so that the fruit becomes chambered. Numerous ethereal oil cells, tannin cells, and stone cells are interspersed in the tissue of the fruit-wall. The pericarp is differentiated into outer tough and coriaceous, and inner spongy, soft regions. The inner region develops numerous ingrowths between the seeds. Only two genera, Drimys and Pseudowintera (see Borgamann, 1964; Darlington & Wylie, 1955; Hotchkiss, 1955; Raven & Kyhos, 1965), are known cytologically. In Drimys (section TASMANIA) » = 13 or 14 has been reported in literature. However, Raven and Kyhos (1965) have shown that reports of 2n = 28 are erroneous since this number is ex- pressed because of a precocious disjunction in one of the bivalents making the configuration as m = 12 2,. Similarly, the court » = 38, for D. winteri (Whitaker, 1933) has been shown to be erroneous and as rein- vestigated by Raven and Kyhos (1965) it ism = 43. Drimys lanceolata occurs in two cytological races, the diploid with 2n = 26, and the triploid having 2n = 39; Pseudowintera colorata also shows n = 43. Degeneriaceae. The Degeneriaceae are monotypic and endemic to the Fiji Islands. The family has a number of primitive features such as the 10 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 broad laminar stamens (without any differentiation into filament, con- nective, and anther) having four microsporangia embedded on the abaxial surface; conduplicate carpel devoid of any style and stigma, and instead, having a decurrent stigmatic crest; laminar ovules; multilacunar node; and a primitive type of wood. Swamy (1949) has worked out the embryology, and Dahl and Rowley (1965) the ultrastructure of pollen of Degeneria vitiensis. In a young anther four groups of archesporium differentiate (FicurE 4 A). The wall of the anther on the abaxial surface comprises an epidermis, 1-layered fibrous endothecium, three or four middle layers, and a single-layered tapetum with binucleate cells (Ficures 4 B, C). At the time of dehiscence only the epidermis and endothecium persist; towards the connective re- gion some more layers develop thickenings like that of the endothecium. The cytokinesis is simultaneous and results in tetragonal tetrads (Fic- uRE 4D). The germinal furrow begins to develop on the distal end even when the microspores are in tetrads (Ficures 4 D, F). The generative cell is cut off towards the distal end (FicurE 4 E). The pollen grains are monocolpate and 2-celled at the shedding stage. There are nearly 30 to 32 ovules in a carpel. They are laminar, anat- ropous, bitegmic and crassinucellate. The micropyle is formed by the inner integument alone (Ficure 4 1); a funicular obturator is very con- spicuous and is comprised of elongated and densely cytoplasmic cells of the epidermis. The 1-celled archesporium is hypodermal and a massive tissue is formed by the primary parietal cell (FicuRE 4 H). A linear tetrad : lade eet: which the chalazal megaspore (Ficure 4 I) functions to e Polygonum type of : and the aneeetl 2 pe oa See 21) The syvereit a short suspensor, a bulbous Ss (Ficurgs 4 R, S) Recently Raven and K ry sue A hos (1965 family and that, therefor 65) reported m = 12 for the monotypic karyoeue » therefore, consti aoa tut : . as not been investigated, the basic haploid number. The 1971] BHANDARI, MAGNOLIALES 11 e, if? ie, SAD 3 CS a. @ imeem, wee Mm ] & J my ro ear . rite S: ‘7 as aie OP, SASGRES CARH CRASS Ty ye a nde, ‘ae, bee IGURE 4, Degeneriaceae. A, Part of transverse section of anther showing two groups of archesporium. B, Same, showing differentiation of wall layers. dja i perm ; ae - Tricotyledonous embryos showing suspensor, hypocotyl, and cotyledons. 1, ‘tudi ‘ Z ner stony re- gions, ruminate endosperm, and ll embryo. U, Part of seed coat enlarged from Ficure T. (A-T, Degeneria vitiensis, after Swamy, 1949.) 12 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 52 Annonaceae. The Annonaceae is the largest of all the woody magnolian families and comprises 120 genera with approximately 2,100 species. The family is cosmopolitan and well represented in tropical and subtropical to temperate regions of the world. (Periasamy & Swamy, 1959), or it is differentiated as a uniseriate row (Ficure 5 A) in Annona (Juliano, 1935), Cananga (Periasamy & Swamy, 1959), Asimina (Locke, In the latter group of plants some of these cells form sterile septae (Fic- URE 5 D) so that individual microspore mother cells or a group of them thia. U F ; Asana & Adatia, 1947), Polyal- Meta ey and Cananga (see Asana & Adatia, 1947). Cells of the r tetranucleate, or the four nuclei fuse to form paleieing v4: nucleus (Ficure § E). In Cananga odorata (Periasamy & (Herms, 1907), and Annona (Juliano, 1935) the , it may be regarded that tapetum is being : ble as the sporogenous tissues. In the ae wamy's (1959) observati c and Miliusa, the sen servations on Cananga an a genera reported to have amoeboid tapetum need a reinvesti- he process of cytokinesis is succes e tri, 1957) and Uvaria (Sastri, 1957 1 noliaceae) ; and the ex ip oi . (rarely with late (see Erdtman, 1952 ; _ J€ Ovules are anatr : ; ? sinucellate, The i trite Theument or by both, the midgn {8 formed by my & Swamy, tpel is condup| 1971] BHANDARI, MAGNOLIALES 13 Annonaceae. (e, epidermis; p, parietal layer; ¢t, tapetum; w, wall FIGURE 5, a layers.) A, Longitudinal section of a young anther locule showing uniseriate ae. | rat) oO wm uo) ° ri o& ee 7] Oo Le. 7 o =} wm o a on me ° =} mn ° Lena wn pil Oo os = aq wn cS OQ 2, oO wn wm _ < oO layers, tapetum and mi ores. G, H, Stages in cytokinesis by furrowing. I, Two-celled pol rains soon after division of microspore nucleus. J, Two- celled pollen; generative cell with granular cytoplasm surrounded by hyaline membrane, (A-J, Cananga odorata, after Periasamy & Swamy, 1959.) 14 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 zation by the traces from the dorsal bundles. A hypostase has been re- ported in Asimina and Artabotrys. An aril is present in some members (Corner, 1948). The hypodermal archesporium is 1- to few-celled (Ficure 6 A), and the massive parietal tissue is formed by the primary parietal cell (FicurE 6 B). At the end of meiosis I a dyad is formed (FicureE 6 B) which results in a linear tetrad, of which the chalazal megaspore functions (FicurE 6 C). The development of the female gametophyte conforms to the Polygonum type. The synergids are hooked and the antipodal cells remain uninucleate (FicurE 6 D). The synergids and 3 small antipodal cells degenerate immediately after fertilization. the octants form a later into a dicot em . 15 have been k : ; lington & Wylie, 1955: worked out cytologically (Dar- ? ) an et al., 1957; Miége, 1954, teear & Mangenot, 1957, 1958; Mangenot 8, 3 have x = 9 an ; A ede with 2n = 18, or triploid, 2” : 1S characterized by th i b : nd 9. Since a lar y three basic numbers, 8, it is likely that the Gihcy ee number of genera and species possess X = nera evolved by addition ion of me, or by a multiple of this base number Seaneie Myristica. ee species which aie ou isticaceae include 15 genera and nearly 250 in new world tropics, Africa, and Asia. Vas- BHANDARI, MAGNOLIALES SUP PY KP = SIS nav Ban C7) ee Vy se at \ e8ins fee) : otau J ee anae S c CIOS re TSS ce Cr \ A los OY, fe Me N hy Ay RARER ee. Ficure 6, Annonaceae. A, Longitudinal section of ovule showing hypodermal archesporial cell and origin of integuments. B, Transmedian longitudinal sec- 16 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 52 cular anatomy of the flower, aril, and embryology of only Myristica spp. has been worked out (Camp & Hubbard, 1963; Joshi, 1946; Nair & Bahl, 1956; Nair & Pillai, 1959; Periasamy, 1961; Sastri, 1955b). The anther wall is composed of epidermis, fibrous endothecium, two ephemeral middle layers, and a single-layered glandular tapetum compris- ing uninucleate cells. At the end of meiosis the successive type of cyto- Kinesis in the microspore mother cells results in isobilateral tetrads. The pollen grains are 2-celled and 1-sulcate. which gives rise to the The endosperm is n Only Myristica fragrans (Mangenot & Mangenot, 1957) have been worked out and these show = 21 and m = 19, respectively. The basic number is doubtful and may turn out to be x be finally said. Canellaceae. The family Canell : . 20 species dis y €laceae comprises six genera and tively from Venezuel hern Florida tropical America, Puerto Rico ela to souther , burgia stuhlmannii, The androecium of the with minute prot surfaces (Ficurr } only the persisting protuh. “ dehisced anther S aft . “ \). The microspore mother cells Rat oie wi 12 — b u \ middle non-functioning megaspore. K, A tetrad of megaspores. L, dhersbgreien embryo sac. M, Mature female gametophyte. (a3, os, M, Schisandra grandi- flora, after Kapil & Jalan, 1964; K, S. chinensis, after Swamy, 1964.) 20 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 has m = 13 whereas the latter possesses n — 14. They suggest that Whita- ker’s (1933) report of J. floridanum and I. anisatum (1. religiosum) as having » = 14 is, perhaps, incorrect. They explain the origin of nm = 13 in I, floridanum by aneuploidy from n = 14 of J. parviflorum. Schisandraceae. Two genera of woody vines, Schisandra and Kadsura, comprise the family Schisandraceae which is distributed in eastern and southeastern Asia and Malesia. Only Schisandra glabra is a native of America. In India there are four species found in the eastern and western Himalayas. In the Schisandraceae amongst many characters, the for- mation of a modified column by the basal connation of the filaments of the stamens, the conduplicate carpel with ventral stigmatic crest, and its distal extension to form a non-vascularized pseudostyle are of interest. The anther wall comprises a persistent epidermis, 1- or 2-layered fibrous endothecium, 2 or 3 middle ] The piles ace e pollen : one or both t Osiphonous, ferti lization js d orogamous an ome crushed durj tt Ing this process, The endo- 1971] BHANDARI, MAGNOLIALES 21 sperm is cellular. The first division is transverse and results in a large micropylar and a small chalazal chamber (Ficure 9 C). Further growth in both the chambers results in a massive endosperm tissue which fills almost the entire cavity of a mature seed (Ficures 9 D, E). The divi- sion of the zygote is either transverse or vertical (Ficurrs 9 F, G). The derivatives of ca form the octant having 4 cells in tier 1’ and 1 (Ficures 9 H-J). Embryogeny follows the Asterad or Onagrad type. The mature embryo is dicotyledonous (FicurE 9 E). The seed coat is contributed by both the integuments, the outer integ- ument comprising an epidermis of macrosclereids, 2 or 3 subepidermal layers of brachysclereids, followed by 2 or 3 layers of parenchymatous tissue; whereas the inner integument persists as a thin 2-layered tissue of thick-walled cells (Ficure 9 K). The fruit wall is fleshy and succulent and comprises 14 to 16 layers of enlarged parenchymatous cells. There are numerous ethereal oil cells interspersed in the outer epidermis of the pericarp. The two genera, Schisandra and Kadsura, of the family Schisandraceae possess nm = 14 (Darlington & Wylie, 1955; Stone, 1968). The basic number was thought by Darlington and Wylie (1955) to be x = 7. According to Stone (1968) the karyotype is different from that of Illiciaceae in being nearly symmetrical and showing an absence of subterminal chromosomes. Monimiaceae. In the Monimiaceae there are about 34 genera and 450 species met with in tropical and subtropical regions with distributional cen- ters in Australia, Polynesia, Madagascar, tropical Africa, South America, and Mexico. Of the larger number of genera and species, the embryology of only Peumus boldus (Mauritzon, 1935) has been worked out in detail, in addi- tion to some fragmentary information that exists on Siparuna eggersii (Heilborn, 1931) and Mollinedia (Peter, 1920). The anther wall comprises epidermis, fibrous endothecium, 2 or 3 middle layers, and a 1-layered glan- dular (periplasmodial in Atherosperma, see Sastri, 1963) tapetum. At the end of meiosis, tetrahedral tetrads of microspores are formed by si- multaneous cytokinesis. The pollen grains are shed individually or they remain in permanent tetrads in Hedycarya angustifolia (Money, Bailey, & Swamy, 1950). They are mono-, bi-, or acolpate and are shed at the two-celled stage. The ovules are anatropous unitegmic (Siparuna) or bitegmic (Peumus boldus) and crassinucellate. The micropyle in Peumus is formed by both the integuments. The nucellar epidermis at the tip undergoes periclinal divisions (Ficure 10 F) so that a 4~6-layered nucellar cap is formed. At the mature embryo sac stage a conspicuous hypostase is present. The fe- male archesporium is multicelled in Siparuna but is one-celled in Peumus (Ficurrs 10 A, F). The archesporial cell forms the primary parietal cell which divides repeatedly to result in a massive parietal tissue so that the megaspore mother cell is deeply buried in the nucellus (Ficures 10 A, F). J 4 [ OL. 52 oH 4 ~~ Sls FicuRE 9, Schisan cell with two oO. , ; ges in de- wing macrosclereids, brachy- » Schisandra chj 964.) F-J, Stage is f testa. (A A, B inensis, after Swamy, er Kapil & Jalan, 1 1971] BHANDARI, MAGNOLIALES 23 After meiotic division in the megaspore mother cell a linear tetrad is formed in Siparuna and Mollinedia (Ficure 10 A). In Siparuna the chalazal megaspore functions and forms a long, tortuous chalazal tube with a vesicular tip (FicurEes 10 A, B). The tip of this tube bursts into another cell and discharges its contents (FicurE 10 C). It forms a 4- or 5-nucleate female gametophyte that remains unorganized in Siparuna. However, in Mollinedia a normal embryo sac is formed. On the other hand, in Peumus, cytokinesis is suppressed after meiosis II, and subsequently the upper of the two dyad cells degenerates (FIGURE 10 G), whereas the lower func- tions to form the embryo sac (FicurE 10 H). The development of the embryo sac is, therefore, of the Polygonum and Allium types. An antip- odal complex of 5 to 20 cells is observed in P. boldus (Ficure 10 I, J). The development of endosperm is cellular. The first division is trans- verse which divides the embryo sac into a large micropylar and a much smaller chalazal chamber. The next division in the chalazal chamber is vertical and is followed by a few transverse divisions in the micropylar chamber (Ficure 10 R). By repeated divisions both the chalazal and micropylar chambers form a massive endosperm tissue (FicuRE 10 S). The zygote divides transversely (F1cuRE 10 K) to form the cells ca and cb. Subsequently these two cells undergo either transverse or vertical divi- sion (Ficures 10 L, M) or the cell ca may divide transversely whereas the cell cb divides vertically (FicurE 10 N), indicating that there is a lot of plasticity in the pattern of divisions during early embryogeny. Later however, the derivatives of the ca form the embryo proper while those of the cb form a short suspensor (FicuRE 10 Q, R). In the mature seed there is a small embryo embedded in a copious endosperm. According to Johanson (1950), the embryogeny conforms to the Asterad type. t must be realized that the foregoing embryological account is based on the information available for only two members out of a large number of taxa. Rather much variation can be expected in the embryological fea- tures matching well with the extensive diversification in vegetative and reproductive organs of taxa included in this family. Chromosome counts of only two members, Laurelia novaezelandiae and Kibara sp. (Borgmann, 1964) have been made. Both genera show n = 22 but no karyotypic studies are available. Calycanthaceae. The Calycanthaceae comprise two genera and about nine species distributed in the southeastern United States, Australia, China, and Japan. The embryological literature has been summarized by Schnarf (1931). Recently Mathur (1969) has studied the development of male and female gametophytes in Calycanthus. The anther wall comprises the epidermis, fibrous endothecium, and glandular or periplasmodial tapetum. Meiosis in the microspore mother cell is abnormal. Cytokinesis is simultaneous re- Sulting in tetrahedral or isobilateral tetrads. The mature pollen grains are 2-celled, rarely in tetrads in Chimonanthus fragrans, sterile and 2- nucleate. The exine is reticulate. 24 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 FIGURE 10. Monin; lazal megaspore of et:_Ay Nucellus wi * ; re ? US W. ing in 2 irk a central tetrad has grown si pe ops ene pes o-sac tube . S, One from a mj Ove the hypostase. B Tubes from Embryo sac Plasma in Pr ao .mMegaspore — note nuclei still in tubes. T cavity, stil] attached to mouth of embryo 1971] BHANDARI, MAGNOLIALES 25 The ovules are anatropous, bitegmic, and crassinucellate; the micropyle is formed by the inner integument alone. The cells of the nucellar epi- dermis divide periclinally to form a massive nucellar cap. The female archesporium is multicelled but parietal tissue is absent. Numerous mega- spore mother cells function to form the linear tetrads, of which the chalazal megaspore functions to give rise to the Polygonum type of female gameto- phyte. Multiple embryo sacs are common. In the mature embryo sac the egg apparatus, one polar nucleus and the antipodal nuclei degenerate; syngamy and triple fusion therefore do not occur. The development of the endosperm is autonomous and is ab initio cellu- lar. Only nucellar embryos develop and polyembryony is frequent in Calycanthus occidentalis. The two genera, Calycanthus and Chimonanthus, possess m = 11. The karyotype has not been investigated. Lauraceae. The Lauraceae comprise about 31 genera and 2,250 species which are distributed in the tropics and warm-temperate areas of both hemispheres, especially Central and South America and southern Asia. Of these, the embryology has been investigated only in Cassytha (Sastri, 1956, 1962), Cinnamomum (Chowdhury & Mitra, 1953; Giuliani, 1925; Sastri, 1958), Laurus (Battaglia, 1947; Mezzetti-Bambacioni, 1935, 1938, 1941), Litsea (Sastri, 1958), Persea (Schroeder, 1942, 1952), Sassafras (Coy, 1928), and Umbellularia (Mezzetti-Bambacioni, 1941). The archesporium is multicelled and the wall layers are formed by the primary parietal layer (Ficure 11 A). The anther wall comprises the persistent epidermis, fibrous or thick-walled endothecium, 2 middle layers, and amoeboid (Cinnamomum, Laurus, Litsea, Persea, Sassafras, Umbel- lularia) or glandular (Cassytha) tapetum (Ficures 11 B—D). The tape- tum is of parietal origin and its cells are bi-, tetra-, or multinucleate (Fic- URES 11 B, C). Successive cytokinesis results in tetrahedral, isobilateral, T-shaped, or linear tetrads (Ficures 11 E-G). The pollen grains are 2- celled, monocolpate or acolpate (Cassytha, Laurus). The exine is minutely spinescent. The ovules are anatropous, bitegmic and crassinucellate. The micro- pyle is formed by both the integuments (Ficure 12 B); in Cassytha (Sastri, 1962), however, the integuments do not grow beyond the nucellus So that the micropyle is broad and the overarching funiculus is in direct contact with the nucellus (FicurE 11 K). The female archesporium is 1-celled (Laurus, Umbellularia) or multicelled as in Cassytha and Cin- “amomum (Ficures 11 H, J; 12 A). A massive parietal tissue is formed either from the parietal cell alone or by it and the nucellar epidermis wi dyad cell degeneratin K- S i bryogenesis. F g. K-Q, Stages in embryog » Oe velopment of endosperm. (A-E, Siparuna eggersii, after Heilborn, 1931; F-S, Peumus boldus, after Mauritzon, 1935.) Oi The JOURNAL OF THE ARNOLD ARBORETUM [Vv 26 ae. Longisection of layer and Sporogenous cells. ye Transection ing wall layers and tapet D showing megaspore mother cell. " ylar two degenerated, the chalazal one at 2-nucleate Stage. J, Longitudinal sec howing elongation of numerous em- 1971] BHANDARI, MAGNOLIALES 27 (Ficures 11 H, J; 12 A, B). Numerous megaspore mother cells func- tion (Ficures 11 J; 12 A) in Cassytha (Sastri, 1962), and Litsea (Sastri, 1958). The chalazal megaspore of the linear tetrad functions (FIGURE 11 I) and gives rise to the Polygonum type of female gametophyte. Mul- tiple embryo sacs are formed in Cassytha (Ficures 11 J, K), although occasionally twin embryo sacs have been observed in Persea americana (Schroeder, 1952) also. In Cassytha filiformis tips of 4 to 6 embryo sacs elongate, bore through the nucellus, and finally lodge themselves in the funiculus or inner integument (FicurE 11 K). The synergids degenerate soon after fertilization but in Litsea iners they persist for some time. The antipodal cells (Cinnamomum, Sassafras, Umbellularia) or nuclei (Cassy- tha filiformis) are ephemeral (Ficure 11 K) but in Laurus nobilis they persist, divide, and form an antipodal complex (Mezzetti-Bambacioni, 1935). In Cassytha filiformis the development of the endosporm is ab initio cellular (Ficures 12 D, E), whereas in all other members it is nuclear (FicurE 12 C). The cell formation in the nuclear endosperm takes place at 4-celled or globular stages of the embryo. In Cassytha the first and the next transverse division in the chalazal chamber results in 3-celled endosperm (FicurE 12 D). The middle cell then undergoes a vertical division. Divisions in all planes result in a massive tissue (FicuRE 12 E) which comes out of the nucellus at its tip and plugs the micropyle. Haus- torial structures commonly found in the parasitic angiosperms are not met with in Cassytha. The development of the embryo conforms to the Piperad (Sassafras), Asterad (Persea americana), or Onagrad (Cinnamomum iners and Litsea sebifera) types. Orientation of walls at the four-celled stage is variable (Ficures 12 F-I), The mature embryo has a well-developed radicle and a shoot apex which develops two small leaf primordia oriented at right angles to the two cotyledons (Ficure 12 J). In Cinnamomum, Litsea, and Cassytha the seed coat is formed by the outer integument alone and comprises an epidermis of radially elongated cells, a few parenchymatous layers, and an inner epidermis of elongated cells having helical thickenings (FicurE 12 L). The pericarp comprises an epidermis of tannin-filled cells, a fleshy zone of 7 to 9 layers of paren- chymatous cells, a single layer of radially elongated cells, and an inner epidermis of lignified and pitted columnar cells. In Cinnamomum and Litsea an additional 4- or 5-layered zone of stone cells is also present be- heath the outer tanniniferous epidermis. Nearly seven genera out of 31 included in the family have been worked Out cytologically (Chung et al., 1963; Darlington & Wylie, 1955; Hair & euzenberg, 1960; Mangenot & Mangenot, 1957, 1958; Sharma & Bhatt, ag; Suzuka, 1953). The genera Beilschmiedia (3 spp.), Cinnamomum bryo sacs K, Longitudi i i ture embryo sacs, two Sacs. K, Longitudinal section of ovule showing ma : ~ ib abutting the lysigenous cavity. (A, B, E-G, I-K, Cassytha filiformis; C, Cim "amomum iners: D, E, C. zeylanicum, after Sastri, 1958, 1962. VOL, 52 JOURNAL OF THE ARNOLD ARBORETUM [ 28 Fae Se SS) : A —— = = == aay ‘ wee embryo sac iageturtae A B, Longitudinal Section of apical reg} f ovule sh i Longitudinal Section of win ry C, two-celled perm nuclei and proembryo, D . F-J, Stages in —, genesis. K, Mature Tyo in Paracotyledonary view s owing primor first two leaves. L, i Longitudinal Section of pa momum iners: D-L, Cassy (9 spp.), Lindera (4 spp.), Perseq (4 spp.) oF lularia (1 sp.) all Possess m = 17. Only Laurus shows 2n = 36 in L. Canariensis and 2n — 42, 48 in L. nobilis. The basic number of the fam- ily perhaps may be x — 42. » Sassafras (1 sp.), and Umbel- 1971] BHANDARI, MAGNOLIALES 29 Trochodendraceae. Trochodendraceae is a monotypic family com- prising Trochodendron aralioides distributed in Japan and Formosa. Bailey and Nast (1945), and Nast and Bailey (1945) have studied the vegetative anatomy and floral morphology of T. aralioides, whereas Yoffe (1962, 1965) has investigated its embryology. At the microspore-mother-cell stage the anther wall consists of an epi- dermis, the endothecium, two or three middle layers, and a two-layered glandular tapetum having binucleate cells (FicurE 13 A). During matura- tion the epidermal cells of the anther wall become protuberant and de- velop cuticular fibrillar projections along their outer surface (FIcuRE 13 B). The endothecium becomes very prominent by developing broad bands of fibrous thickenings; the middle layers become flattened and crushed. The tapetum, however, persists during the complete meiotic process in the microspore mother cells until the formation of two-celled pollen grains (FicurE 13 B). At this stage the tapetal layer adjacent to the middle layers degenerates, whereas the one next to it develops Ubisch granules along the lateral and inner walls of the tapetal cells (Ficure 13 B). Meiosis is normal but occasionally, a cell plate is formed during telophase I which, however, disappears later. The cytokinesis is of the simultaneous type and the resulting tetrads are tetrahedral. The nucleus of the micro- Spore moves towards one side where it divides to form the generative and the vegetative cells. The mature pollen grains are 2-celled and tricolpate having a reticulate exine. The ovules are anatropous, bitegmic, and crassinucellate (FicurE 13 C). The inner integument is initiated first, followed by the outer, but the micro- pyle is formed by the inner integument alone (Ficure 13 C). At the ma- ture embryo-sac stage, the chalazal end of the ovule is transformed into a well-developed long, tapering projection. The vascular supply of the ovule enters this projection and curves backwards to end at the base of the chalazal region (Ficurr 13 C), : A single hypodermal archesporial cell cuts off a parietal cell which divides repeatedly to form a 3- or 4-layered parietal tissue. After meiosis I a dyad is formed which subsequently develops into either a T-shaped or linear tetrad. The chalazal megaspore functions (FicuRE 13 D) to give rise to a 2-nucleate, 4-nucleate (FicurE 13 E) and 8-nucleate embryo sac of the Polygonum type (Ficurr 13 F). The synergids possess a con- spicuous filiform apparatus; the antipodal cells are uninucleate and per- sistent (Figure 13 F). Syngamy and triple fusion occur simultaneously. The development of the endosperm is ab initio cellular. The first division is transverse giving tise to a large micropylar and very small chalazal chamber (FIGURE 13 G). The next division in the micropylar chamber is transverse, while in the chalazal chamber it is vertical (FIGURE 13 H). Repeated divisions in both the chambers result in a massive endosperm tissue. However, the cells formed in the chalazal chamber are small and densely cytoplasmic. ae major part of the endosperm tissue is formed by the micropylar chamber (Ficurr 13 [), VOL. 52 JOURNAL OF THE ARNOLD ARBORETUM [ 30 oe CK: Tce, A eee of@ Pe Ox S883. seid FicureE 13, Trochodendraceae. A, Transection of part of anther lobe at micro- Spore-mother-ce]] Stage. B, Same, at ti i rmis. C, i Sac stage, note circumscutous vascular three degenerated and chalazal m bryo sac. F, Mature embryo sac. G-I, Stages in development of endosperm. J-Q, Stages in development of a dicot embryo, A-Q, after Voffe, 1962, 1965.) 1965.) 1971] BHANDARI, MAGNOLIALES 31 The zygote divides transversely forming the two cells, ca and cb (Fic- uRE 13 J). Subsequent division in both these cells is transverse thus re- sulting in a 4-celled linear proembryo (Ficure 13 K). The terminal cell of this proembryo divides vertically followed by another similar division at right angles to the first. The four cells thus formed divide transversely to form the octant disposed in 2 tiers of 4 cells each (FicurE 13 L). Re- peated divisions in this octant give rise to a globular proembryo (FiGuRE 13 N, O). By this time the other 3 cells divide transversely and longitudin- ally and produce a suspensor (F1icurE 13 N). The globular proembryo dif- ferentiates first into a heart-shaped and later into a dicotyledonous em- bryo (Ficure 13 P). At this stage the procambial strand is well developed (Ficure 13 Q). The seed coat is formed by both the integuments. In the mature seed the inner epidermis of the outer integument becomes thick-walled and sclerenchymatous to constitute the hard part of the seed coat whereas the other layers of the two integuments remain thin-walled. In the ma- ture seeds the tapering chalazal projection becomes transformed into the wing. The basic number of this monotypic family also is m = 19 in Trocho- dendron aralioides (Whitaker, 1933). Cercidiphyllaceae. The genus Cercidiphyllum, which includes two species, C. japonicum and C. magnificum, constitutes the family Cercidi- phyllaceae having a restricted distribution in Japan and China. Swamy and Bailey (1949) have investigated the morphology, anatomy, and embryology of Cercidiphyllum japonicum. The structure and develop- ment of the anther resembles the majority of dicotyledons. The tapetum 1s, however, glandular with binucleate cells. The process of cytokinesis after reduction divisions is of the simultaneous type resulting in microspore tet- tads. The pollen grains at the shedding stage are 2-celled, spherical, and tricolpate; the colpi are conspicuously broad. The exine is finely pitted. The ovules are hemianatropous, bitegmic, and crassinucellate. The two integuments are initiated almost at the same time; the outer integument soon overgrows the inner, although in a mature ovule both the integuments take part in the formation of micropyle. The chalazal end of the ovule develops into a flattened tapering projection which forms a wing in the mature seed. After fertilization the vascular bundle reaches the middle of the tapering chalazal projection and then bends to end at the base of the nucellus. ee The female archesporium is single-celled, forming five or six parietal cells (Ficurr 14 A). Periclinal divisions in the nucellar epidermis add to the formation of a massive parietal tissue, resulting in a deeply seated megaspore mother cell in the nucellus (Ficure 14 A). The megasporé Mother cell undergoes meiosis thus giving rise to a linear tetrad. pie micropylar megaspores degenerate while the chalazal one functions (Ficurr 14 A) to form an eight-nucleate embryo sac of the eran ot type (Ficurr 14 B). The egg apparatus is conspicuous, three antipoda 32 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 cells are uninucleate while the two polar nuclei fuse before fertilization (Ficure 14 B The primary endosperm nucleus moves to the chalazal part of the em- bryo sac where it divides, followed by formation of a transverse wall result- ing in a large micropylar and a small chalazal chamber (FicurE 14 C). Subsequent divisions in the two chambers are essentially transverse until a 10- to 12-celled uniseriate endosperm is formed (Ficure 12 D). Dur- ing this growth period the embryo sac enlarges considerably and comes to be in close contact with the vascular strand (Ficure 14 F). Further divi- sions occur in all planes finally producing a compact mass of endosperm tis- sue (FicurE 14 E), most of which, however, is consumed by the develop- ing embryo. tn ae gument alone. In the region €comes fattened so that the chalazal wi to ex- tend beyond it at either side. Rica anne Of the two species of C ercidiphy been made for C. japonicum, wh karyotypic analysis js available. illum included, chromosome counts have ich is m = 19 (Whitaker, 1933). No morphology and anatomy of the ranalian ee and their further phylo- tk is exhausti i i notte Mapai: —. Stive and voluminous and will S : ? ) , 1960), Degeneriaceae ( wamy, 1949), Himantandraceae (Bailey, Nast & ont 1943), Wett- Canright : ? ? ; Vestal, 1935, Annonaceae (Wyk & iiaaas (aa a b: i Eupomatiaceae (Lemesle, oma. A giias i 48 - Jal Bahl, 1956), Canellaceae, Schisandra- eae rong 68; Smith, 1947), Illiciaceae (Bailey Swamy, 1949) | Trimeniaceae. rings flnage LU Austrobaileyaceae (Bailey & » Amborell nae : Swamy, 1948. |e : aceae, and Monimiaceae (Bailey & amy, 1948; y, Bail y Bailey, 1957; Smith, 1928), Gomaneee re (Algescanthaceae (Fahn & 1971] BHANDARI, MAGNOLIALES 33 Dg Been Se Be * r\ ey ee Cu aN at aoe ew, gee ss an oy, YZ Co CS ‘pe iJ e 69) Smbryo proper. (A-L, after Swamy & Bailey, 1949.) 1950), Lauraceae (Coy, 1928; Sastri, 1952; Stern, 1954), Hernandiaceae ite, 1960), Tetracentraceae and Trochodendraceae (Bailey & Nast, “ed Croizat, 1947; Nast & Bailey, 1945; Smith, 1947), and Cercidiphyl- oe (Swamy & Bailey, 1949). Eames (1961) has given an excellent 34 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 review on Ranales which outlines in brief, the principal trends of spe- cializations in the group. INTERRELATIONSHIPS General considerations. ANTHER WALL. In almost all the families the basic architecture of the structure of the anther wall is uniform and comprises the epidermis, fibrous endothecium, 2 to 4 middle layers, and tapetum. The epidermis is persistent till the dehiscence of the anther; Trochodendraceae stands apart from the rest of the families in having prominent fibrillar cuticular thickenings on the outer surface. The fibrous endothecium (excepting Magnolia stellata; Kapil & Bhandari, 1964) is persistent and mostly 1-layered. However, in Magnoliaceae (Hayashi, 1960; Kapil & Bhandari, 1964; Padmanabhan, 1960), Degeneriaceae (Swamy, 1949), Winteraceae (Bhandari, 1963; Bhandari & Venkatara- man, 1968; Swamy, 1949), and Schisandraceae (Kapil & Jalan, 1964), there is a tendency for it to become irregularly 2-layered along the outer face of the anther wall while some layers of the connective tissue toward the inner side also develop fibrous thickenings and simulate the endo- thecium. Its maximum amplification is found in the relatively more primi- tive families Winteraceae and Degeneriaceae, and perhaps indicates that it Is a primitive feature and that the phylogenetic trend proceeds toward the reduction of the endothecial tissue. The middle layers vary from 2 to 4 in number and again in Magnolia- ceae, Degeneriaceae, Winteraceae, Annonaceae, Illiciaceae, and Canella- sili : ae bik persist even at the time of dehiscence of the anther, ne ast a cpp in the number of the middle layers in Myristi- eral and "oo 5 ime oe ay ‘on Trochodendraceae. They are ephe wr caceae, Lauraceae, pletely crushed during and after meiosis in Myristi- and Trochodendraceae. ANTHER TAPETUM. The tera, and Zygo eg Synum, dari, 1963c) p ~ » Pseudowintera colorata (Bhan- oid tapetum. The other species P, axillaris 1971] BHANDARI, MAGNOLIALES 35 (Sampson, 1963) has glandular tapetum. Furthermore, an earlier report of amoeboid tapetum in Drimys (Kutti Amma, 1938) has been refuted re- cently by Bhandari and Venkataraman (1968). Therefore, Lauraceae is the only family where amoeboid tapetum can be regarded as a family character. Glandular tapetum, therefore, reflects a primitive pattern whereas the amoeboid tapetum is a phylogenetic departure. In the Ranunculales, the herbaceous Ranales, a parallel trend can also be observed, although the amoeboid tapetum occurs more frequently in this group. In Magnoliaceae and Trochodendraceae the glandular tapetum is bi- layered while in Schisandraceae and Illiciaceae it becomes irregularly bi- layered. In the rest of the families it is uniformly 1-layered. The pres- ence of Ubisch granules binds the Magnoliaceae, Trochodendraceae, and Canellaceae together. In the last family, however, the granules are stud- ded along the inner as well as outer faces of the tapetal walls. The cells of the tapetum are uninucleate in Myristicaceae, binucleate in Magnolia- ceae, Degeneriaceae, Annonaceae, Illiciaceae, Trochodendraceae, and Cerci- diphyllaceae, and multinucleate in Magnoliaceae, Annonaceae, Canellaceae, Schisandraceae, Monimiaceae, and Lauraceae with a tendency to form polyploid nuclei by fusion. ORIGIN OF TAPETUM. The sporogenous origin of the tapetum has been looked at suspiciously and this was particularly true in some instances since a reinvestigation proved its parietal origin at least along the outer (pro- tuberant) face of the anther locule (see Maheshwari, 1950). Recently, Swamy and Periasamy (1966) argued that the tapetum is of dual origin, contributed by parietal tissue along the outer side, and by the connective or septal tissue along the inner side of the locule (see also Coulter & Cham- berlain, 1905). This conclusion has been further supported by onto- genetic studies of anther wall development in Anemone rivularis (Bhan- dari, 1968), Bhandari (1968) has suggested a reappraisal of the sporog- enous origin of tapetum and it may be in some instances that this tissue does change its function from reproductive to nutritional as is the case in most of the gymnosperms. Such a situation has been observed in Annona- ceae. In Annona (Juliano, 1935), Cananga odorata, and Miliusa wightiana (Periasamy & Swamy, 1961) the microspore mother cells are separated by septae which originate from the sporogenous tissue and behave like the tapetum developing from the parietal tissue. In such members of Annona- ceae, therefore, the tapetum is contributed partly by the sporogenous tis- sue and partly by the parietal tissue. Hayashi (1960) has reported sporog- enous origin of the tapetum in Magnolia liliflora and M. virginiana. The Magnoliaceae and Annonaceae thus share this feature. CYTOKINESIS AND MICROSPORE TETRADS. Cytokinesis in microspore mother cells at the end of the first or second meiotic division is another im- portant character. It is simultaneous in Austrobaileyaceae, Magnoliaceae, Degeneriaceae, Winteraceae, Annonaceae (some members), Canellaceae, Schisandraceae, Illiciaceae, Monimiaceae, Calycanthaceae Trochodendra- ceae, and Cercidiphyllaceae, whereas in Annonaceae (some members), 36 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Myristicaceae, and Lauraceae it is of the successive type. Sastri (1957) is of the opinion that in Annonaceae a series exists from the successive type of cytokinesis in Annona reticulata, Asimina triloba, and Uvaria kirkii to the simultaneous type in Saccopetalum tomentosum and Polyalthia, through intermediate types like Annona cherimolia and A. squamosa where the first constriction, although initiated at the end of heterotypic division, is delayed, and completed only along with the other after the homotypic division, A similar situation is also met with in Magnolia (Kapil & Bhan- dari, 1964, see also Farr, 1918), Zygogynum baillonii (Swamy, 1952) of the Winteraceae, and Trochodendron (Yoffe, 1962). Periasamy and Swamy (1959) on the other hand, consider the successive type of cytoki- nesis of the Annonaceae to be of modified simultaneous type. The pres- ent author is not in agreement with Periasamy and Swamy (1959) and considers that if a dyad is formed at the end of the heterotypic division as in Asimina and Uvaria, it should be regarded as the successive type irrespective of its formation, by cell plate or by furrowing. The prepon- derance of the simultaneous type of cytokinesis in the majority of the Mag- nolian taxa and the occurrence of the intermediate forms where the first furrow is initiated after heterotypic division, together with the formation of an evanescent cell plate in primitive genera like Zygogynum bailloni (Swamy, 1952), Pseudowintera axillaris (Sampson, 1963), Drimys win- pie ye ae & Venkataraman, 1968), and Magnolia (Kapil & Bhan- , ) might indicate that the successive type is a derived condition (see also Maheshwari, 1950). cae wi aii tetrads are formed in Canellaceae, Illi- mete 40: er eers , * rocho endraceae, and tetrahedral or decussate gnoliaceae, Winteraceae, Schisandraceae, Calycantha- pee and Cercidiphyllaceae. Tetragonal tetrads are found in Degeneriaceae and Annonaceae, while in Myristicaceae onl In the Lauraceae, tetrahedral n Degeneriaceae and Canellaceae, celled within the tetrads. S, bitegmic, and crassinucellate in 1971] BHANDARI, MAGNOLIALES 37 Magnoliaceae, Degeneriaceae, Winteraceae, Annonaceae, Myristicaceae, Schisandraceae, Illiciaceae, Monimiaceae, Calycanthaceae, Lauraceae, Tro- chodendraceae, and Cercidiphyllaceae. In some members of Annonaceae, however, the ovules are tritegmic and in Canellaceae they may also be hemianatropous. In Siparuna of the Monimiaceae the ovules are unitegmic. In the Magnoliaceae, Degeneriaceae, and Annonaceae the outer integument is vascularized whereas in others, the funicular supply ends at the base of the nucellus. The micropyle is formed by both the integuments in Mag- noliaceae, Annonaceae, Canellaceae, Schisandraceae, Illiciaceae, Lauraceae, and Cercidiphyllaceae while in Degeneriaceae, Winteraceae, Myristicaceae, Monimiaceae, Calycanthaceae, and Trochodendraceae it is formed by the inner integument. The micropyle is zigzag in Canellaceae, whereas it is straight in the other families. The family Trochodendraceae (Yoffe, 1965) and Cercidiphyllaceae (Swamy & Bailey, 1949) resemble each other in having a chalazal tapering projection; the funicular vascular supply enters the projection for some distance and then bends back to end at the base of the nucellus. This projection forms a wing-like structure in the seed. The massive parietal tissue is formed by the parietal cell alone in Mag- noliaceae, Degeneriaceae, Winteraceae, Annonaceae, Schisandraceae, IIli- ciaceae, and Trochodendraceae. In Monimiaceae and Cercidiphyllaceae it develops from both the primary parietal cell and the nucellar epidermis, while in Calycanthaceae it originates from nucellar epidermis alone. The family Lauraceae, however, exhibits both the trends where the parietal tis- sue is formed by the parietal cell alone or together with the nucellar epi- dermis. Although the relic feature of producing a massive parietal tissue is retained, the evident phylogenetic trend is towards the suppression of the formation of this tissue from the primary parietal cell, a trend which becomes well established in some of the herbaceous ranalian families such as Ranunculaceae (some members), Berberidaceae, Lardizabalaceae, and some members of Menispermaceae (see Bhandari, 1962, 1963b, 1965, 1968; Bhatnagar, 1965; Johri, 1936; Joshi, 1939; Sastri, 1964; Swamy, 1953). A prominent nucellar cap is met with in Monimiaceae, Calycantha- ceae, and Lauraceae (Heilborn, 1931; Mauritzon, 1935; Sastri, 1963; Schnarf, 1931). A conspicuous hypostase is met with in Magnoliaceae (Kapil & Bhandari, 1964), Degeneriaceae (Swamy, 1949), Annonaceae (Corner, 1949), and Monimiaceae (Mauritzon, 1935). MEGASPOROGENESIS AND FEMALE GAMETOPHYTE. The female arche- sporium is hypodermal, 1-celled in the families Degeneriaceae (Swamy, 1949), Winteraceae (Bhandari, 1963; Bhandari & Venkataraman, 1968; Sampson, 1963; Swamy, 1952), Myristicaceae (Schnarf, 1931), Canella- ceae (Parameswaran, 1962), Illiciaceae (Yoshida, 1962), T rochoden- draceae (Yoffe, 1965), and Cercidiphyllaceae (Swamy & Bailey, 1949) but multicelled in Schisandraceae (Kapil & Jalan, 1964; Yoshida, 1962), Calycanthaceae (Schnarf, 1931), and Lauraceae (Sastri, 1963). In the Magnoliaceae (Kapil & Bhandari, 1964; Padmanabhan, 1960), Annona- ceae (Parulekar, 1967), and Monimiaceae (Heilborn, 1931; Mauritzon, 1935) the archesporium may be 1- to multicelled. The tendency, however, 38 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 is towards the reduction in the number of archesporial cells as is evident in Magnoliaceae (Kapil & Bhandari, 1964) where, although numerous arche- sporial cells differentiate, ultimately only one functions. Only in the Lauraceae and Calycanthaceae do large numbers of megaspore mother cells function simultaneously, these have perhaps retained the ancestral feature of forming functional massive sporogenous tissue. In Siparuna of the Monimiaceae (Heilborn, 1931) the ovules are unitegmic but have a multicelled archesporium, a combination of advanced and primitive fea- tures. In almost all the families the development of the female gametophyte conforms to the Polygonum type. In Schisandra grandiflora (Kapil & Jalan, 1964) and Schisandra chinensis (Yoshida, 1962), the Polygonum type of female gametophytes has been reported, whereas Swamy (1964) observed the Oenothera type, and a unique bisporic type (organizing after 4-nucleate stage) in S. chinensis. The Allium type of embryo sac has also been recorded in Peumus boldus (Mauritzon, 1936) of the Monimia- ceae. It is evident, therefore, that although wall formation takes place after meiosis I and II, resulting in a tetrad of megaspores in a majority of the Magnolian members, there is a tendency towards suppression of the cytokinesis after homotypic division, hence towards a bisporic type. The synergids and the antipodal cells or nuclei (C assytha, Sastri, 1962) are characteristically ephemeral excepting in Peumus where secondary multiplication occurs to form the antipodal cell complex. The polar nuclei fuse before fertilization, which is met with in Magnolia (Kapil & Bhandari, 1964) and Drimys winteri (Bhandari & Venkataram Ent perm is ab initio cellular in the Magnoliaceae (Kapil & Bhandari, 1964; Padmanabhan, 1960), Degeneriaceae (Swamy, 1949), Winteraceae (Bhandari, 1963: Bhandari & Venkataraman, 1968), Annonaceae (Perias y & Swamy, 1958), Schisandraceae (Kapil & Jalan, 1964), Mliciaceae (Hayashi, 1963), Monimiaceae (Mauritzon, 1935), , 1931), Cassytha (Sastri, 1962), Trochodendra- ceae (Yolfe, 1965), and Cercidiphyllaceae : ly in Myristicac : aceae (Parameswaran, 1961) aving Fuminate endosperm. However, the tissue Stologically different in various families. In De- 1971] BHANDARI, MAGNOLIALES 39 generiaceae (Swamy, 1949) and Annonaceae (Periasamy, 1962; Peria- samy & Swamy, 1961) the ruminating projections develop from the outer integument, while in Myristicaceae (Periasamy, 1961) they originate from the meristematic chalaza. For Canellaceae (Parameswaran, 1961), how- ever, such details are not known. The presence of cellular endosperm in most members of the Magnoliales and allies seems to indicate the primitiveness of this feature, and the re- stricted occurrence of the nuclear type may be a phylogenetic advance- ment. This is also confirmed by the fact that in the Ranunculales most her- baceous members have nuclear type. The presence of a haustorium which is not very aggressive is inexplicable; whether this is the preservation of a relic feature or a phylogenetic advancement which was to be well estab- lished in many other groups of angiosperms remains highly speculative. The latter possibility appears to be more sound. EmBrY0GENY. In some members of Magnoliaceae (Kapil & Bhandari, 1964), Degeneriaceae (Swamy, 1949) and Winteraceae (Bhandari & Ven- kataraman, 1968) the early segmentation is rather irregular, although at the preglobular or globular stages the embryogenesis proceeds normally and forms a healthy embryo. In other families, wherever information is available, the pattern of development becomes established and is mostly of the Onagrad (Annonaceae), Onagrad or Asterad (Schisandraceae, Lau- raceae), Asterad (Illiciaceae, Monimiaceae), or Solanad type (see Davis, 1966). However, one basic factor evident throughout is that the basal cell, cb, has little or no role in the organization of the embryo proper and forms the major part of the suspensor. SEED coat. In the Magnoliaceae, Winteraceae, Degeneriaceae, Annona- ceae, Myristicaceae, Lauraceae, and Cercidiphyllaceae the seed coat is con- stituted chiefly by the outer integument while the inner intgument degen- erates and its remnants may persist. In Canellaceae, Schisandraceae, and Trochodendraceae, however, both the integuments take part in forma- tion of the testa. In the Magnoliaceae and Degeneriaceae the seed coat is differentiated into outer fleshy and inner stony regions but in the rest of families the seed possesses a hard seed coat. In the Winteraceae it is mostly the enormously radially elongated outer epidermis which consti- tutes the seed coat while the rest of the layers become slightly thick walled. In Canellaceae and Trochodendraceae the outer epidermis of the outer integument and the inner epidermis of the inner integument become scleren- chymatous whereas the rest of the layers remain thin walled. The Schisan- draceae possess an elaborate seed structure (see Kapil & Jalan, 1964), having outer integument comprising the epidermis of macrosclereids, 2 or 3 subepidermal layers of brachysclereids, and 2 or 3 parenchymatous layers. he inner integument also persists as a thin layer of degenerated thick- walled cells. In Degeneriaceae, Myristicaceae, Canellaceae, and Annona- ceae the ruminations in the seed are produced by the outer integument. The Trochodendraceae and Cercidiphyllaceae possess a tapering chalazal Projection which develops into a wing-like structure in the mature seed. [To be continued | . 52 40 JOURNAL OF THE ARNOLD ARBORETUM [ VOL THE GENERA OF THE URTICACEAE THE SOUTHEASTERN UNITED STATES Norton G. MItuer URTICACEAE A. L. de Jussieu, Gen. Pl. 400. 17 89, (NerrLe Famizy) “Urticae,” nom. cons. P : ioecious, or polygamous; flowers imperfect, rarely perfect: i i in different infloresce ish or whitish, rare] ise, after shedding pollen; tepals 4, 5 [6], “Prepared for a Seneric flora of the southeastern United States, a project of the Arnold Arboretum and the Gray Herhari through the Suppo tt pedicels or sessile, aera equal, valvate, or partly fuse ouisiana. iS region, with additional information from extra-regional ta of treated species are derive examined and partly from data already published, References fol] are those which I have not been able to check. I am grateful to Dr. W G, Schubert, wh owed by an asterisk ood for his continuin 0 kindly examined a number o0 g advice and assistance; to Dr. Bernice ; and to Dr. Gord f specimens at the New York Botanical on P. DeWoll, Jr., who helpfully provided living dana. I have checked Tanges against material in the herbarium of the University of North Carolina and thank j T givi collection under their care, Mrs, Nancy Dunkly has h matters and in the Preparation of t by Miss Rachel A. Wh work of Miss Virginia Savage Served material Supplied by D : ations are base a Iam thankfy} for his commen d on living or pre- - John W. Thieret read the manuscript and 1971] MILLER, GENERA OF URTICACEAE 41 before anthesis [sometimes appendaged]; stamens of equal number and opposite the perianth segments [or a single stamen and a subtending in- volucre]; filaments initially inflexed, slightly flattened, pollen dispersal explosive as the stamens free themselves from the partially surrounding tepals; anthers 2-locular at anthesis, medifixed, thin walled, reniform, oriented with the concave sides inward, dehiscence longitudinal; pollen with 2, 3, or 4[-6] pores, + suboblate, exine stratification obscure, equatorial diameter ca. 15 »; rudimentary gynoecium prominent, conical, obovoid, globose, or cup-like, hyaline or not, glabrous [or woolly]. Carpel- late flowers minute, sessile or subsessile, greenish or reddish; tepals 3 or 4 or perianth + tubular [or rarely absent]; staminodia present or absent; gynoecium unicarpellate; style 1, present or absent, if present, apical at anthesis and rarely becoming laterally displaced in fruit; stigma penicillate, linear, or style prolonged into a filiform stigma; ovule 1, basal, usually orthotropous, sometimes hemianatropous, bitegmic, crassinucellar. Perfect flowers with 4 basally fused tepals and 4 opposite stamens, gynoe- cium with a penicillate or linear stigma. Fruits mostly laterally compressed or ovoid, symmetrical to asymmetrical achenes, rarely drupaceous, gen- erally either loosely or tightly surrounded by the accrescent perianth [which becomes fleshy in some], achene and perianth sometimes dispersed as a unit; style and stigma persistent or not. Seed 1, seed coats thin, brownish, membranaceous; endosperm thin, fleshy or mealy, completely surrounding the embryo; embryo straight, radicle pointing toward the Ovary apex and shorter to longer than the orbicular to ovate, sometimes emarginate, cotyledons. Embryo sac development of the Polygonum type, embryogeny of the Asterad type. Type GeNus: Urtica L. A natural family of about 40 genera and 800 species (ca. 1900 according to Hutchinson), largely confined to tropical and subtropical latitudes of both New and Old Worlds, rarer in temperate regions, and represented by only a few species still farther poleward. Of the six genera native to the United States (excluding Hawaii with four additional genera, two of which, Neraudia Gaud., with milky sap, and Touchardia Gaud., are endemic), five have one or more species in our region. Closely related to Urtica, but differing in the continuous, appressed perianth of the carpellate ower, Hesperocnide Torr. occurs in California (H. tenella Torr.) and Hawaii (H. sandwicensis Wedd.; however, see W. Hillebrand, Fi. Hawaii. Is. 408. 1888, who regards the Hawaiian plants as probably unintentionally introduced from California at an early date). | The family has traditionally been divided into five tribes. Four have representatives in North America. Floral morphology in the exclusively Old World tribe Forskohleeae Gaud. is somewhat atypical for the family as a whole, in that staminate flowers are comprised of a single stamen and a subtending perianth (or involucre), while carpellate flowers may either have a perianth or lack one altogether. The other tribes are distinguished mainly on the nature of the perianth in both staminate and carpellate flowers, Only members of the Urticeae have stinging hairs. As has been 42 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 recently suggested, Cecropia L., Coussapoa Aubl., and Musanga C. Sm. r., which generally have been placed in subfam. Conocephaloideae of the Moraceae, are better treated as a separate tribe of the Urticaceae on account of their basal, orthotropous ovules (cf. Chew, Corner). The oldest tribal name available for these genera would seem to be Cecropieae aud. Wind pollination is characteristic of the family, and the staminate in which epidermal and fibrous layers have been successively lost. Inter- mediate stages in the series are encountered in other members of the Urticales. as el ins.? pert ener Ashed leaf fragments containing cystoliths, calcium tystals, and calcified or silicified cel] walls, have been studied * Sree (adj. remoeetilns bes Nrbge AR 18. 1968) have introduced the term “furuncle” of inhi ns in dried materia] arising from either cystoliths 1971] MILLER, GENERA OF URTICACEAE 43 microscopically, either embedded in celloidin to preserve their original orientation and arrangement in the leaf (a spodogram), or as simple mounts of the ash residue. Cystoliths may be diagnostically useful at various taxonomic levels. There is a general correspondence between cystolith shape and tribes of the Urticaceae (e.g., members of the Elato- stemeae Gaud. for the most part have linear cystoliths), and different species within a genus may have markings of distinctive shape. Chromosome counts reported for 17 genera are 2n = 14, 16; 20,22, 24; 26, 28, 32, 42, 48, 52, 60, and 76-78. The Urticaceae, Moraceae, Cannabaceae, and Ulmaceae (and perhaps the Barbeyaceae) are generally taken to comprise the Urticales whose derivation Ttom the Magnoliales through the Hamamelidales is postulated by many workers. A relationship between the Urticaceae and certain families of the Malvales has been suggested also. Apart from ramie (Boehmeria nivea), a fairly important fiber plant in the Orient, the family is of little importance economically. Several species of Pilea and Pellionia Gaud. are grown as ornamentals, and Soleirolia Soletrolii (Req.) Dandy (Helxine soleirolii Req.), baby tears, 2n = 20, a native of Corsica and Sardinia, is a favored houseplant. REFERENCES: Battton, H. Urticacées, Hist. Pl. 3: 496-537. 1872. [English transl. M. M. Hartoc, The natural history of plants. 3: 497-538. London. 1874.] BECHTEL, A. R. The floral anatomy of the Urticales. Am. Jour. Bot. 8: 386- 410. 8 pls. 1921. [Urticaceae, 397-399, 402-407, pls. 21, 22; Boehmeria cylindrica, Laportea canadensis, Urtica gracilis. BENTHAM, G., & J. D. Hooker. Urticaceae. Gen. Pl. 3: 341-395. 1880. [Tribe Urticeae, 381-395. ] BERNBECK, F. Vergleichende Morphologie der Urticaceen- und Moraceen-In- floreszenzen. Bot. Abh. Jena 19. 100 pp. 1932. [Boehmeria (3 spp. including B. cylindrica and B. nivea), Debregeasia, Dendrocnide, Elatostema, Laportea canadensis, Memorialis, Myriocarpa, Pellionia, Pilea (2 spp.), Pipturus, Procris, Urtica urens. ; ; Bicatxe, H. Die Blattspodogramme der Urticaceae und ihre Verwendbarkeit fiir die Systematik. Beitr. Biol. Pflanzen 21: 1-58. 5 pls. 1933. [In general, type and distribution of cystoliths correlate with classification of the family. | BOLKHovskiku, Z., V. Grir, T. Matveyeva, & O. ZakHARYEVA, Chromosome numbers of flowering plants. (Russian and English prefaces.) 926 pp. Leningrad. 1969. [Urticaceae, 711, 712.] ice § th CHEW, W. L. Florae Malesianae precursores — XXXIV. A revision 0 hg genus Poikilospermum (Urticaceae). Gard. Bull. Singapore 20: = : 1963, [General notes on the family Urticaceae and its circumscription, ¢, 22-29, : : Corner, E. J. H. The classification of Moraceae. Gard. Bull. Singapore 19: 187-252. 1962. [Distinctions between Moraceae and Urticaceae. | Bile CorreLL, D. S., & M. C. Jounston. Manual of the vascular gsc en Xv + 1881 pp. Frontisp. -+ 1 map. Renner, Texas. 1970. [Urticaceae, 498-503. ] 44 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Cronguist, A. The evolution and classification of flowering plants. xii + 396 pp. Boston. 1968. [Urticaceae, 166, 167.] Davis, G. L. Systematic embryology of the angiosperms, x + 528 pp. New York, London, & Sydney. 1966. [Urticaceae, 268, 269. | Duke, J. A. On tropical tree seedlings. I. Seeds, seedlings, systems, and sys- tematics. Ann. Missouri Bot. Gard. 56: 125-161. 1969. [Urticaceae, 146. ] Eames, A. J. Morphology of the angiosperms. xiv + 518 pp. New York, Toronto, & London. 1961. [Numerous references to Urticaceae noted in index. EIcH LER, A. W. Urticaceae. Blithendiagramme 2: 49-54. 1878. Encter, A. Urticaceae, Nat. Pflanzenfam. IIT. 1: 98-118. 1888. ErptMaN, G. Pollen morphology and plant taxonomy. Angiosperms. Corrected reprint and new addendum. xiv + 553 pp. New York & London. 1966. Farcrt, K., & L. vAN DER PrjL. The principles of pollination ecology. x + 248 pp. Toronto & Oxford. 1966. [Anemophily in members of the Urticaceae. | FREISLEBEN, R. Untersuchungen iiber Bildung und Auflésung von Cystolithen a den Urticales. Flora 127: 1-45. 1933. [Boehmeria nivea, Urtica 2 spp.). GUERIN, P. Les laticiféres de ] Bot. France 52: 406-411. 1 . Cellules 4 mucilage chez les Urticées. Ibid. 57: 399-406. 1910. [ Boeh- meria (2 spp. including B. nivea), Urera.| . L’Urera Humblotii H. Baillon et ses affinités. Compt. Rend. Acad. Sci. Paris 168: 517-519, 1919. [Presence of laticifers analogous to those in the Moraceae. | Le mucilage chez les Urticées. Ibid. 174: 480-482. 1922. [Mucilage cells in Urticaceae analogous to those in Tiliaceae. ] ———.. Les Urticées: cellules 4 mucilage, laticiféres et canaux sécréteurs. Bull. Soc. Bot. France 70: 125-136, 207-215, 255-263. 1923. [Family survey; laticifers identified in Dendrocnide spp. | HeEyYwoop, V. H., ed, Urticaceae. In: T. G. Turin, V. H. Heywoon, et al., eds. 9. 1964. aia treatments of Parietaria, Pilea, and ALL 'Urera baccifera Gaud, et leur contenu. Bull. Soc. 905. tralbl. 44: 1-89. 1927. [Leaf Pilea pumila, 48, 49.] abe t cellules annexes chez les Procrideae. Bull. Soc. Hist. SB sien Eo 265-269. 1969 [1970]. [Elatostema, Pilea, Procris. | y=: “ie genera of flowering plants. V. i 2s xi 1967. [Urticaceae, 178-195, ] . 2. eh ee ee & RW. Seueev. Flora of ee “pe _ cee . Gard. 47: 179-193. 1960. [Boehmeria (5 spp. - Cyunarica), Laporte j ] Phew ill hg sae portea (1 sp.), Myriocarpa, Pilea (17 spp.), » E. Urticaceae. Jn: Gy. Lebensgeschichte der Bliitenp [Urtica dioica, U 48: 9-13, 1930. [Chromosome counts in (2 spp.), Pellionia (1 sp.).] Kummer, A. P. Weed seedlings. XxXXiv + 435 Pp. Chicago. 1951. [Urtica 1971] MILLER, GENERA OF URTICACEAE 45 dioica, U. procera, Laportea canadensis, Boehmeria cylindrica, Parietaria pensylvanica, 4-8. | Kuprianova, L. A. The palynology of the Amentiferae. (In Russian.) 214 pp. 48 pls. Bot. Inst. Komarov. Akad. Nauk SSSR. Izdat. ‘Nauka.’ Moskva- Leningrad. 1965.* LAKELA, O., & F. C. CratGHEAD. Annotated checklist of the vascular plants of Collier, Dade, and Monroe Counties, Florida. viii + 95 pp. Coral Gables, Florida. 1965. [Boehmeria cylindrica, B. Drummondiana, Parietaria flori- dana, Pilea herniarioides, P. microphylla, 33.| Lussock, J. A contribution to our knowledge of seedlings. Vol. 2. ii + 646 pp. London. 1892. [Urticaceae, 500-504. | Martin, A. C. The comparative internal morphology of seeds. Am. Midl. Nat. 36: 513-660. 1946. [Urticaceae, 608, 609.] Metcuior, H. Urticaceae. Jn: H. Metcutor, Engler’s Syllabus der Pflanzen- familien. ed. 12. 2: 57-59. 1964. Metcatre, C. R., & L. CHatx. Urticaceae. Anat. Dicot. 2: 1244-1254. 1950. Mopitewsky, J. Zur Samenentwicklung einiger Urticifloren. Flora 98: 423- 470. 1908. [Boehmeria (1 sp.), Dendrocnide, Elatostema, Laportea (1 sp.), Parietaria officinalis, Urera, Urtica (4 spp. incl. U. dioica and U. urens).] MosepacH, G. Uber die Schleuderbewegung der Explodierenden Staubgefasse und Staminodien bei einigen Urticaceen. Planta 16: 70-115. 1932. [Ex- plosive pollen dispersal in Dendrocnide moroides (Wedd.) Chew, Pellionia Daveauana N. E. Br., Pilea Spruceana Wedd., and P. serpyllifolia (Poir.) Wedd.; explosive fruit dispersal in P. Spruceana. | Riptey, H. N. The dispersal of plants throughout the world, xx + 744 pp. 22 pls. Ashford, England. 1930. [Notes on dispersal of Boehmeria, Parie- taria, Pilea, and Urtica species. | Rivizres, R. Fleurs et inflorescence de quelques Urticacées. Nat. Monspel. Bot. 8: 189-204. 1956. [Individual urticaceous flowers interpreted as shortened spicate inflorescences in which perianth segments are subtending bracts; unisexuality secondarily obtained through abortion of the androecium or gynoecium. | ! _ La valeur de la fleur d’Urticacée. Compt. Rend. Acad. Sci. Paris 244: 653-656. 1957. [Floral morphology; see RIvIERES (1956). ] f Rosinson, C. B. Philippine Urticaceae. Philip. Jour. Sa: Bot.” 5: 465-543. 1910; ibid, 6: 1-33. 3 pls. 1911. [General nomenclatural notes and regional revision. SATAKE, Y, Systematic and anatomical studies on some Japanese plants. I. Jour. Fac. Sci. Univ. Tokyo Bot. 3: 485-511. 1931. [Systematic importance of spodograms in the Urticales, 485-507; includes key to Japanese members of the Urticaceae (pp. 500-505), including species of Boehmeria, Laportea, Parietaria, Pilea, and Urtica, based almost entirely on cystolith character- istics. SAYEEDUD-Din, M., & M. Aspus SALAM. On the anatomy of some of the Urticaceae. Jour. Bombay Nat. Hist. Soc. 43: 274-276. 1 pl. 1942. [Pel- lionia Daveauana and Pilea microphylla. Scuretper, A, Urticaceae. In: G. Hect, Illus. Fi. Mittel-Europa. ed. ' ae 296-307. 1 pl. 1958. [Urtica dioica, U. urens, Parietaria o ficinalis. | SHarmA, B. R. Cytology of some north-west Himalayan Urticaceae. Jour. Indian Bot. Soc. 40: 355-364. 1 pl. 1961. [Includes review of chromosome numbers reported for species of Boehmeria, Girardinia, Lecanthus, Myrio- carpa, Parietaria, Pellionia, Pilea, Pipturus, Urtica. 46 JOURNAL OF THE ARNOLD ARBORETUM [voL, $2 SOLEREDER, H. Systematic anatomy of the dicotyledons. Vol. 2. vi + pp. 645- 1182. (Transl. L. A. Boopte & F. E. sta revised D. H. Scort.) Oxford. 1908. [Urticaceae, 775-778, 1056, 1063. STAGER, R. eo Nachweis von Nektarien bei Pollenblumen und Anemo- philen. Beih. Bot. Centralbl. 12: 34-43. 1902. [ Positive chemical test for the occurrence of sugar in staminate flowers of Urtica dioica suggests the presence of vestigial nectaries. ] STRASBURGER, E, Sexuelle und apogame Fortpflanzung bei Urticaceen. Jahrb. Wiss. Bot. 47: 245-288. 4 pls. 1910. [Urtica dioica and Elatostema spp. | TakuHTajAN, A. Flowering plants, origin and dispersal. (Transl. C. Jeffrey.) x + 310 pp. Washington, D.C. 1969 [Urticaceae, 211.] : ~262. 1967. [Urticaceae, 255, 256, 258-260, 261; pollen of Boehmeria (2 spp.), Laportea canadensis, Pilea (2 spp-), Parietaria (2 spp.), Urtica dioica, U. urens (and 4 other spp.) described and illus- tr The vascular plants of southwestern Georgia. Am. Midl. Nat. ; n » & N. R. Lersten. The morphology and toxicology of plant stinging hairs. Bot. Rev. 35: 393-412, 1969 [1970]. [Urticaceae, 394~398, 401-405, Figs. 2-7, 26.] Trepo, O. Comparative anatomy of the Mor Bot. Gaz. 100: 1-99, 1938. [Urticaceae, 28 Unrun, M. Aniso bei Urticaceen aceae and their presumed allies. ene phyllie und Wechselbeziehungen des Wachstums am Spross - Ber. Deutsch. Bot. Ges, 58: 484-488, 1940. [Leucosyke, - A. Revue de la famille des Urticées. Ann. Sci. Nat. Bot. IV. 1: 173-212, 1854, : famille des Urticées. Arch. Mus. Hist. Nat. Paris 9: 1-400. 1856; ibid, 401 592. 20 pls. 1857. [See also Ann, Sci. Nat. Bot. a republication of Weddell’s monograph with the itted.] . Urticaceae. DC. Pp WERNER, 0. Haar- und C Blattgeweben bei Urticales, bei Bryonia dioica wnd Zexmenia longipetiolata. Osterr. Bot. Zeitschr. 80: 81- ] KEY To THE GENERA OF THE URTICACEAE IN THE General characters: monoecious, dioecious, and polygamous annual or perennial herbs, with or without stinging haj exstipulate, mostly SOUTHEASTERN UNITED STATES img hairs; leaves stipulate or nerally orthotropous ; Style o > me or absent, stigma generally 1971] MILLER, GENERA OF URTICACEAE 47 penicillate or linear; fruits mostly achenes wots drupaceous), loosely to tightly surrounded ke the persistent perianth or n ea . Plants with stinging hairs; perianth of carpellate wor at fruiting stage comprised of four separate tepals, 1 pair large, the other small. B. Leaves opposite; style absent, stigma apical, seniellaid: achene erect, symmetrical, nonstipitate; pedicel supporting mature fruit short and PU Sk A ga de 8th io een lah ew ad ae aa . Urtica. Leaves alternate; style and stigma awl-shaped, at first apical but dis- placed to a lateral position during fruit maturation; achene deflexed, ts te stipitate; pedicel supporting mature fruit long and bearing SRT WIR eal pig he ec acne Wen ps 2. Laportea., Plants without stinging hairs, otherwise pubescent or not; perianth of carpellate flowers at fruiting stage either tubular or of three separate tepals. C. Perianth of carpellate flowers with 3 equal or 1 hooded and 2 non- hooded tepals; staminodia present, inflexed beneath the maturing ovary. wei dees trae stair aD wisn eokaig cs las ag UR a age cae ee RN te tae 3. Pilea. Perianth of carpellate flowers = oe continuous or comprised of 4 mostly fused segments; staminodia a D. Perianth of carpellate flowers (aeedeek flowers absent) continuous, tightly covering the mature fruit; style apical and flexed near the tip, stigma linear; plants generally tall, nea: perennials, stem nonsucculent; stipules or stipule scars present. .... 4. Boehmeria?® . Perianth of carpellate and perfect aba accrescent, segments mostly fused; style short or absent, stigma penicillate; plants mostly delicate annuals, stems watery and usually translucent; stipules MIMBO 552 eo ek a ee es Nd es i es 5. Parietaria. a > O o Tribe Urticear [Urereae Gaud. | 1. Urtica Linnaeus, Sp. PI. 2: 983. 1753; Gen. Pl. ed. 5. 423. 1754. Tall, erect or sometimes ascending, fibrous annual or perennial herbs of diverse habitats, including thickets, bottomlands, damp forests, road- sides, shell middens, other waste places, and sometimes fields. Roots in annual species fibrous but usually with a short taproot, or, in perennials, with a horizontal rootstock bearing adventitious roots and aérial branches. Stems generally fluted, mostly hollow, greenish or sometimes purple or red, in certain species bearing numerous stinging hairs on short [or long] ‘pedestals, particularly at nodes. Leaves simple [rarely deeply incised], petiolate, opposite, blade mostly ovate or lanceolate, margins serrate, often with stinging hairs on both upper and lower surfaces; punctiform, rarely linear cystoliths visible on upper surfaces of dried leaves; stipules lanceo- late, generally paired at the petiole base, membranaceous, green [or brown], sometimes ciliate, free [or connate], usually persistent. Plants monoecious or dioecious; inflorescences bracteate, basically cymose, flowers organized into lax racemes or aggregated into loose [or tight] clusters that originate from axils of stipules and leaves, in some species zones of stamin- ms henry zeylanica (L.) J. Benn. (tribe Boehmerieae) from Asia, collected once as a weed in Seminole County, Florida, would key out here. For its separation from es shecad we of the family see footnote 5. 48 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 X te. Tacs ey ee Vise Pho 5°) - FIGURE 1. Urtica, a—k U portion of adaxial lea , » xX A; Dd, f surf hair from petiole, X 12; d, inflorescenc WwW ; e, post-anthesin staminate flower , X 40; g, longitudinal section of ovary with orthotropous o ee h, achene and accrescent peri- anth, X 15; i, mature achene, x 20: 50 e % , i mature achene in longitudinal section Tyo, endosperm stippled, < 20; k, wet achene, mucilage halo dotted, X 20.1, U. dioica: mature achene, X 20. icropyle formed from the inner integu- 1971] MILLER, GENERA OF URTICACEAE 49 ment. Fruit a thin-walled, compressed, ovoid achene, loosely inclosed by enlarged inner tepals, outer tepals remaining small; apical stigmatic tuft persistent or not. Seed coats thin; endosperm fleshy, forming a thin layer around the embryo, some species with a chalazal endosperm haustorium; embryo straight, radicle generally shorter than the + orbicular, emarginate cotyledons. Lectotype spectrs: U. dioica L.; see Britton & Brown, Illus. Fl. No. U. S. ed. 2. 1: 634. 1913. (Classical Latin name, from wrere, to burn, surely in reference to the stinging hairs.) — NErtte. A genus of about 45 species, mostly indigenous to temperate parts of the Northern and Southern Hemispheres, but a few species in the New and Old World tropics, usually in the mountains at middle and upper elevations. Urtica dioica L. and U. urens L. are exceptionally widespread weeds. In North America U. gracilis Ait. extends northward through the Boreal Forest to the edge of the subarctic (also in central Alaska). The genus is badly in need of monographic study, especially in North America where regional floristic treatments of the perennial species vary dramatically. About 10 species of Urtica generally have been recognized on this continent north of Mexico, and of these, seven are perennials. Urtica californica Greene, U. holosericea Nutt., and U. Lyallii S. Wats. occur in western North America; U. serra Blume is a species of the South- west; U. dioica, U. procera Muhl. ex Willd., and U. viridis Rydb. are generally northern and mostly restricted to the central and eastern sections (strictly dioecious plants of U. dioica may be entirely introduced from Europe); and U. gracilis is northern and transcontinental. Urtica dioica, U. gracilis (sensu Small, Man. Southeast. Fl. 433. 1933), and U. procera (sensu Fernald) have ranges that include a portion of the southeastern United States. The latter two species are often treated as U. dioica var. procera (Muhl. ex Willd.) Wedd. (Hermann; however, see Hitchcock for a different interpretation). Urtica dioica in the broad sense seems clearly to be a circumpolar species complex. The temperate Asian representative of the complex is U. angustifolia Fisch. ex Hornem. Two annual nettles occur in the southeastern United States. The Eura- sian Urtica urens is sporadically naturalized in Florida, South Carolina, and probably elsewhere in the region. The native U. chamaedryoides Pursh, generally a plant of flood plains and rich woods often over lime- stone, occurs from central Florida west to Texas and Mexico. It is na- tive northward to South Carolina, West Virginia, and Missouri, and oc- casionally is found as an adventive beyond these states. Disjunct populations have been reported from Argentina. A third annual, U. gra- cilenta Greene, occurs in Arizona, New Mexico, and Texas. Cystoliths are prominent in the leaves of many Urtica species, and both punctate and bacilliform types are present. The effervescence of those in U. chamaedryoides and U. urens when dilute hydrochloric acid is added indicates that the cystoliths are probably mineralized with calcium car- bonate. Specimens of U. chamaedryoides generally have bacilliform cysto- liths, although a few collections from Texas have been seen with puncti- form markings. 50 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Stinging hairs are present or absent on leaves, stems, inflorescence axes, and tepals, and occur on both annual and perennial species. Each hair consists of a single, long, narrow, tapering stinging cell with a sac-like base embedded in a multicellular emergence. The hair is closed at the tip by a small bulb that, when pressure is applied, breaks off along a pre- determined line of fracture producing an extremely sharp, bevelled point. Hair cells are reported to be both silicified and calcified. Upon contact with the skin, the hair punctures the surface and compression of the base forces the contained fluid into the wound. Contact with nettles results in a three-phase reaction involving formation of a wheal (associated with itching), reddening, and pain. The chemical basis of these responses has not been fully determined but experimental data obtained from the hairs of Urtica urens indicate that in this species histamine and acetylcholine are responsible for the itching and burning sensations, respectively. Base chromosome numbers for the counted species are x = 11, 12, and 13. Diverse chromosome numbers appear to be documented in Ur- tica dioica, sensu lato, 2n = 22, 52 (U. kioviensis Rogowitch — U. divica?), 48, 52 (Live & Love), and 52, 76-78 (U. platyphylla Wedd. = U. dioica?); diploid and tetraploid races occur in U. urens, 2n = 24, 26, and 52. Sex chromosomes tentatively identified in U. dioica by Meurman have not been confirmed subsequently (Westergaard). Little has been written about the dispersal of Urtica species. According to Ridley, the achene and perianth (which surrounds the fruit to maturity) may be transported as a unit by passing animals. In some species the stiff, curved, downward-pointing hairs that cover outer tepal surfaces provide the means of attachment. Achenes from the annuals U. chamae- dryoides and U. urens become mucilaginous upon wetting, but those of the perennial nettles U. dioica, U. gracilis, and U. procera, in contrast, do not develop mucilage when placed in water (Miller), Annual and perennial species from other parts of the world should be examined to document this difference further, REFERENCES: tees ek see BECHTEL, BERNBECK, ERDTMAN, FREISLEBEN, A BE, KUMMER, ARTIN, Mopr ; LEWSKY, RIDLEY, SATAKE SCHREIBER, SHARMA, STRASBURGER, TARNAVSCHI et al. and THURSTON & LER- STEN, and WERNER. ; Bayer, G. Die grosse Brennessel (Urtica dioica L.) in ihrer Bedeutung fiir die 1971] MILLER, GENERA OF URTICACEAE 51 Gesundheit von Mensch, Tier und Pflanze. Planta Med. 5: 12-23. 1957. [ Economic botany. CuHEsNuT, V. K. The stinging nettles. Pl. World 1: 116, 117. 1898. [Urtica, Laportea. | Cottier, H. O. J., & G. B. CuesHer. Identification of 5- -hydroxytryptamine in the sting of the nettle (Urtica dioica). Brit. Jour. Pharmacol, Chemother. 11: 186-189. 1956.* Dapper, H. Mondézie und Hermaphroditismus bei Urtica dioica L. Flora B. Morphol Geobot. 157: 97-99. 1967. EMMELIN & W. Fetpperc. The mechanism of the Ps of the common nettle ai en urens). Jour. Physiol. 106: 440-455. FERNALD, M. L. Urtica gracilis and some related North are species. (Jn: M. i. FERNALD. Two summers of botanizing in Newfoundland.) Rhodora 28: 191-199. 1926. & A. C. Kinsey. Edible wild plants of eastern North America. Rev. ed. y R. C. Rotts. New York, & Evanston, Illinois. xvi + 452 pp. 1958. Urtica, 164-166. | GRAVIS, "7 fen ipa anatomiques sur les organes végétatifs de l’Urtica dioica . Cour. Sav. Etr. Acad. Roy. Sci. Lett. Arts Belg. Quarto Sci. 47. = a 256 pp. 23 pls. 1884. [Reprint edition, Librairie neice et scientifique de A. Manceux. Brussels. 1885.] rts, S.C. Sait A growth in Urtica dioica L. Agra Univ. Jour. Res. Sci. 9: 93-102. 1960. Havuman, L. Notes asia deuxiéme série (Dicotylédones de |’Argentine). Anal. Mus. Nac. Hist. Nat. Buenos Aires 32: 395-475. 1923-25. [U. chamaedryoides, 413-415, ] HERMANN, F. J. The perennial species of Urtica in the none States east of the Rocky Mountains. Am. Midl. Nat. 35: ser 778. 6. Hircucock, C. L. Urticaceae: Ja: C. L. Hircucock, A. Fomaeny et al. Vasc. Fi. Pacific Northwest 2: 90-95. 1964. fRevision of West Coast Urtica species bearing on the taxonomy of the genus throughout North America.] Husson, P. Etude de l'appareil stomatique chez la plantule d’Urtica mem- bronacea Poir. Bull. Soc. Hist. Nat. Toulouse 104: 125-130. 1968. [Anomo- cytic and anisocytic stomata in seedling. ] Jounson, C. P. The useful plants of Great Britain. 324 pp. 25 pls. London. 1861 [1862]. [U. dioica, 228-230.] LAUDERMILK, J. The plant with the hypodermic needle. Nat. Hist. 61: 350- 1952. [Urtica.] LesicuE, A. Embryogénie des Urticacées. Development de l’embryon chez l'Urtica dioica L. Compt. Rend. Acad. Sci. Paris 242: 923-926. 1956. [Embryogenesis in Sganirgs of the Urticaceae similar to that known in members of the Malvace & R. Porx. Le ‘evclopoenaat i ope a chez VUrtica urens L. Bull. Soc. ne France 103: 587-589. 1 Love, A., & D. Love. Cytotaxonomical einer of the Icelandic flora. Acta Horti sai 20: 65-290. 1956. [U. dioica, 204, 214, 215; U. urens, 204.] Mepvepev, P. F. On the economic and bivlogical significance of dioecious speci- mens of Urtica dioica L. (In Russian.) Bot. Zhur. 43: 1704-1707. 1958.* MEtcatre, C. R. Economic value of the common stinging nettle. Nature 150: . 1942 MevuRMAN, 0. The chromosome behaviour of some dioecious plants and their 52 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 relatives with special reference to the sex chromosomes. Comment. Biol. Helsinki 2(3). 104 pp. 1925, [U. dioica (2n = 48), U. urens (2n = 24), Navas B., E. El género Urtica en Chile. Bol. Soc. Argent. Bot. 9: 395-413. 1961. [10 spp. including U. dioica and U. urens.] Necopi, G. Sporofilli e gametofiti in Urtica caudata Vahl. Ann. Bot. Roma 18: 325-328. 1930. NESTLER, A. Zur Kenntnis der Wirkung der Brennhaare unserer Urtica-Arten. Ber. Deutsch. Bot. Ges. 43: 497-504. 1925. [Stinging hairs of U. dioica and U. urens.| Prtcrmm, R. L. C. Some properties of the sting of the New Zealand nettle, Urtica ferox. Proc. Roy. Soc. London B. 151: 48-56. 1959. SELANDER, S. Urtica gracilis Ait. in Fennoscandia. Sv. Bot. Tidskr. 41: 264— 282. 1947. [Considered indigenous. | Sovices, R. Embryogénie des Urticacées. Développement de l’embryon chez Y'Urtica pilulifera L. Compt. Rend. Acad. Sci. Paris 171: 1009-1011. 1920. . Développement de l’embryon chez l’Urtica pilulifera L. Bull. Soc. Bot. France 68: 172-188, 280-294. 1921. TxHurston, E. L., & N. R. Lersten. The ontogeny and histochemistry of sting- ing hairs in Urtica dioica. (Abstr.) Am. Jour. Bot. 55: 705. 1968. Tsuxapa, M. The fine sculpturing of pollen surfaces and some terminological problems. (In Japanese; English abstr.) Bot. Mag. Tokyo 81: 385-395. 1968. [Electron microscopy of Urtica dioica pollen. | WESTERGAARD, M. The mechanism of sex determination in dioecious flowering plants. Advances Genet. 9: 217-281. 1958. [U. dioica, 225.] 2. Laportea Gaudichaud-Beaupré in Freycinet, Voy. Monde Bot. 498. 1830 (‘1826’), nom. cons ?), restricted to ‘ petiolar, bifid toward the apex, fugacious. Plants monoecious; inflorescences axillary, lax bracteate cymose panicles; carpellate and staminate flowers in separate inflorescences [rarely to- gether], the carpellate inflorescences borne higher on the axis than the staminate. Staminate flowers small, greenish, pedicellate [or sessile] with 5 [4] saccate, longitudinally ridged tepals; stamens 5 [4]; rudimentary gynoecium large, obovoid, hyaline. Carpellate flowers small, zygomorphic borne on a thick pedicel that develops lateral [or dorsi-ventral] wings as the fruit matures; tepals 4, free, the i 1971] MILLER, GENERA OF URTICACEAE 53 . VS Olu SSEERE SE WO’ ~ [fet toh Ripe : oS a= [Shira pdey SS baad Sno Ss Res spc 25 2-22: hs Mery Sethe . titer? 2.55 Se = Ficure 2. Laportea. a-h, L. canadensis: a, leaf, adaxial surface, X %4; b, portion of carpellate inflorescence, cymules containing both flowers (at arrows) and fruits, X 3; c, staminate flower bud, X 30; d, carpellate flower, X 25; e, young fruit in longitudinal section showing position of ovule shortly after fertil- ization, stigma mostly removed, X 25; f, enlarged perianth with mature achene and winged pedicel, X 10; g, mature achene, X 8; h, mature achene in longitu- dinal section showing embryo, endosperm white, x 8. mostly hidden by the mature fruit, abaxial tepal only slightly larger than at anthesis, base of style persistent and displaced laterally by unequal development of the achene and embryo. Seed coats thin; endosperm mealy, sparse; cotyledons large, orbicular, radicle short. (Urticastrum Heister ex Fabr., nom. rejic.) Lectotype species: L. canadensis (L.) Wedd.; see Hitchcock & Greene, Prop. Brit. Bot. Int. Bot. Congr. Cam- bridge, England, 1930. 101. 1929. (Named for Francois L. de Laporte, a 19th century French entomologist.) — Woop-NETTLE. A predominantly Old World genus of 22 species, 12 restricted to either the Malagasy Republic (5 species endemic) or Africa south of the Sahara (Chew, 1969). Other Laporteas occur in Asia (4 species), Polynesia and/ or Malesia (2 species), and widely throughout the Old World tropics (1 species), Of those in the Western Hemisphere, Laportea (Fleurya) cun- 54 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 eata (A. Rich.) Chew, is known only from Cuba and Hispaniola, L. (Fleurya) aestuans (L.) Chew is widespread in the Caribbean and Central and northern South America (also in the Old World tropics), while L. (Laportea) canadensis (L.) Wedd. is widely distributed in eastern North America from Nova Scotia to southern Manitoba, south to Louisiana and northern Florida. Disjunct stations occur in the eastern Mexican highlands (Nuevo Leon and Tamaulipas). in the Piedmont and Mountain districts. Typically, it is a species of mesophytic hardwood forests. Although apparently not reported from the United States, ZL. aestuans should be sought in southern peninsular Florida and the Keys. segregated as the genus Dendrocnide (Miq.) Chew, one member of which, ‘edd. Ww, an endemic of Queensland and New South Wales, Australia, attains a height of 35 m. REFERENCES: Under family references see BECHTEL, BERNBECK ? , Kiturp, Kummer, Martin, SATAKE, TARNAVSCHI et al., THORNE (1954), and T N. HURSTON & LERSTE CHEsNUT, V. K. The stinging nettles. Pl, World 1: 116, 117. 1898. [Urtica Laportea. | : uch W. L. Laportea and allied genera (Urticaceae). Gard. Bull. Singapore 1; 195-208. 1965, [Segregates Dendrocnide, unites Fleurya and Laportea; 23 spp. of Laportea, including L, canadensis. | - A monograph of Dendrocnide (Urticaceae). Jbid. 25: 1-104. 1969. . A monogra ph of Laportea (Urticaceae). Ibid, 111-178 [T me : 5 —178. reatment Parone _ same as Chew, 1965, except Sceptrocnide Maxim. united with FRANCIS, W. D. Australian stingin trees 1956.* [Dendrocnide.] § trees. Queensland Nat. 15(4/6): 66-70. -MAcrFar.ane, W., V. The stinging properties of Laportea. Econ. Bot. 17: 303- 311. 1963. [Irritating fluid from L. moroides (= Dendrocnide moroides) partially characterized but not identified chemically; not the substa producing the response in Urtica spp. | nei 1971] MILLER, GENERA OF URTICACEAE wat wn Tribe ELATOSTEMEAE Gaud. [Procrideae Wedd. | 3. Pilea Lindley, Collect. Bot. ad pl. 4. 1821, nom. cons.4 after pollination; ovule orthotropous. Fruit a + symmetrical, laterally compressed, light brown to black achene, surface smooth and mottled, or papillose, or covered with prominent bosses. Endosperm fleshy; cotyledons Ovate, not emarginate. (Adicea Raf. ex Britton & Brown.*) TYPE SPECIES: Pilea mucosa Lindl.; see Collect. Bot. ad pl. 4. 1821, nom. illeg. = P microphylla (L.) Liebm, (Name from Latin, pileus, a felt cap, in refer- ence to the hooded tepal in the type species.) — RICHWEED, CLEARWEED. A much ignored genus, primarily pantropical, with a few species in temperate parts of the Northern and Southern Hemispheres; absent from Europe, Australia, and New Zealand. Pilea is by far the largest genus of the Urticaceae, with over 600 species described, but the actual number probably considerably less than this. The American tropics, especially the Caribbean region, are rich in species. Killip records 112 species from northern and western South America; Leén & Alain report that of the 64 species occurring in Cuba, 51 are endemic. Four species are found in 56 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 eastern North America. The genus is not represented in the western half of this continent. Weddell’s two outdated treatments (1856, 1869; 136 and 159 spp. recognized, respectively), in which Pilea is artifically broken down into three sections and a number of lesser categories, mainly on the basis of leaf characteristics, are the only monographs of the entire genus. Killip has recognized 12 informal species-groups based similarly on vegetative traits. Our native species fall into two, perhaps more, natural categories according to the presence or absence of a hooded third tepal. Section Pirea (Integrifoliae Wedd.), containing species with entire leaves, is represented in the Southeast by Pilea microphylla (L.) Liebm., artillery plant, 2n = 36, and P. herniarioides (Sw.) Wedd., both restricted to the southern parts of our area but fairly widespread in the American tropics. The former, a plant of pinelands, forests, hammocks, and some- times weedy places, has been found sporadically throughout Florida (in part escaped from cultivation) and may occur at climatically favorable places elsewhere in the Southeast. It is present in the Old World tropics, probably also as a garden escape, and is grown commonly in greenhouses in temperate regions. Pilea herniarioides is known from damp ground and hammocks in southern peninsular Florida and the Keys. It differs from the more fleshy and ascending P. microphylla in being repent and in having filiform stems and deltoid to +. orbicular leaves with petioles as long as or longer than the blades (vs. obovate or lanceolate leaves with short and South America seem to include several elements, one of which ap- proaches P. microphylla. The small, entire-leaved pileas are in need of further study, as a comparison of the descriptions of P. herniarioides in Killip (1939, 1960) and Fawcett and Rendle shows. Small (Man. South- east. FI. 434. 1933) lists a third member of the section, P. serpyllifolia (Poir.) Wedd., from southern Florida and coastal Louisiana, but speci- mens so-named by Small lack the ciliat sli e leaves characteristic of the type (see Killip, 1939, pp. 479, 480) and appear to be only robust plants of P. microphylla, Species in sect. HETEROPHYLLAR Wedd, have distinctly anisophyllous paired leaves, Both New and Old World members are cited by Weddell (1869), but anisophyllous species of Pilea are especially abundant in the American tropics. Our other two species belong to sect. DENTATAR Wedd., members of which have neither entire nor anisophyllous leaf-pairs. Pilea pumila (L.) Gray (including var, Deamii (Lunell) Fern.), 2n = 16, a plant of moist, shaded woods, occurs widely from New Brunswick to North Dakota, south to Texas, Louisiana, Mississippi, Alabama, Georgia lori The less frequently collected P. fontana (Lunell) usually gtows in open, Swampy or marshy habitats. Although perhaps of wider distribution, literature Teports and specimens examined establish its presence from Prince Edward Island westward to North Dakota and south to Nebraska and Indiana in the Midwest, and to the Ridge and Rydb., in contrast, 1971] MILLER, GENERA OF URTICACEAE 57 Valley Province of Virginia, and along the Coastal Plain of North and South Carolina, southeastern Georgia, to northeastern Florida in the East. These species are of similar aspect and differ mainly in fruit characters. In P. fontana mature achenes (FIGURE 3, 0) are black or dark brown, only slightly less broad than long, and often have a narrow, translucent edge and prominent low bosses that do not coincide with a color pattern, should one exist. Fruits of P. pumila (Figure 3, n) are much narrower, light brown in color, and have a contrasting mottling that appears only slightly raised in relation to the unmottled achene surface. Pilea has a unique type of fruit dispersal in which the achene is force- fully ejected from the enlarged, but non-adhering, perianth by staminodia. This phenomenon has been investigated in two tropical species, P. Spruce- ana Wedd. (Mosebach) and P. stipulosa Miq. (Goebel) and should be studied in our native species as well. For example, the distance the achene is propelled from the parent plant is a critical yet unknown fact in our plants. The mechanism (Mosebach) involves the simultaneous build-up of tension in the three inflexed staminodia, a result both of their anatomy and their being restricted to the confined space beneath the developing fruit. The central cells of the staminodia are larger (to 560 zu long) than the cubical cells of the adaxial epidermis (90 «) and the longer spindle-shaped abaxial epidermal cells (ca, 275 w long). When intercell pressure reaches the critical point during the elongation phase of growth, the fruit breaks off the receptacle and is hurled out of the perianth as the staminodia pop outward. (See Ficure 3, e, k-m.) An adhesive substance produced at the tips of the staminodia apparently helps to hold them under the fruit and prevents their premature extension. Only a few chromosome counts (9 spp., 27 = 16, 24, 36, 48, and 52) have been reported for the genus, and, with two exceptions, all for species occurring beyond our region. At least five species, including P. Cadierei Gagnep. & Guillaumin, 2n = 48, the aluminum plant, are cultivated as ornamentals. REFERENCES: Under family references see BERNBECK, Heywoop, Horm, Kite, Laketa & CRAIGHEAD, MartTIN, MOosEBACH, RIDLEY, SATAKE, SAYEEDUD-DiIn & SALAM, SHARMA, SOLEREDER, TARNAVSCHI et al., THORNE (1954), WEDDELL (1856, 1869), and WERNER. Apams, C. D. Notes on Jamaican flowering plants I. New species and varieties in Pilea. Mitt. Bot. Staatssam. Miinchen 8: 99-110. 1970. Cappel, E. D. A study of variation within the genus Pilea (Urticaceae) of temperate North America. (Abstr.) Diss. Abstr. Int. 30: 3046-B. 1970. [P. fontana considered a variety of P. pumila; ecotypes recognized in P. pumila; P. pumila (and P. fontana?), 2n = 16; studies on achene dis- persal. Fawcett, W., & A. B. RENDLE. Flora of Jamaica. Vol. 3. xxiv + 280 pp. 5 pls. London. 1914. [42 spp. of Pilea, 60-81, including P. herniarioides, P. microphylla, 61-64, pl. 1.] | ae JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 * Shs Fic . Pj oe : ; widvein eb nes : Bs iy microphylla: a, partially cleared leaf with cystoliths, ’ with inframarginal hvdath ary staminate and carpellate inflorescences, leaves d, carpellate ti al hydathodes, 10; c, post-anthesin staminate fl : , wer, X 40; e, nearly mature achene and enlarged perianth (one 1971] MILLER, GENERA OF URTICACEAE 59 FERNALD, M. L. Contr. Gray Herb. 113. II. Pilea in eastern North America. Rhodora 38: 169, 170. pl. 413. 1936. [P. pumila var. Deamii GOEBEL, K. Morphologische und biologische Bemerkungen. 27. Schleuderfriichte bei Urticifloren. Flora 108: 327-336, 1915. [Fruit dispersal in P. stipulosa.] HERMANN, F. ia The geographic distribution of Pilea fontana. Torreya 40: 118-120 Kixiip, E. P. gi ew species of Pilea from the Andes. Contr. U. S. Natl. Herb. 26: 367-394. 1936. [Includes keys to species described in Killip (1939).] he Andean species of Pilea. Ibid. 26: 475-530. 11 pls. 1939. [112 spp., including P. herniarioides and P. microphylla.] Leén, Hno., & Hno. Atatn. Flora de Cuba. Vol. 2. Contr. Occas. Mus. Hist. Nat. Col. La Salle 10. 456 pp. 1951. [Pilea, 59-67. LUNELL, J. Adicea. Am. Midl. Nat. 3: 6-12. 1913. a Adicea fontana, A. opaca, A. Nieuwlandii, and A. Deamii, spp. n Morton, C. V. New species of Pilea from Cuba. ea Soc. Cuba. Bot. 2: 2-10. 1945. [Seven new spp.; generalized distributional data for the genus in the West Indies; notes on P. microphylla. | Raprorp, A. E., H. E. AHLES ,& C. R. Bett. Manual of the vascular flora of the Carolinas. lxi + 1183 pp. Chapel Hill, North Carolina. 1968. [P. fontana in North and South Carolina, 395.] Tribe BOEHMERIEAE Gaud. Subtribe Euboehmeriineae Wedd. 4. Boehmeria Jacquin, Enum. Syst. Pl. Ins. Carib. 9. 1760. Erect, fibrous, perennial herbs [or shrubs to small trees] of swamps, marshes, lake margins, stream banks, and, less frequently, drier situations. Roots fibrous, sometimes thick and fusiform; plants spreading by rhizomes. Stems weakly fluted, glabrous to densely pilose. Leaves decussate or al- ternate, if decussate, then sometimes alternate on branches, blades broadly ovate to narrowly elliptic, triplinerved, margins serrate or dentate [or shallowly incised], upper surface smooth to strongly scabrous with rigid, forward-pointing hairs, lower surface pilose to white tomentose; cystoliths punctiform, restricted to upper epidermis; stipules lanceolate, brown, usually caducous, and either paired at the petiole base, or intrapetiolar with the apex bifid. Plants monoecious [or dioecious]; inflorescences axillary and bracteate, either + erect and spicate with carpellate and Staminate flowers occurring mixed in remote to congested clusters borne along the main inflorescence axis, the individual staminate flowers sur- rounded by groups of carpellate flowers within one cluster, or, paniculate tepal removed), X 40; f, mature achene, X 40. g-n, P. pumila: g, leaf, adaxial surface of blade, X 1; h, a xillary infructescence, staminate flowers represented 60 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Ficure 4. Boehmeria. a-i, B. cylindrica: a, plant apex with axillary inflores- cences, X 4; b, cystoliths at edge of adaxial leaf surface, X 10; c, cluster of mi from inflorescence, X 8; d, staminate flower > &, Dp 40; g, mature fruit covered by persistent perianth, X 20; ion, endospe hite, endocarp hatched (semidia- grammatic), X 20; i, fruit ™m Ww 1, fruit in cross section, endosperm and endocarp as in h (semidiagrammatic), < 20, and composed of either stamin inflorescences borne lower on soon after anthesis. stamens 4; rudimenta 1971] MILLER, GENERA OF URTICACEAE 61 around the edge, or, a + ovoid, hard-walled achene; in both cases the perianth and withered style persistent |perianth strongly inflated in some}. Endosperm fleshy, sparse; cotyledons ovate, longer than the radicle, not emarginate. (Including Ramium Kuntze.) Type species: B. ramiflora Jacq.; see Enum. Syst. Pl. Ins. Carib. 31. 1760; Select. Stirp. Am. Hist. 246, 247. pl. 157. 1763. (Named in honor of George Rudolph Boehmer, 1723-1803, professor of anatomy and botany at Wittenberg University, Saxony.) — FALsr NETTLE. A genus of about 100 species, mostly confined to tropical and subtropical latitudes in both New and Old Worlds (poorly represented in Africa), but with species native also to temperate portions of the Northern and Southern Hemispheres; absent from Europe; Australia and New Zealand with one species each. Two species occur in the southeastern United States. Boehmeria (subg. TILOCNIDE) nivea (L.) Gaud. (Ramium niveum (L.) Kuntze), ramie, 2n = 22, 28, and 42, a native of China, has been widely introduced in the tropics and subtropics as a source of fiber. It has been found as an escape in South Carolina, Florida, and parts of the Gulf Coastal Plain (also near Washington, D. C.), and may be naturalized in some of these areas. The native B. (subg. Durett1a) cylindrica (L.) Sw., bog-hemp, 2” = 28, is readily distinguished from ramie by its gla- brous to weakly pilose lower leaf surfaces (vs. densely white tomentose) and spicate inflorescences (vs. paniculate). Boehmeria cylindrica (includ- ing var. Drummondiana (Wedd.) Wedd., and probably B. decurrens Small described from specimens collected in northern Florida) ranges from southern Quebec and Ontario west to Minnesota, south to Florida (throughout) and across Nebraska, Kansas, and Oklahoma to Texas. It is known also from southeastern New Mexico and Arizona, Mexico (Coa- huila and Tamaulipas to Morelos and Tabasco), Guatemala, Honduras, Costa Rica, Panama, the Greater Antilles, Bermuda, Trinidad, and as a disjunct in southeastern Brazil, Uruguay, Argentina, and Paraguay. In North America, B. cylindrica is extremely variable in petiole length, leaf blade shape, roughness of upper blade surface, and character of the inflorescence. Forms with rigid, strongly scabrous leaves and short petioles are sometimes separated as var. Drummondiana. Although apparently less frequent in the north, it is largely coextensive with var. cylindrica but seems to prefer more open habitats. Transplant studies would help to establish whether the variety is merely a sun-form, as has been suggested. Subgenera were defined by Blume (Mus. Bot. Lugd. Bat. 2: 194~227. 1857) on the basis of a combination of vegetative, floral, and fruit char- acters, especially the form of the perianth surrounding the mature fruit. In a treatment of the Asiatic members of subg. DuretT1a Blume, Satake proposed seven sections based on shape and pubescence of the perianth. The usefulness of this classification on a world-wide basis remains to be determined by monographic study. Chromosome numbers for species investigated to date are 2n — 22,28, 42, and 52, with 28 and 42 the most frequently reported counts. Okabe 62 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 were triploids (2n = 42) which, upon embryological study, were found to taken into account the occurrence of apomixis (cf. Satake, who recognized 28 species in subg. Duret1a, many of which are geographically restricted endemics, with Ohwi, who accepts only 10), Hooked hairs on the outside of the fruiting perianth in B. cylindrica probably facilitate dispersal. The fleshy, drupaceous fruits of this species may also be eaten by birds. The genus is of little economic importance except in China, Formosa, Japan, and the Philippines, where ramie or China grass is grown com- mercially. Among the several cultivated varieties, B. nivea var. tenacissima fibers are difficult to separate from abundant associated gums, Willimot usehold linens, lace, furnish- t rubber tires, sewing thread » 4S mantles, paper for bank notes, and packing commercial use were obtained between 1935 and 1943 from this region. t eparation are devised, high labor costs will ep ramie a minor fiber plant in the United States, REFERENCES: Under family references see iy > L BEN, Kummer, LakeLa & CRAIGHEAD, Martin, 1971] MILLER, GENERA OF URTICACEAE 63 MopiLewsky, RIDLEY, SATAKE, SHARMA, SOLEREDER, TARNAVSCHI et al., THORNE (1954), WEDDELL (1856, 1869), and WERNER. AvpaBa, V. C. The structure and development of the cell wall in plants I. Bast fibers of Boehmeria and Linum. Am. Jour. Bot. 14: 16-24. 3 pis. 1927. [B. nivea Betisario, M. R. A comparative study of four varieties of ramie. Philip. Agr. 32: 185-214. 1949.* DeoprKar, G. B., & S. D. ParwarDHAN. Wild ramie Boehmeria platyphylla D. B. from Western Ghats —a possible source of textile bast fibre. Curr. Sci. Bangalore 28(1): 29, 30. 1959.* Dewey, L. H. Ramie, a fiber-yielding plant. U. S. Dep. Agr. Misc. Circ. 110. 12 pp. 1929. FERNALD, M. L. Note on Boehmeria cylindrica var. Drummondiana. Rhodora 12: 10, 11. 1910. Hitz, A. F. Economic botany. ed. 2. xii ++ 560 pp. New York. 1952. [B. nivea, ae Hotm, T. Boehmeria cylindrica (L.) Sw. —a morphological study. Am. Jour. Sci. V. 13: 115-122. 1927. Joyner, J. F., E. O. Ganestap, & C. C. SEALE. Ramie — Chinese silk plant. Gard. Jour. N. Y. Bot. Gard. 3: 18, 21. 1953. [Economic botany. ] Kunpv, B. C., & N. S. Rao. The shoot apex of Boehmeria nivea during morpho- genesis. Cellule 58: 217-228. 2 pls. 1957. Supir SEN. Origin and development of fibres in ramie (Boehmeria nivea Gaud.). Proc. Natl. Inst. Sci. India B. 26(Suppl.): 190-198. 2 pls. 1960 Maekawa, F. Nyctinasty in Sanicula, Boehmeria etc. (In Japanese.) Jour. Jap. Bot. 29: 286, 287. 1954. Oxwt, J. Flora of Japan. (In English.) F. G. Meyer & E. H. WALKER, eds. ix + 1067 pp. 17 pls. Washington, D. C. 1965. [Boehmeria, 390-392. OxaseE, S. Chromosome numbers and apomixis in Boehmeria. (Abstr.; in Jap- anese.) Jap. Jour. Genet. 31: 308. 1956. . Cytological studies of the apomixis in angiosperms I. Apomixis in the genus Boehmeria. Sci. Rep. Téhoku Univ. Biol. 29: 207-215. 1963. RaBEcuautt, H. La ramie — études morphologique et taxonomique en vue de la sélection. Théses, Doct. Sci. Univ. Paris No. 532 (Sér. A. 508). 133 pp. 33 pls. 51 photogr. 1951. Rosrnson, B. B. Ramie fiber production. U. S. Dep. Agr. Circ. 585. 14 pp. 1940, RoyaL Botanic GARDENS, Kew. Selected papers from the Kew Bulletin. I. Vegetable fibres. (Second reprint.) Bull. Misc. Inf. Kew Addit. Ser. 2. 280 pp. 1912. [Ramie, 52-95.] SATAKE, Y. Boehmeria Japonica. Jour. Fac. Sci. Univ. Tokyo Bot. 4: 467-542. 1936. [40 spp. in 2 subg. and 7 sects., mainly Japanese but some spp. from other parts of Asia; also data on petiole anatomy. Scott, F. M. Cystoliths and plasmodesmata in Beloperone, Ficus, and Boeh- meria, Bot. Gaz. 107: 372-378. 1946. [B. nivea.] SEALE, C. C., E. O. Gancstap, & J. F. Joyner. Agronomic studies of ramie in the Florida Everglades. Florida Agr. Exper. Sta. Bull. 525. 30 pp. 1953.* Supa, S. Features of male and female flowers on the stalk in ramie. (In Jap- anese; English summary.) Bull. Miyazaki Coll. Agr. Forestry 3(1/2): 9-17. 1958.* 64 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 StncH, D. The ramie in Florida. Indian Farming 6: 27-29. 1956.* Situ, C. E., & T. Kerr. Pre-conquest plant fibers from the Tehuacan Valley, Mexico. Econ. Bot. 22: 354-358. 1968. [Yarn made from urticaceous bast fiber, probably Boehmeria sp. | Stusss, W. C. Ramie (Boehmeria nivea). Uses, history, composition, cultiva- tion, etc. Bull. Agr. Exper. Sta. (Baton Rouge) 32: 1126-1146. 1895.* Wittimor, S. G. Ramie fibre — its cultivation and development. World Crops 6: 405-408. 1954. Tribe PARTETARIEAE Gaud. 5. Parietaria Linnaeus, Sp. Pl. 2: 1052. 1753; Gen. Pl. ed. 5. 471. 1754. Erect, ascending, sometimes decumbent, often multi-stemmed, taprooted annual or perennial herbs [or subshrubs] of moist forests and shaded ground. Stems in annual species translucent, watery, sparsely to densely pubescent with soft, sometimes hooked hairs; stems usually somewhat woody in perennial species. Leaves petiolate, pinnately to + palmately veined, opposite in seedlings, alternate in mature plants; blades lanceolate, ovate, or deltoid, generally pubescent, margins entire; cystoliths dense, punctiform [rarely otherwise], restricted to the upper leaf epidermis; stipules absent. Plants polygamous; inflorescences few-flowered axillary cymes, one cymule borne on each side of the petiole base, flowers subtended y 1 to 3 green, linear [or deltoid and somewhat foliar], often ciliate [and sometimes basally-fused] bracts, usually bearing cystoliths; basal (first opening) flower in each cymule often bisexual, others generally carpellate; staminate flowers also said to occur, but apparently rare or absent (?) in American species, Carpellate flowers sessile with 4 equal, pubescent tepals, fused toward the base, but free above; ovary short- stipitate or not, attached to a small disklike receptacle; style central, apical or slightly displaced from the ovary tip, short or long, stigma penicillate [or linear]; ovule orthotropous (?), micropyle formed by both integuments. Bisexual flowers apparently proterogynous; tepals 4, equal; stamens 4, anther sacs fallin A genus of about 20 species in two subgenera, occurring mainly within temperate and subtropical | : atitudes, but occasionally also in the tropics, where often restricted to montane habitats. Subgenus Parterarta (subg. 1971] MILLER, GENERA OF URTICACEAE 65 VS FicureE 5. Parietaria. a, b, P. floridana: a, habit of young plant, x %; b, achene (note flanged stipe), X 25. c-h, P. praetermissa: c, adaxial surface of leaf apex with punctiform cystoliths, 10; d, i perfect the perianth accrescent, and two carpellate flowers, X 10; e, perfect flower just after pollen discharge (note absence of stigma, cf. f), 12; f, carpellate flower, 3,8, mature achene, X 25; h, embryo, X 25. i, j, P. pensylvanica: i, leaf, 5 X 1; j, mature achene, < 25. Euparietaria Komarov ex Jarmolenko) has long, filiform styles supporting a brush-like stigma, and is comprised mostly of perennials that occur from western Europe and northwestern Africa eastward to central Asia (al- though Parietaria macrophylla Robins. & Greenm., a species of south- western Mexico, would seem to belong here also). Two members of this Subgenus, P. officinalis L.> and P. judaica L., have been infrequently re- *Specimens named Parietaria officinalis from Sanford, Seminole County, Florida [shaded places, Rapp, April 11, 1929 (ny), sandy, waste places, Rapp, July 1929 (Nx)] on which Small probably based his statement, “sandy pinelands, old fields, and adsides, pen. Fla.” (Man. Southeast. Fl. 435. 1933), actually represent an apparently unreported adventive in the United States, Pouzolzia zeylanica (L.) J. Benn., a native of Asia that ranges from Japan to Formosa, China, Indochina, Malaysia, and India has stipulate leaves and typically a large taproot. Carpellate and staminate flowers subtended by small, scarious bracts are aggregated in tight axillary clusters, and the 66 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 ported as adventives at places in the Southeastern United States and else- where in North America. Both have black achenes, in contrast to the light to dark brown fruits of our native species. Subgenus FREIREA (Gaud.) Komarov ex Jarmolenko, represented in both the New and Old Worlds, has a sessile stigma or one supported by a very short style. Most species in this subgenus are annuals. The significance of the difference in loca- tion of carpellate and bisexual flowers within an inflorescence, used by Jarmolenko (1936) as another distinguishing feature between the two subgenera, needs confirmation. Africa within the range of variation of the widespread P. debilis Forst. f.). The most reliable taxonomic characters appear to be shape of the achene and form of the accrescent perianth of both carpellate and bisexual flowers. Dispersal of certain European species by ants attracted to the achenes by the contents of the basal stipe, which they utilize as a food source, is mentioned by Ridley. filiform styles of carpellate flowers dro i isting i : p off soon after anth : Boehmeria). At maturity, the black, shiny, ovoid ache fee Oca ua Posie: ic taria officinalis) are tightly enclosed by a thick, dry, continuous perianth bearing cae . S. é peje the Sanford plants have persisted, and therefore can be considered naturalized, remains to be determined, as ; or? seine » aS does the status of Small’s report ° Treatment of these two Species follows Hinton (196 : 8). The t eci te nummularia and P. floridana represent the same saieer Danie onesie 1971] MILLER, GENERA OF URTICACEAE 67 Chromosome counts of 2n = 14, 16, and 26 have been reported. The genus is of no economic importance, except possibly for the infre- quent use of P. officinalis medicinally. Expressed juice and boiled extracts obtained from this species have been used as diuretics and to lower fevers. REFERENCES: si Raed references see HEYwoop, KoLUMBE, KRAUSE, KUMMER, LAKELA & CRAIGHEAD, Martin, Mopitewsky, Rw IDLEY, SATAKE, SCHREIBER, SHARMA, TiN ctr et al., WEDDELL (1869), and WERNER. Decaux, F. Une antique panacée: La pariétaire ou perce muraille (Parietaria officinalis D.C.). Revue Phytothérap. 7: 117-120. 1943. [Medicinal uses.] Hepserc, O. Afroalpine vascular plants, a taxonomic revision. Symb. Upsal. 15. 411 pp. 12 pls. 1957. [Parietaria, 73, 74, 268. Hinton, B. D. The native annual Parietaria (Urticaceae) of the conterminous United States and Canada. M.S. thesis; Univ. of Southwestern Louisiana. 68.* ——. Parietaria praetermissa (Urticaceae), a new species from the south- eastern United States. Sida 3: 191-194. 1968. . Parietaria hespera (Urticaceae), a new species of the southwestern United States. Jbid. 293-297. 1969. [P. hespera var. hespera and P. hespera var. californica; notes on P. pensylvanica. JARMOLENKO, A. V. Urticaceae. Fl. URSS 5: 384-405. 1936. [ Parietaria, 2 subg. 6 spp., 398-405; 315-320 in English transl., N. Lanpav, Fl. U.S.S.R. 5. xxvil + 593 pp. Jerusalem. 1970. An attempt at the construction of a system of the genus Parietaria and related forms. (In Russian; English summary.) Acta Inst. Bot. Acad. Sci URSS. 1. Syst. 5: 319-330. 1941. [Parietaria, sensu stricto (5 spp. inchud- ing P. officinalis), and Freirea Gaud. (13 spp., including P. pensylvanica and P. floridana) based on inflorescence and stigma differences; numerous in- frageneric categories LEBEGUE, A. Embryogénie des Urticacées. Développement de l’embryon chez le Parietaria officinalis DC. Compt. Rend. Acad. Sci. Paris 243: 817-820. 1956. [Embryogeny in Parietaria and Urtica differ somewhat; additional evidence for their assignment to two different tribes. ] McVaueu, R. Suggested. phylogeny of Prunus serotina and other wide-ranging phylads in North America. Brittonia 7: 317-346. 1952. [Generalized dis- tributional data on Parietaria in the southwestern U.S. and northern Mexico (species not sariobeues ene 330, Map 11 p. MILtLrr, N. G. new species of Parietaria (Urticaceae) from Mexico. Jour. Arnold Arb. 51: 529-533. 1970. [P. decoris from Coahuila, Nuevo Leén, Tamaulipas, and San Luis Potosi.] Navas B., E. Monografia sobre las ri nape de Chile. Revista Univ. Catdl. Chile 48: 9-17. 1 pl. 1963 [1964]. [4 sp Pactt, J. Nachtrag zu meiner Pucaneds Baas. Phyton ee 5: 242-246. 1954, [Critique of Jarmolenko’s subgeneric classificatio Prreyre, N. Action du strontium sur la croissance de la Pais 0 Hien L.. et le développement de ses cystoliths. Compt. Rend. Acad. Sci. Paris 250: 579-581. 1960. [Includes spectrographic analysis of stalis) SEGAL, S. Ecological notes on wall vegetation. 325 pp. 4 tab. 3 append. The: Hague. 1969. [European Parietaria spp. mentioned hinted 68 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 SHINNERS, L. H. Notes. Field Lab. 18: 42. 1950. [P. pensylvanica var. obtusa (Rydb. ex Small) Shinners. ] Suxopp, H., & H. Scuouz. Parietaria pensylvanica Miihlenb. ex Willd. in Berlin. Ber. Deutsch. Bot. Ges. 77: 419-426. 1964 [1965]. [Widely established in weedy ae includes line drawing of plant. Tureret, J. W. Twenty-five eg of vascular plants new to Louisiana. Proc. Louisiana Acad. Sci, 32: 78-82. 1969. [P. judaica L. (as P. diffusa Mert. & Koch, see Townsend), a native of Europe, nw. Africa, and the Near and Middle East, on ballast, Port Eads, Plaquemines Parish, 79. | TOWNSEND, C. C. arietaria officinalis and P. judaica, Watsonia 6: 365-370. 19 4 oe nomenclature, and taxonomy; illustrations of both speci Woop, H ¢; Ja., Co, LaWALt, et al. The dispensatory of the United States of America. ed. 22. xix + 1894 pp. (Supplement, iv + 76 pp.) Philadelphia & a 1940. [P. ce eealis 1508] ARNOLD ARBORETUM Present address HARVARD UNIVERSITY DEPARTMENT OF Borany UNIVERSITY OF NORTH CAROLINA CHAPEL HI, N. C. 27514 1971] NEVLING & ELIAS, CALLIANDRA 69 CALLIANDRA HAEMATOCEPHALA: HISTORY, MORPHOLOGY, AND TAXONOMY Lorin I. NEVLING, JR., AND THOMAS S, ELIAS ONE OF THE STRIKING ORNAMENTAL TREES Of tropical and subtropical climates is the leguminous powder-puff tree. Since its description more than a century ago it has become a widely distributed ornamental plant. The curious background associated with this species has led to considerable taxonomic and nomenclatural confusion over its correct scientific name. As a result of general interest in the genus Calliandra, particularly in its floral biology, we began a routine re-evaluation of C. haematocephala Hassk. and C. inaequilatera Rusby, the two scientific names commonly associated with the powder-puff tree. Our conclusions differ in several significant respects from those of previous investigators, and we present them in the hope that we have clarified a difficult systematic problem. One of the most critical problems in equating Calliandra haemato- cephala Hassk. with Calliandra inaequilatera Rusby has been that of geographic distribution. Calliandra haematocephala was described by Hasskarl from material in cultivation at the botanic’ garden in Java. In Hasskarl’s original publication he cited his material as coming from the botanic garden at Calcutta under the name of “Jngae haematoxyli.” The origin of the material in Calcutta was reported as unknown. This “spe- cies” had not been found in the wild, although it had been widely and correctly assumed to be a plant of American origin The type material of Calliandra inaequilatera was described by Rusby from material collected by Miguel Bang in Bolivia near the town of Guanai. This locality is also spelled Guanay, Huanai, or Huanay and is apparently critical in the resolution of the problem (see Ficure 1). Guanai lies between the fork of the Rio Tipuani and the Rio Mapiri at an altitude of about 700 meters. It was established by the Spanish as a mission and as a place of exile for political prisoners. At the confluence of the Tipuani and Mapiri the river becomes the Rio Kaka, which soon joins the larger Rio Beni. One of the difficulties in developing evidence for the conclusion that the two names apply to the same plant has been the failure to resolve the geographic disparity between Bolivia and Java. It seems unlikely that a Bolivian Calliandra would have made its way to Java via Calcutta entirely on its merits as an ornamental shrub by the year 1855 (the date of Hasskarl’s publication). We believe that such a “migration” can be explained only through an understanding of the outstanding botanical problem of the times, namely Cinchona. The de- mand for Cinchona bark was tremendous at the turn of the 19th century, and field botanists of the era were sending an alarm to their respective governments. Exploitation of the bark of the great Cinchona trees was advancing at a much too rapid rate. By 1830 the governments of Britain, JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Racial BRAZIL \ Islay } ae, E PARAGUAY be a ay ae ; > ARGENTINA tery @ Ixiamas Rurrenabaque Mapiri o Tipuani San Miguel de Huachi Cochabamba Buena Vista / Santa Cruz Ficure 1. Map showing important localities and rivers significant in some historical aspects of Calliandra haematocephala. Inset is enlargement of stippled area. France, and Holland began to believe the fiel cerned that their bark source wo ing for high-yielding bark was un because of lack of understanding 1971] NEVLING & ELIAS, CALLIANDRA 71 was to isolate a species that would yield high percentages of sulphate of quinine. At approximately the same time, the Bolivian government re- alized the threat to her Cinchona resources, which were a primary source of income from the outside, and began to take repressive steps to prevent the over-exportation of Cinchona bark. In 1845, it granted a five-year monopoly to the Spanish firm of Jorge Tesanos Pinto and Company in return for $119,000. The agreement limited the amount of bark to be taken. In 1849, by presidential decree, the monopoly was broken because of low wages paid to the bark gatherers (cascarilleros). Several subse- quent monopolies were established, but all collapsed in a few years be- cause of financial difficulties. In spite of the series of impending monop- olies, or perhaps because of them, the major powers banded together in various attempts to get seeds or seedling plants of Cinchona from various parts of the New World. One of the main objectives was to be Cinchona calisaya Wedd. (= C. officinalis L. fide Standley). This species was known from southern Peru and Bolivia, but the exact limits of its dis- tribution were unknown. Accordingly, in 1843, the Count de Castelnau, the Viscount d’Osery, and Dr. H. A. Weddell entered Bolivia via an over- land route through Brazil to collect material of Cinchona and particularly to establish the southern geographical limit of Cinchona calisaya. Through a series of mishaps Dr. Weddell was eventually separated from the other members of the party; the Count returned to Europe, while the Viscount disappeared into the upper Amazonian Basin, never to be seen again. Weddell explored the Bolivian Cinchona forests from Cochabamba through Ayopaya and Enquisivi. In 1847, he entered the province of Caupolican, went down the Rio Tipuani, visited the mission of Guanai (the type locality of Calliandra inaequilatera), and then went up the Mapiri. In June o that year, he entered the Peruvian district of Caravaya and went down to Cuzco. Weddell met a Mr. Delondre in Cuzco and, after several days of pleasant but guarded conversation, discovered that he also was searching for Cinchona calisaya and, further, that he was a manufac- turer of sulphate of quinine. Foreseeing difficulties with the government of Bolivia, Delondre had disassembled a factory for the manufacture of quinine and brought it by ship to the town of Valparaiso where he re- assembled it. However, because of the strict limitations put on by the Bolivian government he was unable to obtain bark legally. He had con- tacted Pinto and Company, but they had sold their total output to a firm in New York for processing. Weddell and Delondre collected in the valley of Santa Ana. They re- turned to the coast late in 1847, and Weddell went on to France. In Paris, Weddell’s Cinchona seed lot was apparently divided, with some seeds being sent to England. The English sent their seeds on to Calcut- ta for growing, in an attempt to establish a quinine industry in India. After the seeds germinated in Paris the French turned over seedlings to the Dutch who, in turn, sent them directly to Java to start their planta- tions. The remaining seedlings were sent to Algeria. It seems to us quite possible that Weddell would have collected material other than that of 72 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Cinchona and quite possibly could have collected the Calliandra because of its extraordinary beauty and potential use as an ornamental plant. This material could have gone the route from Bolivia to France to England exchanged regularly. In addition, part of the Weddell seed collection ap- parently was turned over to 4 Parisian horticultural firm, since the it in 1851. Seeds of ornamental Species may have followed the same route. Leiden sent the Cinchona to Java where it arrived in 1852. In 1849 Weddell published a classification of Cinchona based on his field studies. He was, as a result better qualified in 1851 when he re- turned to South America. In 1852, he entered the Bolivian Cinchona ranged at the end of the first one. In 1852, the Dutch Minister of Colonies decided that a Cinchona grow- ing area should be established in the colonies, Charles Hasskarl (1811-1894) a botanist who ha » Hasskarl sailed from The Hague on December 4, 1852, charged “to ¢ : collected by a native near Ychu-corpa on the Bolivian fronti Once he had the Islay, which is just north of the town of Mollendo. He had sent ahead word for help, and the Dutch warshi j j in Islay. The Plants were placed in Wardian Cases, and the ship set sail August 27, 1854. It was routed directly to Java and arrived December 1971] NEVLING & ELIAS, CALLIANDRA 73 13, 1854. It is widely reported that only two Cinchona plants survived, but a contradictory report indicates that there were 75. The latter num- ber is probably correct. On his return to Java Hasskarl was put in charge of the Cinchona cul- tivation project, and in connection with this he received additional Bolivian Cinchona seeds from Paris, presumably some of those collected by Weddell in 1852, and seeds of other species were sent from Peru. Seeds collected by Karsten in Colombia were received through the Gov- ernor of Curacao. The significant point is that Hasskarl was indeed in Bolivia and in the general vicinity of the locality for Calliandra in 1854. His publication describing the newly cultivated Calliandra was in print by the end of the year of 1855. We think that the evidence, although highly circumstantial, indicates that Hasskarl may have collected the plants in Bolivia, have seen them flower aboard ship in Wardian cases, and at that time may have written his very extensive Latin description. This technique, i.e., description aboard ship, was well known to botanists of that era. Alternatively, it is possible that he did indeed obtain his ma- terial from Calcutta, and this, in turn, may have been received from the Weddell expeditions; or perhaps Hasskarl wished to cover his tracks in Bolivia and ascribed the plant to Calcutta without receiving seeds from there. In 1860, the English botanist C. R. Markham was sent to Bolivia to collect additional material of Cinchona. He arrived on April 19th near the town of Crucero in the province of Caravaya and met there a Peru- vian, Don Manuel Martel. Don Manuel “had a good deal to say, not very complimentary, about the Dutch agent who had come over to obtain Cinchona plants in 1854 for cultivation in Java.” The bad feel- ing generated by Hasskarl affected Markham’s expedition, and it was remarkable that he was able to bring his plants and seeds out of Bolivia safely. It should also be mentioned that the Bolivian agent with whom Miiller-Hasskarl had dealt was forced to remove himself to another part of the country because of the threats and in fact, acts of physical vio- lence by the local populace. It seems quite possible, then, that the ma- terial that was described either on shipboard or from cultivated material in Java was obtained through one of the tortuous routes described. In postscript it should be mentioned that because of low yields none of the materials collected by these expeditions formed the basis for the rich Java Cinchona plantations. The honor of introducing a high-yield- ing strain fell to Charles Ledger or, perhaps more correctly, to his ser- vant, Manuel Incra Mamani, who collected fourteen pounds of seeds of high-yielding Cinchona. One pound was sold to Holland for 100 francs plus £24 on germination. The remaining thirteen pounds of seed were partly sold on the streets of London, and the residue was purchased by a Mr. Money, of India, but on his return to India this lot failed to ger- minate. After all of the preceding efforts it seems ironical that the Profitable Dutch Cinchona industry was founded on the purchase of a single pound of seed. 74 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 The history of the introduction of Calliandra inaequilatera into culti- vation is almost as convoluted as that of Calliandra haematocephala but fortunately is better documented. Again the geographical focal point is lieved that he had arrived in Guanai approximately “three weeks too late,” since the forest was primarily in fruit. This seems to be a curious wholly of mimoseae in prodigious variety.” He said further that “the slope just back of the town was heavily forested, many of the trees be- ing of really gigantic size. They belong chiefly in the senna, mimosa, fact that most of the legumes ap- parently were in fruit, extensive collections were made, but most of these were lost in a boat mishap. In 1889, Rusby hired Mi him. Bang had been trained in gardening at Kew and had gone to 0 collect living orchids to send to England. ; Bang began collecting for Rusby. The her- the vicinity of Guanai. alliandra inaequilatera in 1896. The only, and no fruit was seen. vent was the Mulford Expedition for Biological azon Basin headed by Rusby (on his final Bo- de along the Rio Bop. wh; “ass : Beni 9p!, which drains into the Rio €nl above the confluence of the Rio Kaka. They established several 1971] NEVLING & ELIAS, CALLIANDRA 75 base camps, including ones at Huachi and Rurrenabaque; made extensive collections along the Rio Beni, and at Rurrenabaque they had a tem- porary reunion with Miguel Bang. We think that it was in the Rio Bopi-Rio Beni drainage that White collected seeds of Calliandra. Indirect evidence (see discussion of chromo- some numbers) points to the origin as “Huachi” (San Miguel de Huachi). Because of illness Rusby was forced to leave for New York in December, leaving White in charge of the faltering expedition. White continued collecting until March 15, when the expedition disintegrated, from the biological point of view, and he and some of his companions left for New York, via the Amazon River. Shortly after White’s return on April 14, 1922, the dried specimens were sent to Rusby at the New York Bo- tanical Garden, while White took the seeds to the Brooklyn Botanical Garden. Eleven years later the Brooklyn Botanical Garden exhibited a vase of flowering C. inaequilatera at the January 20th meeting of the Horticultural Society of New York and won an Award of Merit for the entry. One of the plants from the original seed lot, now a repeatedly pruned tree with a trunk eighteen inches in diameter, is still growing in the Brooklyn collection. According to Hayward (1955), the first record of the shrub’s intro- duction into Florida was three plants of Calliandra inaequilatera that were sent in 1933 to Dr. George Tyrell, a horticulturist at LaBelle, Florida. He, in turn, gave stock to Dr. E. A. Menninger, who is re- sponsible for its wide distribution and status as a desirable ornamental for tropical climates. The primary method of propagation of the plant is by the seed (Menninger, in Jitt.), although the progeny are variable and not reliable. The plants seem to be quite variable in regard to the num- ber of fruits set, and some plants seldom fruit. The principal horti- tural selections have been made on the basis of the color of the staminal filaments, the most conspicuous parts of the showy heads of flowers. Recently (1963), Cowan in reviewing the taxonomy of the cultivated powder-puff trees concluded that two species, Calliandra haemato- cephala and C. inaequilatera are being grown. He presented keys, de- scriptions, and discussion of each species. His maintenance of the two as distinct species was based, not on opposing qualitative characters, but on variations of the same quantitative characters, usually with consider- able overlapping. We have carefully re-examined the vegetative charac- ters, including petiole length, terminal leaflet size, and degree of leaflet pubescence, which supposedly were useful in distinguishing the two species. The petioles of C. haematocephala were reported to be 10-20(-25) mm. long, while those of C. inaequilatera were (15—)25-40(—50) mm. long. In addition, his study indicated that the median and terminal leaflets of C. haematocephala are smaller, [20-]23-47 mm. long by [5—]7-17 mm. broad, and more numerous than those of C. inaequilatera, which are [45—] 50-75 mm. long by [15—]20-30 mm. broad. Our own study shows that petiole length, a highly variable character, is independent of leaflet size 76 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 and number, with both short to long petioles being found on specimens attributed to each of the two “species” in question, Although selected elements of Old World cultivated “populations” of C, haematocephala can be separated on the basis of leaflet size from similar selected elements of C. inaequilatera, there is complete gradation between the two on this character. Indeed, many specimens have intermediate leaf- let sizes. Examination of living plants substantiated the unreliability of this character, for leaves with small leaflets can be found on the same branch with leaves with large leaflets (see Ficurr 2). The Old World specimens typically have the leaves glabrate and ciliolate, whereas New World ones have varying amounts of leaflet pubescence. We suspect that this is due primarily to a more restricted genotype in the Old World ma- terial, probably because of a smaller original sample size. of the holotype and five isotypes of C. variation in the degree of pubescence of An examination inaequilatera showed considerable the leaflets. The floral characters. corolla-lobe size and position in age, length of the Ficure 2. Outli of f 2 4. Outlines of four leaves ¢ i aves f as > vegetativ t Calls dra haematocephala showing size varlition ee Pane of Colin 1971] NEVLING & ELIAS, CALLIANDRA 77 corolla tube, and peduncle length were also examined and found to be unreliable in distinguishing the two “elements.” Although a longer pe- duncle length is usually found in plants with smaller and narrower leaf- lets, no clear separation could be made, for the peduncular lengths of Calliandra haematocephala overlap with those of C. inaequilatera. Like- wise, the longer lanceolate corolla lobes attributed to C. haematocephala, as well as shorter, more ovate ones, could be found on specimens of C. inaequilatera. The length of the corolla tube, like the other characters examined, showed a continuous gradation from 3 to 7 mm. The position of the corolla lobes, erect versus reflexed, could not be adequately as- sessed. This condition does not appear to be genetically controlled. Some of the lobes becoming reflexed may result from the pressing techniques of the collector. In sum, the taxonomic characteristics of the flowers, as well as those of the vegetative parts, fail to differentiate two taxa after a careful analysis of a sizable sample of specimens, both wild and cultivated. The unreliability of the characters previously used to distinguish the two “species” of Calliandra is not in itself sufficient reason to treat them as a single species because of the peculiar circumstances that led to the plant’s being known to science in cultivation prior to its being known to science in nature. In the following paragraphs we point out a few pe- culiar morphological features which link the specimens into a single species and when considered in conjunction with the historical aspects serve to confirm further the morphological conclusions. On the totality of these considerations we believe that there is no justification for recognition of more than one species, the correct name of which is C. haematocephala Hasskarl, The most unique floral feature of Calliandra haematocephala is the structure of the androecium. As in all species of the genus, the stamens are basally connate into a staminal tube of varying length. In C. haema- tocephala the tube generally equals or exceeds the corolla lobes (see Fic- URE 3.) An irregular number of small fimbriae are inserted at the mouth of the staminal tube. These fimbriae, which have been referred to in previous literature as a “quasi corona,” are irregular in both size and shape (generally not more than 1 mm. in length) and often erose at their apices. They are derived from staminal tissue at the orifice of the fila- ment tube, and they may either represent aborted stamens or be irregular enations that are not vascularized. The fimbriae are oriented with their apices directed towards the center of the staminal tube orifice and form a somewhat dome-shaped structure covering it in functionally staminate flowers or permitting only the style to pass in bisexual flowers. They are white, as is the staminal tube, in stark contrast to the variously red- hued filaments. That such an abrupt color change occurs at the filament tube orifice implies more than casual biological significance. This sig- nificance probably can be determined properly only in field studies. It has been long recognized that in Calliandra, as in many other genera of the Mimosoideae, the pollen is found in “massulae,” or clumps. In Callian- dra two types of “massulae” are found: a sixteen-grained, disc-like polyad 78 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Ficure 3. Calliandra haematoce section of functionally staminate flower, ternal view of upper portion of filament tube sho tudinal section of ova psoidal polyad. The polyad of C. Aaema- Ficure 4). Although it has been figured at right angles to the axis of the fila- of the anther is directed away from the center 1971] NEVLING & ELIAS, CALLIANDRA 79 Ficure 4. Photomicrograph of fresh polyad of Calliandra haematocephala stained with aniline blue in lactophenol. The adhesive droplet in upper right- hand quadrat distorted from normal spherical shape because of attachment to microscope slide. Actual length of polyad ca. 160 pe of the inflorescence. Flowering proceeds from the base of the inflorescence to the apex but the process is so rapid that, for practical purposes, it may be considered synchronous. As soon as the filament straightens, the anther begins to open by two longitudinal slits. Within each of the four anther locules lie two yellow eight-grained polyads. Each flattened polyad is composed of two median pollen grains and six peripheral ones. One of the peripheral grains is modified slightly and has a somewhat eccentric and pronounced projection. This special grain, the foot grain, lies at the prox- imal end of each polyad and the two foot grains of the polyad pair are directed toward one another. All previous published illustrations of the eight-grained Calliandra polyad show this detail but not the orientation within the anther. The significant point that has not been mentioned is that between the apices of the “foot grains” an extremely viscous sub- Stance (that cannot be seen in acetylized material) is formed. As the anther opens the polyads are not only exposed, but the polyad pairs are separated, and their sticky apices are rotated until they are directed away from the anther by at least 90 degrees. The polyad bases are held in the old locules. The mechanism of this maneuver is not known, but it may be related to a rather large disc-like structure on the back of the anther that surrounds the filament. At anthesis the polyads are held with their apices directed outwards, and each apex is covered by a glistening drop of ad- 80 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 hesive material. Analogy to the pollinia of Asclepiadaceae and Orchidaceae is irresistible. In an inflorescence of some forty flowers (which seems to be an average number in an inflorescence of a well-grown C. haematocephala) there may be approximately 25 anthers per flower (varying between 20 and 30), making approximately 8000 polyads available for pollination. wards the end of the secon third day. Individual inflorescences vary in the total number of flowers and in their functional i coun and megasporogensis (lackin any men- tion of the pollination mechanism) has — cosa by veo 1958). He concluded that the eight-grained polyad of Calliandra haema- sobilateral tetrads. The individual pollen to Dnyansagar but Brewbaker (1967) has (in C, inaequilatera). Most likely one of 1971] NEVLING & ELIAS, CALLIANDRA 81 these reports is incorrect as a result of misidentification of the specimen. There are two chromosome number reports for Calliandra haemato- cephala, both made by Earlene Atchison from mitotic divisions. The first count, 2n = 16, was made from material grown at the Blandy Experi- mental Farm from Huachi, Bolivia (Jour. Elisha Mitchell Soc. 65: 118— 122. 1949). This count is based on the O. E. White material and provides the only precise information on the origin of the material introduced at Brooklyn. We believe this locality information came directly from Dr. White to Dr. Atchison. The second count, 2m = 16, was made from ma- terial grown at the Atkins Garden, Cuba (Am. Jour. Bot. 38: 538-546. 1951). An examination of the voucher specimen, Atchison 91 (GH), shows that it is not C. haematocephala but another species of the genus. The only other chromosome count reported for the genus is that for C. pittieri, which has 2n = 32 (Jour. Agri. Sci. Tokyo Nogyo Daigaku 8: 49-62. 1962). Calliandra haematocephala Hasskarl, Retzia 1: 216-219. 1855. Calliandra inaequilatera Rusby, Mem. Torrey Bot. Club 6: 28, 29. 1896. Anneslia haematocephala (Hasskarl) Britton & Wilson, Sci. Survey Porto Rico & Virgin Islands 6: 348. 1926. Shrubs or small trees to 5 m. high, frequently spreading and forming a ++ rounded crown; branches and branchlets terete, the latter with nu- merous inconspicuous lenticels, glabrate to densely puberulent, with minute multicellular glandular trichomes on most vegetative parts; conspicuous short shoots often present. Leaves alternate, bipinnate, eglandular, with one pair of pinnae, petiolate; petioles canaliculate, 0.7—3.25(—4.75) cm. long, slightly swollen at the base, sparingly puberulent to densely pilose, with a small apiculum between the insertion of the pinnae; rhachides 4.5—-11.5(—15) cm. long, canaliculate, sparsely to densely pilose, terminated by a small lanceolate apiculum; pinnae with (4—)5-8(—10) pairs of leaf- lets; the basal pair of leaflets reduced, subopposite, the inner leaflets smaller, the terminal leaflets narrowly ovate, elliptic, to obovate, 1.5—-8.4 cm. long, 0.53.5 cm. broad, + falcate, oblique and rounded at the base, obtuse to acute and mucronulate at the apex, glabrate to pilose, generally ciliolate, with 3 primary veins (2 prominent and | subprominent) palmate- ly arranged, chartaceous; stipules ovate to infrequently lanceolate, 2—10 mm. long, 1-3 mm. broad at the base, acute at the apex, persistent, often crowded on the short shoots. Inflorescences capitate, 20—45-flowered, borne singly or in pairs in the leaf axils, pedunculate, the peduncles erect, 1.2-3.3(-5.3) cm. long, pilose, often densely so. Flowers bisexual or functionally staminate by gynoecial abortion, 5-merous, pseudo-actino- morphic, sessile, bracteate, the bracts elliptic, to 1.5 mm. long, striate and ciliolate. Calyx tubular to subcampanulate, 1.5-3 mm. long, striate and sparsely puberulous to glabrate, greenish, the lobes ovate, 0.5—1 mm. long, Obtuse at the apex, ciliolate, erect, aestivation valvate. Corolla tubular and flaring slightly towards the orifice, 3.5-8 mm. long, glabrous or with 82 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 a few unicellular trichomes interspersed with minute, ae aa dular trichomes, reddish, the lobes lanceolate, 2—3 mm. long, ca. fete broad at the base, sparsely ciliolate, usually erect at see a wa valvate, Stamens ca. 25, basally connate into a staminal tube, the tu ae ally equalling or often surpassing the corolla lobes, with ee aed i eo larly shaped fimbriae, to 1.5 mm. long, inserted within the ee forming a dome-like structure at the tube apex, the filaments aa free at various levels, filiform, 3-4 cm. long, white at base to Ma 2 tensities of red towards apex; anthers oblong, attached on the bac dy a a disc-like structure surrounding filament at the point of agenesis oe hiscing by longitudinal slits directed away from base of ripen a united into 8-grained polyads, the polyads elliptic and flattene , 4 long, and ca. 88 » broad, with one grain eccentric and bearing a a pe adhesive material, each anther bearing 8 polyads. Disc cupuliform, 0. 0.75 mm. tall, fleshy, glabrous. Gynoecium 1-carpellate, the ovary ie gonal, 3-4 mm. long, glabrous, the style filiform, 34 cm. min r = glabrous, the stigma dome-shaped, ca. 0.5 mm. in diameter, ad — anthesis; ovules anatropous, 8; in functionally staminate flowers aborted gynoecium of various sizes, Legume elastically dehiscent, ree oblanceolate, 8-12 cm. long, 1-1.5 cm. broad, with a broad valve and ee row but conspicuous margins, brown, usually 5—6-seeded, the ee oblong, flattened laterally, 0.8-1.2 cm. long, 0.4—0.6 cm. broad, each si marked with a conspicuous pleurogram, hilum eccentric. SPECIMENS EXAMINED (SPONTANEOUS): Bolivia. La Paz: Vicinity of Guanai, July 1892, Bang 1586 (holotype ny isotypes A, GH, Mo, Ny); Ixiamas, Dec. 15, 1921, Cardenas 2011 (BKL, GH); Isapuri (on Rio Kaka) Oct. 5, 1901, Cruz: Buena Vista, April 1- 92 Steinbach 7154 (A, GH); Santa Cruz, May 1892, Kuntze s.n (us), June 1892, Kuntze s.n. (ny), Jul ¥ 2, 1968, H. Adolfo M. (us); between Puerto Chuelo and Buena Vista, Mar. 14, 1964, Badcock 10 (Kk), SPECIMENS EXAMINED (CULTIVATED): United States, Connecticut: Danielson, Lo tt Gibson 1246 (aan). N : Brooklyn Botanic Garden, Mar. 1932, Evere 5.m. (BH-2 sheets), Mar. 1970. igi » GH, NA), Jan. 1961, Huttleston 1838 (origin, Edinburgh Bot. Gard.) (aa, BH), Jan. 1961, Huttleston 1839 (origin, Florida) (BH), Jan. 1961 (origin Sao Paulo, Brazil) (sx), Jan. i i , Cuba) (BH), Jan. 1964, Peele 1254 (BH), Jan. 1964, Huttleston 2021 (origin, Atkins Garden, Cuba) (Na). FLORIDA: : . » Howard 14929 (aan), Feb. 1969, Gillis 7582 (NA); Homestead i : . , Cooley & Brumbach 11815 (AAH) ursery, Jan. 1932,T.R 32 (BH), Apr. 1932 (z Oneco, Reasoner Bros. N Palm Nurseries, Jan. 19 1971] NEVLING & ELIAS, CALLIANDRA 83 (GH, NA); Winter Haven, Dec. 1964, Mazzeo 846 (wA), Dec. 1967, Mazzeo 2419 (NA); Winter Haven, Cypress Gardens, Jan. 1968, Mazzeo 2468 & 2473 (NA); Winter Park, Lakemont Gardens, Feb. 1956, Wyndham Hayward (kK). TEXAs: Houston, Dec. 1945, E. Teas s.n. (cas). CALIFORNIA: Azusa, Monrovia Nur- sery, Nov. 1960, McClintock s.n. (cas); University of California campus, Mar. 1964, McClintock s.n. (BH, cas); La Jolla, garden, Mar. 1956, McNeill s.n. (cas); West Los Angeles, Vavra Estate, Jan. 1959, Raven, Turner & Whitesel C103 (cas); Santa Barbara, Harvey Estate, Jan. 1957, C. Smith s.n. (cas); Santa Barbara, Hillside Park, Dec. 1956, McClintock s.n. (BH, CAS); Santa Monica, Evans and Reeves Nurseries, Nov. 1942, McClintock s.n. (BH). Bermuda. Paget, Botanical Gardens, Nov. 1963, Manuel 120 (AAH). Dominican Republic. Santiago City, May 1956, Jimenéz s.n. (us). Puerto Rico. Agricultural Experiment Station, Feb. 1921, Hume 2146 (na); Hato Rey, Pennock Garden, Feb. 1967, Howard 16321 (us), 16339 (AAH); Km. 22 on Route 191 through Luquillo Mts., Jan. 1965, Wagner 741 (aan). Curacao. Museum Garden, Dec. 1959, Fr. M. Arnoldo 2257 (us). Ecuador. Hda. Angelica, Chobo, Aug. 1921, Popenoe 1351 (us). British Guiana. Georgetown, Botanic Gardens, Feb.-Mar. 1922, Bailey 566 (BH). Brazil. Rio de Janeiro, Jardim Botanico, Dec. 1928, Ducke 16078 (us); Sao Paulo, Parque Instituto Agronomico, Apr. 1938, Santoro 1574 (us). India. Lucknow, National Botanic Gardens, Dec. 1966, Sharma 68601 (AAH); Dehra Dun, 1928, Singh 80 (aan); Bally gunge, Calcutta, Nov, 187 7, Gamble 5792A (kK). Singapore. Botanic Garden, Oct. 1924 (BH), 1929, Clemens & Clem- ens 22525 (AAH). Java. Buitenzorg, Botanic Garden, Hasskarl s.n. (Aa, frag- ment of type material); 1903 (us), Oct. 1903, Sargent s.n. (AAH). Philippines. Luzon: Los Bafios, Feb. 1954, Steiner 493 (us); College of Agri- culture Campus, Nov. 1946, Sulit PNH 7022 (sau), Sept. 1949, Sulit PNH 9350 (aaH), Dec. 1949, Sulit PNH 12189 (saw), Mar. 1958, MacDaniels 21 (BH). United States. Hawat: Honolulu, Old Pali Rd., Aug. 1961, McClintock s.n. (CAs). England. Royal Botanic Garden, Kew, Mar. 1959, E.N. 211/45 (origin, Peru) (kK), Feb. 1938, E.N. 581/32 (origin, Brooklyn Botanic Garden) (xk), Jan. 1966, E.N. 211/45 (xk), Jan. 1966, E.N. 342/56 (origin, U.S. Dept. Agriculture) (Kk). France. Les Cédres, St. Jean, Cap-Ferrat, Jan. 1963, Marnier-Lapostolle (K) (white-flowered form). South Africa. Pretoria, Botanic Gardens, Aug. 1967, Mills 84 (aaH). Perhaps the impression has been given that because all the cultivated material known as Calliandra haematocephala and C. inaequilatera must be considered as a single species (and the name Calliandra haematocephala applies) that no taxonomic problems remain. This is not the case as there seems to be greater difficulty in determining the range of variation of the species in the wild than that encountered in collating the cultivated with the wild material. In this connection several populations of Calliandra must be considered. The first and the most perplexing, is a population that occurs at Guanai and Mapiri. The Mapiri population is represented by Rusby 1315 (Ny) and the Guanai population represented by Rusby 1314 (Ny) and Bang 1603 (a, cH). As the basis for his Calliandra boliviana Britton used the Rusby collections which were made in May, 1886 and Which must have been among the specimens not lost in the river. There is no question that the material above belongs to this complex. The 84 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 question is whether or not C. boliviana sho uld be a synonym of C. haema- tocephala. The flowers are almost larger than the leaflets found in the collections of C. haematocephala. The t these represent two ecological types, i.e., whether or not they are ecologically separate but geographically sym- patric species. A second population of Calli andra boliviana occurs in the basin of the Rio Bopi in the de partment of La Paz. This population is represented (A), Rusby 654 (BKL, GH), and White 443 (BKL, GH, orifice of the staminal tube. A very interesting population the town of Buena Vista in th is typified by somewhat sm, e plants occur both in the forest and in the Pampas region. It is possible that these plants represent a separate but very closely related species, but we have chosen to treat them as C. haema- tocephala. We are indebted to the directors and curators of those institutions which er A sity, Jamaica Plain BH Bailey Hortorium, Cornell University, Ithaca i Brooklyn : Francisco GH Gray Herbarium, Harvard University, Cambridge ens, Kew an Institution, Washington A number of individuals have assisted us in q variety of ways, and we are grateful for their help. Among these are Mr. Russell Edwards, of 1971] NEVLING & ELIAS, CALLIANDRA 85 the Blandy Experimental Farm; Dr. Patricia Holmgren, of the New York Botanical Garden; Dr. Donald G. Huttleston, of Longwood Gardens; Dr. E. Milne-Redhead, of Kew; Dr. Frederick McGourty, Jr., of the Brooklyn Botanic Garden; Dr. Edwin A. Menninger, of Stuart, Florida; Dr. S. N. Mitra, of the Botanical Survey of India; Dr. Howard Pfeifer, of the University of Connecticut; and Dr. C. G. G. J. van Steenis, of the Rijksherbarium, Leiden. We are much indebted to our colleagues and students of the Arnold Arboretum and Gray Herbarium for extended discussion and suggestions which led to the final manuscript. BIBLIOGRAPHY Anonymous, Calliandra haematocephala Bot. Mag. 86: t. 5181. 1860. Boertacr, J. G. Justus Karl Hasskarl en zijn botanische werken. Teysmannia 5: 129-148. 1894. BREWBAKER, J. L. The distribution and phylogenetic significance of binucleate and trinucleate pollen grains in the angiosperms. Am. Jour. Bot. 54: 1069- 1083. 1967. Cowan, R. A. Correct name of the powder-puff tree. Baileya 11: 94-98. 1963. Dwyansacar, V. R. Embryological studies in the Leguminosae. VIII. Acacia auriculaeformis A. Cunn., Adenanthera pavonina Linn., Calliandra haema- tocephala Hassk., and Calliandra grandiflora Benth, Lloydia 21: 1-25. 1958, Duran-Reynots, M. L. The Feverbark Tree. Doubleday. 1946. Haywarp, W. Calliandra, what kept you? Tropical Homes and Gardens 6(1): 27, 39. 1955. Lowis, L. Familiar Indian Flowers. ¢. 7. 1878. [C. haematocephala as “Caliste- mon Linearis.”’ | MacCreacu, G. White Waters and Black. Century Co., New York and Lon- don, p. 404. 1926. Marxuam, A. H. The life of Sir Clements R. Markham. John Murray, London, p. 384. 1917. Marxkuam, C. R. Peruvian Bark. London, p. 550. 1880. Mason, H. Try your hand at Calliandras. Tropical Homemaker and Gardener. 7(1): 16. 1956. Ruspy, H. H. An enumeration of the plants collected by Dr. H. H. Rusby in South America. 1885-1886.—I. Bull. Torrey Bot. Club 15: 178-184. 1888. ———. On the collections of Mr. Miguel Bang in Bolivia. Mem. Torrey Bot. Club 3: 1-67. 1893. ———. Jungle Memories. McGraw Hill, New York. p. 388. 1933. Taytor, N. Cinchona in Java. Greenberg, New York. p. 87. 1945. WeppeLL, H. A. Voyage dans Le Nord de La Bolivie. Paris, p. 571. 1853. HITE, O. E. Botanical Exploration in Bolivia. Brooklyn Bot. Gard. Rec. 11: 93-105. 1922. ARNOLD ARBORETUM AND GRAY HERBARIUM ITY CAMBRIDGE, MASSACHUSETTS 02138 AND ARNOLD ARBORETUM Harvarp UNIVERSITY CaMBRIDGE, MassACHUSETTS 02138 86 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 THE ECOLOGY OF AN ELFIN FOREST IN PUERTO RICO, 14. THE ALGAE OF PICO DEL OESTE 1 JOHN W. Foerster Puerto Rico is the most easterly and the smallest of the Greater An- tilles. The Sierra de Luquillo highlands are located in the northeastern interior. The study site, described in ecological detail by Howard (1968), called Pico del Oeste, is located in these highlands. This peak, 1050 meters above mean sea level, has a vegetational veneer referred to as an elfin for- est (Howard, 1968). Gleason and Cook (1927) have also termed this area of windswept ridges above the 700 meter elevation the mossy or elfin forest. Pico del Oeste extending above the 700 meter line into the clouds is exposed to the constant effects of th moisture-laden clouds u forest constantly humid (Liboy & Sabio, 1966; Baynton, 1968, 1969). This high moisture content. cou educed 40 percent. In the two papers by itions and influences within the d well with this study Algae in the tropics have had limited study from an ecological stand- * This investigation was supported by research grant GB-3975 from the program of Environmental Biology of the National Science Foundation allocated to Dr. R. A Howard of arvard University, Deep appreciation is extended t h m i Universit 1971] FOERSTER, ELFIN FOREST, 14 87 point. A study of the algal flora of elevated elfin-like forest regions from around the world, as well as in the Caribbean area, has not been made. Most tropical investigations are carried out at lower elevations and are usually concerned with the aquatic habitat. Other Caribbean islands have been studied taxonomically by West and West (1894, 1899) and Drouet (1942). Taylor (1935) studied the alpine regions of Colombia, but his investigations concerned various lakes which were far above the elevation of the elfin habitat and away from the influence of the north- east trades. Fritsch (1907) noted the extensive subaérial algal flora in the tropics. He reported that Cyanophyta predominate, and Chlorophyta are reduced in numbers, with the green alga Trentepohlia being the most suc- cessful of the group. No reference was made to subaérial diatoms or flagellates. The investigation reported here contributes to our knowledge of algae from a tropical area. After completing identification of a series of samples it was noted that the algal flora of the region was not uniformly distributed, but divided into intergrading phyco-synusia. Data are presented to show the epiphytic algal taxa of several habitats including four endemic vascular species, vascular and non-vascular epiphytes, and various ground loca- tions. Further, data are offered to demonstrate gross ecological influences on spatial distribution. METHODS Terrestrial algal epiphytes were collected during the months of Feb- ruary, March, April, May, and August 1966, and during February 1968 from the elfin forest on Pico del Oeste, Sierra de Luquillo, Puerto Rico. Each sample was preserved for later analyses and unpreserved co-samples were used to prepare cultures. Specimens were identified and cross ref- erenced, utilizing the present, most up-to-date and available monographic works, The nomenclature adopted is as follows: for the Cyanophyta, Desikachary (1959); for Bacillariophyta, Patrick and Reimer (1966); for the Chlorophyta, Prescott (1962), Randhawa (1959), Saxena (1962), and Islam (1963); for the Chrysophyta, Pyrrophyta, and Euglenophyta, Pres- cott (1962). Some of the other works employed included Tilden (1910), Drouet and Daily (1956), Van Huerck (1896), Tiffany and Britton (1952), and many other short papers on individual species and genera. Vascular plants were identified by Dr. R. A. Howard of the Arnold Arboretum, Harvard University. Analyses were made of the probability that numbers of species collected from each location may exhibit preference of substrate, and/or of micro- climate. Analysis of variance (ANOVA) was applied to the difference i numbers of algal species found on the various plants and to the differ- ence in number found on the vascular stem at various heights above the ground. Linear regression was further employed to test the height-species number variation and relationship. These gross ecological parameters were chosen as representative of multiple factors, such as light, moisture, 88 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 nutrients, etc., which this portion of the study was not equipped to measure directly. RESULTS AND DISCUSSION , height above the ground, and meteorological conditions only. Lyford (1969) has reported that in an area-to-area comparison the trunks of the trees then light, moisture, and tem factors, In all samplings the effect of shield appears to be a signi perature become more important limiting cloud cover would significantly influence sity. When the sun is most direct the at- st, and when its zenith is lowest for P ni hese are related to the vascular host and to their Position on the host. Note the high proportion of Cyanophyta which dominate the microhabitats studied. Diatoms were also present. yund (1945) and Fritsch (1922) reported diatoms in the terrestrial ted that the amount of moisture appears to be moisture in the elfin forest seems to be adequate - er, the presence of flagellates in the samples an appropriately moist 2 habitat, They viding some of the alga 1 colonizers. 1971] FOERSTER, ELFIN FOREST, 14 89 producing precipitation in the form of rain, clouds, dew, mist, and fog, establishing an environment suitable for algal epiphytes. The precipita- tion would deposit any of the airborne algae which were trapped as nu- clei of condensation during evaporation at lower levels. Further, the mi- gration of temperate zone birds and the wanderings of native bird and other animal species must not be ruled out as a possible source of algal introduction. ABLE | depicts the principle phyco-synusia encountered in the Feb- ruary-March, 1966, collections relating them to the vascular host and TABLE 1. Subaérial Epiphytic Phyco-synusia of Four Vascular Species Sampled during February-March, 1966 DISTANCE ABOVE VASCULAR Mayor GrounD (METERS) Host APPEARANCE PHYCO-SYNUSIA 0-1 1-2 Cyathea white globular a diatom-filamentous xX pubescens m bluegreen association in an exten- sive Asco- mycete fungal mat Miconia always moist gelat- Stigonema-Lyngbya- x pachyphylla inous flat mass, Hapalosiphon (chiefly greenish-gray Stigonema informe) Ocotea green material on almost pure association spathulata dead leaves of Hapalosiphon luteolus white, globular Mougeotia caimani chief X gelatinous mass species green, gelatinous Hapalosiphon-Trentepohlia x growth T abebuia flat-green Stigonema-Hapalosiphon- X rigida gelatinous Oedogonium (Stigonema is principle) solid mass Stigonema-Hapalosiphon- X Scytonema (Stigonema panniforme principle) flat mass Chroococcus-C ylindro- x cystis-Synechococcus flat mass Synechococcus-Cylindro- X cystis-Frustulia gelatinous, white Chroococcus-Hapalosi- x colony phon-Trentepohlia 90 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 52 their relative height above the ground. It is interesting to note that “ inant species change with height on the tree. Particularly, this is ee on Tabebuia rigida as the dominance shifts from a Stigonema —— to a Trentepohlia association (a shift from a blue-green to a — s synusium). The reason for the shift in association dominance, and e diminishing number of species present undoubtedly rests in the micro- habitat itself. It is not the purpose of this paper to analyze these a limiting factors, but to point out the composition of the subaérial a ga epiphytic flora on the principle vascular plant hosts, and to note the ee cacies of the spatial distribution of component species. What causes the variation is still uncertain but physical data indicate that moisture as wetness or as a light shield may be a principal factor. Grubb and bee more (1966) report that fog is an important limiting factor in tropica montane forests. For the algae this suggests not only a high atmospheric moisture content, but also light attenuation. TABLE 2 is a representation of the mean climatic conditions of the sample site for February-March, 1966. The mean temperature is moderate, solar TABLE 2. Mean Climatic Data for the February-March Period of Sampling * MEASUREMENT Mean READING Temperature 1B": Minimum Relative Humidity 80.6% Langleys (gram calories/cm.) Daily Precipitation above canopy 283.45 (above canopy) 12.7 mm. below canopy 9.7 mm. Percent of Time Humidity less than 100% above canopy 28.4% Humidity at ground level below canopy 100% * Means computed from data supplied by Dr. R. A. Howard. energy is available, precipitation is adequate, and the relative humidity seldom drops below 100 per 1971] FOERSTER, ELFIN FOREST, 14 91 cantly and the hypothesis was rejected at the 99 per cent level of confi- dence. A definite variation appears to exist between the samples. However, it must be pointed out that, while samples were randomly collected, the number of samples on which the statistics are based is of necessity small. The ideas of microhabitat are introduced to demonstrate that the algal flora of the cloud forest is not a superficial veneer, but, as in an aquatic habitat, is distributed in relation to the stress of the micro- climate in which it exists. Varying the microclimate may in some instances accentuate or retard various species and thus shift the community struc- ture both quantitatively and qualitatively. A regression analysis was performed on the height above the ground where samples were collected (X) and the relation of this height to the number of species found (Y). Samples 9 and 10 (see TABLE 6) are not used in this computation because they are isolated growths not found on the trunks of the trees. Though the data are limited, a definite relation- ship appears to exist and is presented here to demonstrate the possibility of microhabitat existence. The regression equation (Y = 45.5 — 18.2X) was computed to test the height factor hypothesis, and the line derived (FicurE 1) showed that as samples were taken higher on the tree trunks the number of species diminished. This is an indirect relationship further qualified by the correlation coefficient (r = — .757), a high negative correlation, Other samplings reported here were made during April-May, 1966, August, 1966, and February, 1968. Tas e 3 is a listing of the algal species from five habitats not previously described. Specimens from a flowing Stream (1), from bryophytic ground-covering adjacent to the access road (2), from cement drains at the experimental site on the access road (3), from bryophytic ground-covering receiving some ground water run-off (4), and from a horizontal trunk of Tabebuia rigida (5). Again the re- sults of the qualitative survey indicate the existence of microhabitats. The major phyco-synusia are listed in TABLE 4. It is interesting to note that the diatom Frustulia rhomboides var. capi- tata is the most prevalent species in the habitats listed for April-May, 1966. Further, it is thought that this species exhibited polymorphism in the samples investigated. In all samples studied a small number of F. rhomboides var. viridula were found, a larger number of F. rhomboides var. saxonica, and an extremely large number of F. rhomboides var. capi- tata. In light of Stoermer’s recent report (1967) on polymorphism in Mastogloia, the probability exists that these three varieties are growth forms, each variety being reduced in size by approximately half in the following order: var. viridula, var. saxonica, var. capitata. Culturing Frus- tulia was not successful, and therefore only field observations can be of- fered as evidence. The April-May, 1966, collections were composed mainly of diatoms (1, 2,3,4). The habitats from which the samplings were made were on the ground and bathed in moisture. Specimens from Tabebuia rigida were dominated by the blue-green alga Hapalosiphon, with many lesser impor- 92 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 FOUND ° | ou lL NUMBER OF SPECIES oO | Y= 45 5-18.20 ¥r"*-.757 rs. Re) 12 1.5 1.8 2.1 EPIPHYTE HEIGHT ABOVE GROUND (METERS) Ficure 1, Relationship of the height of collection on the vascular host to the number of species from the February-March, 1966 sampling. orizontal trunk of T. rigida was approximately ground under the canopy where light and : rlous synusia, number of species height of collection. A shift from i an association dominated by Cyano- phyta to one with increased domi 1971] FOERSTER, ELFIN FOREST, 14 93 TABLE 3. Species Found in Five Samples Collected during April-May, 1966 SPECIES Hasirat * CYANOPHYTA Nostocales Anabaena sp. (3) (4) Aulosira fritschii Bharadw. (3) (4) A. prolifica Bharadw (4) Oscillatoria foreaui Frémy (2) Stigonematales Heuee son welwitschii W. & G. West (5) CHLOROPHYTA Ulotrichal Stichococcus subtilis one) Klercker (1) (2) Ulothrix variabilis Kiit (1) (2) Zygnematales Euastrum sp, (1) (2) Cosmarium sp. (1) (2) CHRYSOPH Bhisochera dales Chrysidiastrum catenatum Lauterb. (1) (2) Eo Centra Biddulphia favus ee Van Heurck (1) (2) (3) (4) G) Cyclotella glomerata Bachm (5) Melosira varians C. Ag. (1) (2) Pennales Achanthes hauckiana Gru (5) A. microcephala (Kiitz.) “Grun, (1) (2) Diatoma vulgare Bory (1) (2) (5) Eunotia naegelii Migula (4) E. praerupta Ehrenb. (1) E. tenella (Grun.) Hust. (1) (4?) (5) E.s (4 earl rhomboides var. capitata (A. ede a Patr. (1) (2) (3) (4) (5) F. rhomboides var. saxonica (Rabenh.) DeTon (1) (2) (3) - (5) FP. rhomboides var. viridula (Bréb.) Cl. (1) (2) (3) (4) pga tenelloides Hust. ‘o (2) (4) () iach chong (?) Hohn & Hellerman (1) (2) N. hantzschiana Rabenh. (3) Shales Sudetica fone (1) (2) . viridis var. minor Cleve (4) Rhopalodia gibberula (Ehrenb.) Mueller (1) (2) (S) R. ventricosa (Kiitz. ) Mueller (2) *Sample numbers correspond to the following: (1) Specimens found growing one inch deep in the water of a flowing stream. (2) From bryophytic growth covering oO adjacent to access road. (3) From cement drains a access r (4) From Baldr ground-covering kept wet by rain and clouds, as well as some water (5) From a Sew: on a horizontal trunk of Tabebuia rigida (two feet off ground). 94 JOURNAL OF THE ARNOLD ARBORETUM [voL, 52 TABLE 4, Phyco-synusia of Samples from April-May, 1966 Torat No. Haszirat * SYNUSIA SPECIES 1. In small flowing stream Ulothrix-Frustulia 18 2.) On bryophytes near road Oscillatoria-Frustulia 18 3. From cement drains Aulosira-Frustulia 6 4. On bryophytes receiving Aulosira-Frustulia 13 water run-off . On horizontal trunk Hapalosiphon-Frustulia- of Tabebuia rigida ohlia Trentep 11 * See also footnote to Table 3. TABLE 5. Phyco-synusia of Samples from August, 1966 DISTANCE ABOVE NUMBER Hasirat GROUND (METERS) OF (vascular host) Synusia 0-1 1-2 SPReang Tabebuia rigida Stigonema-Chroococcus x 14 T. rigida Aulosira-Zygnemo S78 8 Ocotea spathulata Hapalosiphon-Phormidium x 13 O. spathulata Stigonema-Frustulia xX 8 Miconia pachy phylla Stigonema-Chroococcus x 14 Cyathea pubescens Navicula-Frustuli ° T. rigida Phormidium-Frustulia xX 8 T. rigida Stigonema-Frustulia x 15 O. spathulata Aulosira-Frustulia x if cated. The Samplings for F ebruary- period of less rainfall and ] of August, This is tuation on the peak . ain performed to test the same relationship the February-March samples. The line derived (FIGURE y the Tegression equation (Y = 6.30 + 4.55X) showed that a positive n wW : imiting factor. increased, so has ing, a factor not as prevalent in the February- March samplings, 1971] FOERSTER, ELFIN FOREST, 14 95 TABLE 6. Subaérial Epiphytic Algal Species and Their Vascular Hosts from Pico del Oeste, February-March, 1966 Tabebuia rigida Ocotea spathulata Miconia wn pachyphylia || Cyathea pubescens fo) ~r oo w as © — ° SPECIES 1 2 CYANOPHYTA occus is persus (Keissler) Lemm. x pea, bi ag S$ var. minor . M. Sm evitent W. West x im x ana pallidus ema a Nag. r. Dactytcoecops smithii R. & F. Chodat x cae a Am MAM 4 ~ aeruginosa (Carm.) Kiitz. iS Kiitz. x Gloeot Pati var. composita G. Merismopedia ree hae Bréb. ex Kitz. a pe a ~ as i of o =] i Nt od ZA po: ga val re 5 ghia esc x Rhabdode sas Synechoco pias Nag. Xx x X m _ eon pea rostaffinskii Hansg. JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 TaBLe 6. Subaérial Epiphytic Algal Species and Their Vascular Hosts from Pico del Oeste, February-March, 1966 (continued) Tabebuia rigida spathulata Miconia a pachyphylla Cyathea pubescens Ocotea SPECIES 1 2 6 7 8 3 > Oo — ea istorcat andi nabaena anomala Fritsch x laxa (Rabenh.) A. Br. x epee Bory ex x Bor Flah. ors Fritsch xX » Caan. ¢ Ca epiphytica W. & G. West Xx weberi Schmidle echinulata (J. E, a0 Be longiarticulata Gs ala ngoyva limnetica Lemm. x ha agardhii Gom. xX annae van Goor ees Lemm, xX x Peck midiu tenue (ene, Gom. xX x Pseudana stent xX rien omnia Lauterb, x amplum W. & G. West x x hofmannii C. Ag. ex Born. & Flah. Stuposum (Kiitz.) Born. ex orn. & Flah. x toly pothricoides Kiitz, ex orn. & Flah. x ss seo Sis hyapii (Bharadw.) Geitler Salah Subtilissima Kiitz. 1971]. FOERSTER, ELFIN FOREST, 14 97 TABLE 6. Subaérial Epiphytic Algal Species and Their Vascular Hosts from Pico del Oeste, February-March, 1966 (continued) Tabebuia rigida Ocotea spathulata Miconia wn pachyphylia || Cyathea pubescens ON ~ oo w a oOo — °o SPECIES 1 2 Tolypothrix distorta Kiitz. ex Born. & Flah. Pleurocapsales Xenococcus erneri Hansg. x Stigonematales hibernicus W. & G. West x 2 = 5 y & a: im Q $2 4 La stuklmannii Hieron. welwitschii W. & G. West Nostochopsis x lobatus Wood em. Geitler ahana onema aderugineum Tilden hormoides (Kiitz.) Born. & Flah. informe Kiitz. mesentericum Geitler b ae panniforme Harv ex Born. & Flah. Stigonema turfaceum (Berk.) Cooke iz Flah. A A Mo eee A A CHLOROPHYTA au (L.) Martius x x ita (L.) Wallroth x mbrina (Kiitz.) Born. a, Ankistodes falcatus (Corda) Ralfs x Characium x stipitatum (Bachm.) Wille 98 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 TABLE 6, Subaérial Epiphytic Algal Species and Their Vascular Hosts from Pico del Oeste, February-March, 1966 (continued ) Tabebuia ° rigida Cyathea pubescens Miconia ow pachyphylla Ocotea * spathulata ry o SPECIES 1 2 cain, CI Se a de Chlorella vulgaris Beyerinck p ee. Chlorococcum humicola (Nag.) Rabenh. x > a. x cystis © borgei Snow pyriformis Prescott x x submarina Lagerheim Pediastrum muticum var. crenulatum Prescott x x ala Fiillemann) G. M. Sm lacustris = G.M. Sin. Scenedesm wemieas var. platydisca G. M. Sm. anal Tetraedron tumidulum (Reinsch) Hansg, x Cylindrocapsales Cylindrocapsa conferta W. West x x geminella Wolle x Oedogoniales Oedogonium sp. x rae Ra Sphaero cular (Roth) C, Ag. Tetrasporales loeocystis Sphaerocystis Schroeteri (?) Chodat x x xX 7 ybonematales proses (Kiitz.) Hazen oa Ulotrichales Stichococe: subtilis (Kiitz.) Klercker x Ulothrix aequalis Kiitz. 1971] FOERSTER, ELFIN FOREST, 14 99 TABLE 6, Subaérial Epiphytic Algal Species and Their Vascular Hosts from Pico del Oeste, February-March, 1966 (continued) SPECIES Tabebuia © rigida Ocotea shal pathulata Oo Miconia wn pachyphylila | Cyathea pubescens = °o Zygnematales Cosmarium dentatum Wolle C se he ystis issonii var. minor West Hyalotheca iin Nordst. M ougeo caimani aaa floridana Transeau Netrium digitus (Ehrenb.) = ip os. & Rothe Zygnemopsis desmidioides (W. & G. West) Transeau CHRYSOPHY Heterococcales Peroniella panconi G. M. Sm. Pleurogaster lu Pascher a ae gigas Pascher Rhyzochrysidales Lagynion reductum Prescott BACILLARIOPHYTA s Centrale crenulata (Ehrenb.) Kiitz. tium Tricera semicirculare Brightw. undulatum Ehrenb botlaae 0) (Kiitz.) Hust. ~a Dia Hemale (Roth) Heib, Dia aeowshe Grev. oneis oculata (Breb.) Cl. a 100 JOURNAL OF THE ARNOLD ARBORETUM [VoL, 52 TABLE 6. Subaérial Epiphytic Algal Species and Their Vascular Hosts from Pico del Oeste, February-March, 1966 ( continued ) Tabebuia rigida Ocotea * spathulata Cyathea pubescens Miconia uw pachyphylla SPECIES a cre © iy oO unotia tenella (Grun.) Hust. ragilaria rhomboides var. capitata (A. Mayer) Patr. Nitzschia dissipata (Kiitz.) Grun. ignorata Krasske x x Stauroneis ignorata (?) Hust. EUGLENOPHYTA Euglenales Euglena ial elastica (?) Prescott x polymorpha Dangeard Xx Glenodinium kulczynskii (?) (Wolosz.) Schiller ” Trachelomonas charkowiensis (?) Swirenko ToTat NuMBER oF SPECIES = 98 38 38 46 10 19. .16 47 36 Key to Numpers 1 — gelatinous white colony on Tabebuia 6 — solid mass on T. rigida rigi a 7— gelatinous mass on T. rigida 2 — gelatinous gree n colony on 7. rigida 3 — green gelatinous mass on Ocotea thulata $ d) 4— white gelatinous mass on O. spathy- 10 — har ass growing on dead stump lata of Cyathea pubescens 5 — always wet gelatinous mass on Mi- conia pachyphylla TABLE 7 qualitatively reflects the major synusia as distributed more randomly, It would appear that moisture, light during the wetter i ar epiphyte and from the om squeezing the epiphytic nks was analyzed. In TasLEe 8 the species 1971] FOERSTER, ELFIN FOREST, 14 101 7 SURZ 25:<) a = = ro) uw 2023) w — oO uJ a Léa) eat LX fo) 10) a oO uw 0 o = = 10 ad e89) A Y* 6:3 4+ 4:5X r=.548 5 Oo (@) | T T T T 52 i 39 II, 5 1.8 23 EPIPHYTE HEIGHT ABOVE GROUND (METERS) Ficure 2. Relationship of the height of collection on the vascular host to the number of species from the August, 1966 sampling. are listed from each sampling. The dominant organisms forming the phyco-synusium in V. sintenisii were Diatoma-Stigeonema-Pithophora, and the diatom Diatoma vulgare was the most prevalent species. The leaves of these epiphytic bromeliads overlap forming a natural catch basin for moisture. The species found in this basin are more predomi- 102 JOURNAL OF THE ARNOLD ARBORETUM [voL, 52 TABLE 7. Species found on Endemic Vascular Species during August, 1966 SPECIES Vascutar Host * + CYANOPHYTA Chroococcales Aphanocapsa biformis A. Br. C1) €2) Aphanothece clathrata W. & G. West 3 Chroococcus dispersus (Keissler) Lemm, (1) (8) C. minor (Kiitz.) Nag. 5 C. minutus (Kiitz.) Nag. (1) (3) (5) C. pallidus Nig. C. turgidis (Kiitz.) Nig. (1) (3) (5) (8) (9) Gloeothece palea (Kiitz.) Nag 5 Microcystis aeruginosa Kiitz. (1) (3) (5) (8) Synecococcus aeruginosus Nig. (1) (3) (5) (8) Nostocales nabaena circinalis Rab (1) (8) Aulosira prolifica (?) taad ch) 02) (5) (9) Cylindrospermum catenatum 0) Ralfs Phormidium tenue (Menegh.) Gom. (3) (5) (7) (8) Scytonema ected ot (Kiitz.) Born. (2) (5) S. schmidtii (5) Stigonema ales Bite luteolus W. & G. West (3) ema growth form of Stigo- es) nema hormo Stigonema dendroideum Frémy Ss. ( hormoides gee Born. & Flah. (1) (2) (3) (4) S. panniforme (C. Ag.) Ha (5) CHILO OPHYTA ptowpaieer Zygnemopsis desmidioides (W. & G. West) (2) Transeau BACILLARIOPHYTA ales Biddulphia alternans (Bail.) (2) (4) Van Huerck Campylodiscus echeneis Ehrenb. fay) C5) Cyclotella glomerata Bac (7) (9) Melosira crenulata sent Kiitz. i M. ere oho ier Ralfs (6) (9) M. (3) (possible auxospore noted) Paleiae” Cymbella ventricosa Kiit (7) (9) Eunotia exigua a (Bréb) Rabenh, (3) (4) (8) E. naegelii Mig (7) (9) : 4 praerupta Ehrenb. Sa (9) . Lee (Grun.) Hust. C7) (9) ) rustulia oe var. capitat: 8) 9 ee ar. capitata (1) (2) (4) (5) (6) (7) ¢ F, ppd os va tates var. Peo (2) (4) (9) F. rhomboides var. viridula (Bréb. Isthmia enervis Ehrenb, seri atiee iG) (9) 1971] FOERSTER, ELFIN FOREST, 14 103 TABLE 7. Species found on Endemic Vascular Species during August, 1966 (continued) SPECIES VASCULAR Host * + Na vicula carniolensis (?) p pang sp. (radiate, capitate form) N. sp. (different from ab Nitzschia dissipata (Kiitz.) Grun. N. hantzschiana Rabenh. N.s Sp. Pinnularia sudetica Hilse Rhopalodia here Mueller Surirella ovata Kiitz (1) (5) (7) (1) (8) (9) (1) * The numbers identify the vascular host species as follow: (1), (2), (2), Sd Tabebuia rigida; (3), (4), (9), Dot spathulata; (5), Mi- conia pachyphylla ; (6), Cyathea pubescens { For relative position of sample on vascular host see Table 5. TABLE 8. Algal Species Components of Non-algal Epiphytes Collected in August, 1966 SPECIES BROMELIAD BRYOPHYTE SQUEEZINGS CYANOPHYTA Chroococcales Aphanocapsa grevillei (Hass.) Rabenh. Chroococcus dispersus an. minutus (Kiitz.) Nag. Ss poe ge Nag. G Pa Ry Fascicularis Lemm lo ic hii aeruginosus Nag. Nosto Pe Pris Bharadw. Nostoc m C. Oscillatoria suet ada Elenkin O. subbrevis Schmidle Pleurocapsales xosarcina spectabilis Geitler Stigonematales Hapalosiphon delicatulus W. tg G. West W. & G. Wes HA. welwitsc. Stigonema aerugineum Tider S. hormoides (Kiitz.) Born. & Flah. S. minutum (C. Ag.) Hass var. minor G. M. Sm. rr a ae mM OM ahaa am wm RM MM a ana 104 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 TABLE 8. Algal Species Components of Non-algal Epiphytes Collected in August, 1966 (continued) SPECIES BROMELIAD BRYOPHYTE SQUEEZINGS CHLOROPHYTA Chaetophorales Cephaleuros ee Kunze Protococcus viridis C. Ag. Stigeoclonium Sie (?) (Franke) Heerin Trentepohlia aurea var. tenuior Bri ; ana T. aurea (L.) Martius T. torulosa De Wildeman Chlorococcales Characium obtusum (?) Braun Chlorella ellipsoidea Gern Crucigenia jisersfaesl Morren Pediastrum muticu Trochiscia raillica (Reinsch) Hansg. Cylindrocapsales Cylindrocapsa conferta W. West Oedogonial Pithophora oedogonia (Mont.) Wittr. Xx bgp cannes pore Tribonema utriculosum (Kiitz.) Hazen x Ulotrichale Geminella interrupta Turpin xX os ales a mA PA re a BACILLARIOPHYTA Centrales Coccinodiscus lacustris (?) Grun. x Cyclotella bodanica (?) Eulenst. x Melosira crenulata (Ehrenb.) Kiitz B M. granulata (Ehrenb.) xX ea semicirculare Brightw. xX x Penn ere x Diatoma Hemate (Rat Heib. x x D. hiem meso (Ehre run. xX Diatoma ae var. elongatum Lyngb. x D. vulgare Bo xX Page Pectinalis f, elongata Van Huerck xX - pectinalis var. minor (Kiitz.) Rabenh. ) Rabenh, Ha (?) (Grun.) th ragilaria brevistriata Gru meee 1971] FOERSTER, ELFIN FOREST, 14 105 TABLE 8. Algal Species Components of Non-algal Epiphytes Collected in August, 1966 (continued) SPECIES BROMELIAD BRYOPHYTE SQUEEZINGS Frustulia rhomboides var. capitata (Mayer) Patr. Xx Hantzschia amphioxys (Ehrenb.) Grun. xX xX Navicula sp xX x Nitzschia tryblionella Hantzsch. x Surirella ovata Kiitz. Xx Tropidoneis sp. x PYRROPHYTA Dinococcales Cystodinum cornifax (Schilling) Klebs x Torat NuMBER OF SPECIES 42 32 diatoms plus the fact that the trunks of the vascular hosts transmit to the ground 21.4 per cent of the rainfall reaching the forest canopy may be one of the principal factors in determining the presence of this diatom- dominant association. During February of 1968, mucilaginous material from around aérial roots of Hedyosmum arborescens was analyzed for algal components. One sample was fluid, containing a large amount of water, while the sec- ond sample was highly consolidated and, in comparison to the other sample, low in water. Taste 9 compares the algal species found in the two samples differing in moisture content. The sample containing the most moisture (1) has the highest number of species. Species such as Trentepohlia aurea, Melosira varians, and Frustulia rhomboides var. capitata are represented in both samples. Throughout the entire study these species have exhibited eurybiont characteristics. The green alga T. aurea was the major component in both samples, while none of the other species appeared more than occasionally. As a gross ecological parameter the height above the ground seems to become a limiting factor to many diatoms and blue-green algae. Height would include the above mentioned factors of light, moisture, and tem- perature as well as many others. A general shift from a Bacillariophyta- dominated to a Cyanophyta-dominated to a Chlorophyta-dominated algal synusium is noted as samples are taken from points higher above the ground. In concluding, it is important to compare the regression and corre- lation analyses of the February-March, 1966, to August, 1966, samples. A shift in environmental utilization is evident. This could be called a Seasonal response and it may be a reaction to light attenuation caused by creased cloudiness during August. It is possible that the algal species are re-distributing to a higher portion of the vascular stem in response to a more favorable light condition. Only samples taken at higher levels 106 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 TABLE 9. Algal Species Found among Mucilaginous aérial root secretions of Hedyosmum arborescens SPECIES SAMPLE NUMBER * CYANOPHYTA Chroococcales Chroococcus dispersus (Keissler) Lemm. 1 C. turgidis (Kiitz.) Nag. 2 Microcystis ramosa Bharadw. 1 Nostocales Anabaena circinalis (?) Rabenh. 1 Aulosira prolifica (?) Bharadw. 1 illatoria foreaui (?) Frémy 2 Stigonematales Stigonema hormoides (Kiitz.) Born. & Flah. 1 CHLOROPHYTA Chaetophorales Protococcus viridis C. Ag. 1 Trentepohlia aurea var. tenuior Briihl & Biswas 1 - Chlorococcales Chlorella-like BACILLARIOPHYTA Centrales Biddulphia Melosira v Pennales Eunotia fallax Cleve-Euler Frustulia rhomboides var. capitata (Mayer) Patr. Navicula cincta (Grun.) Cl. N. contenta Grun. Nitzschia parvula Lewis As Th Toran NuMBER oF SPECIES 14 6 alternans (J. W. Bailey) Van Heurck 1 arians C. Ag. 1 -—_ = SS a * Sample (1) refers to a very watery mucoid material, and Sample (2) refers to 4 very thick and highly viscous mucoid material ty-March, 1966, are compared to August, ct a “wetness” requirement but may be re- 1971] FOERSTER, ELFIN FOREST, 14 107 lated to the amount and duration of light, which is affected by moisture in the form of clouds and fog. Further, Fritsch (1922) reported Tren- tepohlia to be very resistant to fluctuating moisture. Therefore, Trente- pohlia dominance higher on the vascular trunk may be related not only to a tolerance for changing moisture but also to a higher light tolerance (or requirement). In all the investigations of nearly aquatic ground samples and highly moist aérial roots there is a high incidence of diatoms, while samples from drier habitats shift to non-diatom dominant forms. It is reasonable to conclude on the basis of the samples investigated that the large number of algal epiphyte species studied have become successfully established and exist in distinct microhabitats. The algae exhibit seasonal variation in possible response to meteorological conditions. The gross ecological parameter of height above the ground appears to be a proper method of assessing environmental influences in a vertical direction. SUMMARY The cloud forest can be likened to a large natural culture unit having uniformly constant meteorological conditions, and, judging from the rela- tively large number of species found, adequate nutrients. Though all would appear uniform at the first casual investigation, careful analysis indicates that the subaérial algal epiphytes are not uniformly distributed, but are segregated into phyco-synusia. These seem to be microhabitats related to height above the ground and type of host material. The data depict an apparent random horizontal distribution but a less random vertical pattern of distribution. This vertical distribution and apparent seasonal succession pattern appears related to slight changes in meteoro- logical conditions. Samples taken higher on a tree trunk during drier months contain fewer species than those collected closer to the ground. The probability exists that epiphytes located higher above the ground are subjected to greater fluctuations in temperature, moisture, and light. However, the lower parts of a tree trunk are older and perhaps more Stable chemically. F urther, during wetter months the species-count on the host appears to be rearranged and seasonal species variation appears. During this period increased cloud cover and fog reduce light penetra- tion which may be instrumental in this change. Subaérial epiphytic algal growth in all samples studied, and especially on the four principle vascular host species in the Luquillo elfin forest on Pico del Oeste, appears not as a uniformly distributed algal veneer, but rather as associations of species representing various intergrading microhabitats. LITERATURE CITED Atmopovar, L. R. 1963. The fresh-water and terrestrial Cyanophyta of Puerto Rico. Nova Hedwigia 5: 429-435. 108 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 Baynton, H. W. 1968. The ecology of an elfin forest in Puerto Rico, 2. The microclimate of Pico del Oeste. Jour. Arnold Arb. 49: 419-430. . 1969. The ecology of an elfin forest in Puerto Rico, 3. Hilltop and forest influences on the microclimate of Pico del Oeste. Jbid. 50: 80-92. DestkacHary, T. V. 1959. Cyanophyta. Indian Council of Agricultural Re- search, New Delhi. 686 pp. Drovet, F. 1942. The filamentous Myxophyceae of Jamaica. Fieldiana Bot. 20: 107-122. & W. A. Datry. 1956. Revision of the Coccoid Myxophyceae. Butler Univ. Bot. Stud. 12: 1-218. Fritscu, F. E. 1907. The subaerial and freshwater algal flora of the tropics. Ann. Bot. 21: 235-275. . 1922. The terrestrial algae. Jour. Ecol. 10: 220-236. Garpner, N. L. 1927, New Myxophyceae from Porto Rico. Mem. N.Y. Bot. Gard, 7: 1-144. . 1932, The Myxophyceae of Porto Rico and the Virgin Islands. Jn: N.Y. Acad. Sci., Scientific Survey of Porto Rico and the Virgin Islands. 8: 249-311. GLeason, H. A., & . T. Cook. 1927. Plant ecology of Porto Rico. Jn: N.Y. Acad. Sci. Scientific Survey of Porto Rico and the Virgin Islands. 7: 1-173. Gruss, P. J.. & T. C. WHIrmore. 1966. A comparison of montane and lowland rain forests in Ecuador. II. The climate and its effects on the distribution and physiognomy of the forests. Jour. Ecol. 54: 303-333. Hatickt, P. 1964. Observations on algae at El Verde, Oct. 1-9, 1964. Puerto Rico Nuclear Center. The Rain Forest Project Annual Report Fy-64: 106-108. Howarp, R. A. 1968. The ecology of an elfin forest in Puerto Rico, 1. Intro- duction and composition studies. Jour. Arnold Arb. 49: 381-418. Istam, A. K. M. N. 1963. A revision of the genus Stigeoclonium. Nova Hed- wigia 10: 1-164 + 47 pl. Linoy, J. A. G. & H. Sasro. 1966. A preliminary study of the epiphytic algae in ) University of Puerto Rico Agricultural Ex- periment Station, 10 - REIMER. 1966. The Diatoms of the United States. Vol. 1. 688 pp. Acad. Nat. Sci. Philadelphia. Monogr. No. 13 Prescott, G. W. 1 962. Algae of the Western Great Lakes Area. Wm. C. Brown Co., Dubuque. 977 pp. RANpDHAwA, M. S. 1959...7 search, New Delhi, 478 pp. SAXENA, P. N. 1962. Algae of India. I. Chaetophorales. Bull. Natl. Bot. Gard. Lucknow 57: 1-59. SCHLICHTING, H. E. 1960. The role of water fowl in the dispersal of algae. rans. Am. Micr. Soc. 74: 160-166 ygnemaceae. Indian Council of Agricultural Re- 2 ie Viable species of algae and protozoa in the atmosphere. Lloydia 1971] FOERSTER, ELFIN FOREST, 14 109 . 1964. Meteorological conditions affecting the dispersal of airborne algae and protozoa. /bid. 27: 64-78. STOERMER, E, F. 1967. Polymorphism in Mastogloia. Jour. Phycol. St 73-71, Szeicz, G. 1966. Field measurements of energy in the 0.4-0.7 micron range. In: Light as an ecological factor. Eds. BAINBRIDGE, EvANs, and RACKHAM. pp. 41-52. Blackwell Scientific Publ., Oxford. Taytor, W. R. 1935. Alpine algae from Santa Marta Mountains, Colombia. Am. Jour. Bot. 22: 763-781. TrrFany, L. H. 1936. Wille’s collection of Puerto Rican freshwater algae. Brit- tonia 2: 165-176. & M. E. Brirron. 1944. Freshwater et oe and Xanthophyceae from Puerto Rico. Ohio Jour. Sci. 44: 39- . 1952. The Algae of Illinois. Univ. of ae Press, Chicago. 407 pp. Tren, J. 1910. Minnesota Algae. Vol. 1. Univ. of Minn. Press, Minneapolis. 328 pp. + 20 pl. Van Huerck, H. 1896. A eee on the Diatomaceae. Wm. Wesley and Son. London. 558 pp. + 35 pl. West, W., & G. S. West. 1894. On wir siege’ algae from the West Indies, Jour. Linn. Soc. Bot. 30: 264— - 1899. A further contribution to te ‘freshwater algae of the West Indies, Ibid. 34; 279-295. Witte, N. 1915. Report on an expedition to Porto Rico for collecting fresh- water algae. Jour. N.Y. Bot. Gard. 16: 132-146. DEPARTMENT OF BIOLOGICAL SCIENCES GouCHER COLLEGE Towson, Maryitanp 21204 110 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 A REVISION OF THE BORAGINACEAE OF WEST PAKISTAN AND KASHMIR * S. M. A. Kazur 18. Rochelia Reichenb. Flora 7: 243. 1824; Icon. Crit, 2: 13. t. 123; 1824. Raclathris Raf. Sylva Tellur. 167. 1838. TYPE sPEctEs: R. saccharata Reichenb. Usually annual, rarely biennial, villous or hispid herbs. Leaves ace Inflorescence terminal, simple or branched, bracteate. Flowers ap o or white. Calyx usually 5-fid or rarely 9-10-fid (not in the West Pakis . or Kashmir species), usually enlarged in fruit, lobes linear, oblong s lanceolate, rounded auricled or sagittate at the base. Corolla oe : campanulate, infundibuliform or tubular, tube usually cylindric, aed with or without appendages, lobes small, spreading. Stamens 5, inclu a anthers ovate, usually obtuse. Gynobase columnar, slightly develop 7 Style filiform, stigma indistinctly capitate. Nutlets 2, 1-seeded, ren oblong-ovate, glabrous, smooth to tuberculate, affixed along the whole length of the gynobase, A genus of about 15 species, centering in southwestern Asia and rang- . ° * . . in ing from central Asia to the Mediterranean basin. One species occurs 1 Australia. KEY TO THE SPECIES a. Nutlets glabrous, smooth, shi ning. a. Nutlets minutely tuberculate to st b. Calyx lobes ovate-lanceolate in 8 oe ts 1. R. leiocarpa. ellate-papillose. fruit, auriculate-cordate at base. ....-- - ¢. Calyx lobes 6-7 mm. long in fruit, erect or slightly recurved 2 ae ds the base, subunguiculate, midrib nie ; th stiff erect trichomes. ... 3. R. macrocalyx. Qo € 3 e B i cn Pate. b — : tly developed, trichomes various. f m. long in fruit, at minimum twice the length A 2 te or 4. R. peduncularis. 3-8 mm. long in fruit, at maximum not twice the length of Fre calyx, . Pedicel calyx, €. Pedicel in fruit shorter or subequalling the calyx. ......---- Ro tS CO 5. R. disperma. * Continued from volume 51, p. 520. 1971] KAZMI, BORAGINACEAE 111 e. Pedicel in fruit always longer than the calyx, f. Calyx lobes, in fruit, longer to much longer than the nut- lets; style exceeding the nutlets by 0.5-1 mm. .......... POR Ree ape I oi ae ie wR eee ee 6. R. stylaris. f. Calyx lobes in fruit usually shorter than or equal to, rarely longer than the nutlets; style exceeding the nutlets by 0.5 mm. g. Calyx lobes linear in fruit, to 0.3 mm. broad, curved, bearing usually straight, sometimes slightly curved, tri- chomes; nutlets 2—2.5 mm. long, attached subhorizontal- Wii eh re oe he ee RL ee epee 7. R. bungei. Calyx lobes oblong-lanceolate in fruit, 0.5-0.8 mm. broad, usually erect, sometimes curved, bearing usually uncinate, sometimes curved, trichomes; nutlets 3-3.5 mm, long, attached vertically or slightly tilted. ........ EVER ERE S CVE VS a alae ees Soe 8. R. laxa. by. 1. R. leiocarpa Ledeb. FI. Alt. 1: 172. 1829; DC. Prodr. 10: 176. 1846; Boiss, Fl. Orient. 4: 245. 1875; M. Pop. Fl. URSS 19: 557, 1953. R. incana Kar. & Kir. Bull. Soc. Nat. Moscou 15: 710, 1841. R. stellulata var. Kar. & Kir. loc. cit. R, lissocarpa Dunn, Kew Bull. 1924: 386. 1924, syn. nov. Type: Hab. in sterilibus ad fl. Kurtschum deserti Soongaro-Kirghisici orientalis, without citation of collector’s name (8). Icon.: Ledeb. Icon. Pl. Fl. Rossica 3: t. 244. 1847; M. Pop. l. c. 19: t, 27. fig. 4. 1953. Annual, erect herb, branched near the base, 5~25 cm. tall. Stem and branches slender, brownish, covered loosely with short, appressed, antrorse, white trichomes. Basal leaves sessile, linear, oblong-ovate or oblong— spathulate, entire, obtuse, 10-25 mm. long, 1-2 mm. broad, covered on both surfaces with thin, spreading white trichomes 0.2-0.7 mm. long, arising from large tuberculate bases, lower surfaces sometimes less hairy or rarely subglabrous; cauline leaves gradually decreasing in size upwards. Inflorescence lax, unilateral, short in flower, elongated in fruit, bracteate, bracts lanceolate, acutish. Flowers subsessile, fruits pedicellate, pedicels up to 2.5 mm. long, deflexed. Calyx divided nearly to the base, lobes lanceolate, erect, 1-1.5 mm. long in flower, linear, elongated to 4 mm., curved their whole length, pubescent in fruit, trichomes straight to slightly curved or sometimes uncinate. Corolla blue, campanulate, slightly nar- Towed at the throat, tube nearly equal to the calyx, lobes ovate to rotun- date, erect or patent, 0.5 mm. long. Nutlets 2, yellowish- to pinkish-white, glabrous, smooth, shining, 2 mm. long. Style exceeding the nutlets by 0.5 mm. DISTRIBUTION: Soongaria, Turkestan, West Pakistan, Kashmir. West Pakistan: Gricir AcENcy: Gurais valley, Minimorg, 3000 m., Jnayat 25716 (xk); Baltistan, upper end of Hushe valley, Chondakoro glacier basin, at 112 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 the foot of the Masherbrum, hillside near base camp Shahircho, 3650 m., Web- ster & Nasir 5974 (GH, Raw). Kashmir: Upper KISHENGANGA VALLEY: Purane Tilal, R. R. Stewart 4398 (RAW, K-type of R. lissocarpa Dunn). The specimens of Rochelia leiocarpa are usually mixed with a R. bungei in the collections cited above. Rochelia leiocar pa mee a Ri related to R. bungei from which it can easily be distinguished by smooth, glabrous, shining nutlets. 2. R. cardiosepala Bunge, Mém. Acad. Sci. St. Pétersb. sai - 7: 420. 1851; Boiss. Fl. Orient. 4: 246. 1875; C. B. Clarke in per f. Fl. Brit. India 4: 166. 1883; M. Pop. Fl. URSS. 19: 563. 1953; Riedl in Rechinger, Fl. Iranica 48: 90. 1967. Type: In rupibus graniticis ad Bakali Turkestaniae, Lehmann s.n. (18). Icon.: M. Pop. 1. c. 19: t, 27. fig. 1. 1953. Annual, up to 35 cm. tall. Stem and branches covered loosely or densely with antrorsel : line y appressed, short, white trichomes. Basal and zit oo faves to 80 mm. long and 7 mm. broad, oblong-ovate to spathulate, ’ di to patent. Nutlets 2, grayish-green, to 4 mm. long, Stellate-papillose. : h- DISTRIBUTION: Iraq, Iran, Afghanistan, West Pakistan, Kashmir, nee west India; Caucasus, Tur estan, Pamir Alaj, Kizil Kum, Syr Daty® Amu Darya, Dsungaro-Tarbagatai, Tien Shan. T West Pakistan: Currpar State: Drosh, 1500 m., Stainton 2277 (BM). yi StaTE: Jabba Valley, Rechinger 30733a, 30761 (w); between Saidu Shari ais oe tee 5 miles from Baherain, Kazmi s.n. (PEs). Reported a Nall SENCY: Astor district, Gudai i Marys Dade Sie Tict, Gudai, R. R. Stewart; Baltistan, naps, purane Tilal, 2600 m., RR. & 7. D. Stewart 4448 (K, MICH); Srinagar, 1709 m., R. R. Stewart 12546 (cH). 3. R. macrocalyx Bun 7: 419, 1851; Bois 19: 562, 1953+ Ri ge, Mém. Acad. Sci. St. Pétersb. Savants se S. Fi. Orient. 4: 246, 1875; M. Pop. Fl. ¥ edl in Rechinger, Fl. Iranica 48: 92. 1967. 1971] KAZMI, BORAGINACEAE 113 R. rectipes Stocks in Jour, Bot. Hook. 4: 176. 1852; C. B. Clarke in Hook. f. Fl. Brit. India 4: 166. 1883. Type: In rupestribus granites Turkestaniae ad Balsali, Lehmann s.n. (LE). Icon.: M. Pop. Fl. URSS 19: #. 27. fig. 3. 1953. Annual, erect or decumbent herb, up to 15 cm. tall. Stems many, much branched, branches short, both stem and branches usually covered with straight, sometimes curved, thin, spreading trichomes to 1 mm. long, which may arise from conspicuous tuberculate bases on the lower part of the stem. Basal leaves obtuse or roundish, slightly attenuated at base, to 1 cm. long and 1.5—2.5 mm. broad, covered on both sides with trichomes like those of the stems, arising from tuberculate bases; cauline leaves narrower, usually longer than the basal leaves. Inflorescence, lax, few- flowered, bracteate. Pedicels in flower short, in fruit elongated to 7-12 mm., narrowed towards the base, hairy, trichomes spreading. Calyx divided to the base, lobes lanceolate, acute, 6-7 mm. long, 1.5—2 mm. broad in fruit, narrowed towards the base into a short claw, usually 0.5—1 mm. long, stout, midrib prominent below, covered on both sides with short, erect trichomes (sometimes arising from tuberculate bases). Corolla blue, cylindrical, narrowed at the throat, ca. 1.5 mm. long, lobes short, suberect. Nutlets bluish-white to light brown, ca. 3.5 mm. long, 1.5-2 mm. broad at base, densely stellate-papillose. Style exceeding the nutlets by ca. 1 mm. DistripuTion: Iran, Afghanistan, West Pakistan, Kashmir, northwest India, Pamir Alaj, Kizil Kum. West Pakistan: Katat STATE: near Surab, Kazmi 1303b (pes); Rechinger 28329 (w). Quetta Dist.: 5 miles west of Ziarat, Kazmi s.n. (pes); Ziarat Proper, 2600 m., Jafri & Akbar 2124 (£); Spin Karez, Kazmi 1618b (PEs); Rechinger 29213 (w). Reported from: Quetta: Inter Bostan et Saran Tangai, 1800 m., Rechinger 29174 (w); in jugo W. Ziarat in junipereis, 2200 m., Rech- inger 29335 (w); Baluchistan, Gurghina, Stocks s.n. Kashmir: Zaskar, Char, Ichor, 4300 m., Koelz 5663 (GH, MICH, US); Kargia, Kargah, 4500 m., Koelz 5581 (GH, MICH, US). 4. R. peduncularis Boiss. Diagn. Pl. Orient. 7: 35. 1846; Fl. Orient. 4: 246. 1875; M. Pop. Fl. URSS 19: 561, 1953; Riedl in Rechinger, Fl. Tranica 48: 91. 1967. Type: In humidis alpis Kuh Delu, Persiae australis Kotschy 531a (w). Icon.: M. Pop. Fl. URSS 19: #. 27. fig. 2. 1953. Annual, erect herb, to 20 cm. tall. Stems simple or branched above, covered with short, white, subpatent trichomes. Basal leaves few, sessile, oblong, obtuse or roundish, entire, sometimes slightly revolute at margins, to 2 cm. long and 2-3 mm. broad, both sides covered with, thin, short, spreading white trichomes, usually arising from tuberculate bases; cauline €aves narrower and longer, upper cauline leaves shorter and broader, usually acute. Inflorescence lax, few flowered, bracts minute. Pedicels 114 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 DISTRIBUTION: Iran, Afghanistan, West Pakistan, Kashmir ?, Pamir Alaj, Tien Shan. West Pakistan: CurrraL Strate: Lutko, 2550 m., Bowes Lyon 879 (BM). Quetta Dist.: Quetta to Ziarat, at the top of Pass, 2200 m., Kazmi 1660a E); Rechinger 29317 (w); Spin Karez, 1900 m., Kazmi 1632b (PES); Rechinger 20210 (w) ~ 5 ad & 3 S a = s S = 5. R. disperma (L. f.) C. Koch, Linnaea 22: 649. 1849, emend. Kary- agin, Isv. Akad. Nauk Azerb. USSR 12: 13, 14. 1945; M. Pop. Fl. URSS 19: 551. 1953: Riedl in Rechinger, Fl. Iranica 48: 94. 1967. Lithospermum dispermum L. f. Descr. Pl. Rar. 13. 1762. Rochelia stellulata Boiss. Fl. Orient. 4: 244. 1875, pro parte non Reichenb. 824. Type: “Habitat in Hispania inter Gaditanum et Madritium,” without citation of collector’s name, Herb. No. 181.11, 181.12 (LINN). Icon.: L. f. 1. c. #. 7. 1762. Annual, erect or decumbent, to 20 cm. tall. Stem usually branched at the base or sometimes above, branches long, slender; stem and branches ; lower cauline leaves linear-lanceolate, obtuse, 12-30 mm. long, 13.5 mm. broad; upper cauline leaves gradually decreasing in size, more ort in flower, later elongated to 15 cm., bearing S-L.7 mm. long, lobes small, ovate to rounded, suberect. Nutlets 2, Sone, white, stellate-papillose. Style exceeding the nutlets by 0.5- mm. 1971] KAZMI, BORAGINACEAE 115 DistRIBUTION: Turkish Armenia, Caucasus, Iraq, Iran, Afghanistan, West Pakistan, Kashmir, Turkestan. West Pakistan: KurraAm AcGENcy: Kurram Valley, Parachinar, Kazmi s.n Khel, Kazmi 185b (pes). Reported from: South Waziristan: Kaniguram, J. L. Stewart s.n.; Baluchistan, Stocks s.n.; Gitcrr AGENcy: Naltar Valley, R. R. Stewart 26339; Baltistan, Ludlow s.n. Kashmir: Reported from: Nubra, Herb. Ind. Or. Hook. f. & Thoms., Thom- SOM S.N, Closely related to Rochelia disperma is R. bungei from which the former is differentiated by its pedicels always shorter or sometimes sub- equal to the calyx in fruit, its much longer fruiting calyx lobes (up to 5 mm.) and its subhorizontally attached nutlets. In R. bungei the fruiting pedicels are always longer than the calyx, the calyx lobes in fruit do not exceed 2 mm. in length, and the nutlets are attached horizontally. Rochelia stellulata Reichenb. in Hook. f. Fl. Brit. India 4: 166. 1883, is a misidentification. 6. R. stylaris Boiss. Fl. Orient. 4: 245. 1875; C. B. Clarke in Hook. f. Fl. Brit. India 4: 166. 1883; Riedl in Rechinger, Fl. Iranica 48: 93. 1967. Type: Afghanistan: circa Pusht, Griffith s.n. (x-holotype, w-isotype). Icon.: Riedl in Kgie & Rechinger, Biol. Skr. 13(4) : fig. 196, 197. 1963. Annual, usually erect, sometimes decumbent herb, to 15 cm. tall. Stems many, branched, branches long, spreading, stem and branches covered with short, thin, subappressed trichomes. Basal leaves many, petiolate, oblong-spathulate, obtuse to rotundate, entire, gradually narrowed toward the petioles, petiole longer, equal to, or shorter than the lamina, (including petioles) to 35 mm. long and 5 (—6) mm. broad, both surfaces covered with thin, short, patent, white trichomes arising from prominent tuberculate bases; middle cauline leaves sessile, linear to linear-oblong; upper cauline leaves much shorter and narrower. Inflorescence lax, bearing many uni- lateral flowers, much elongated in fruit. Pedicels in flower 1-1.5 mm. long, hairy, erect, to 6 mm. long, slender, pubescent, usually recurved in fruit. Calyx divided nearly to the base, lobes lanceolate, 1-1.5 mm. long in flower, nearly equalling the corolla, pubescent, erect, linear, curved inside, especially at the tips, spreading, to 4 mm. long in fruit. Corolla bluish- white to blue, equal or slightly exceeding the calyx lobes, ca. 2 mm. long, cylindrical, lobes short, suberect. Nutlets 2, grayish, 2-2.5 mm. long, recurved, attached subhorizontally, much shorter than the enlarged calyx, stellate-papillose. Style exceeding the nutlets by 0.5—0.8 mm. 116 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 DistriputTion: Afghanistan, West Pakistan, Kashmir, northwestern ndia. (pM). JueLUM Dist.: Salt Range, Sone Sihian Valley? (Sayedan Valley), Fleming 29 (e). Kuyper AGENCY: Ali Masjid, Kazmi s.n. (PEs); near Landi Kotal, Kazmi 2500 (pes); Khyber Pass, Torsapper, 1760 m., H. H. Johnston 60 (&); Khyber Pass, H. H. Johnston 13, 44, 45 (E). Dry rocky area near Afghan border, R. J. Rodin 2563 (us). KurrAm AGENCY: south of Parachinar, Kazmi 2506 (prs). PesHawar DIst.: near Palosi village, Kazmi 2489 (PES); be- tween Peshawar and Pabbi, Kazmi 2525 (PES); Peshawar, J. L. Stewart 137 Stewart 27420 (Raw); Manglaur, 1000 m., R. R. Stewart 27420a (BM). WAzIRIS- TAN AGENCY: Miran Shah, Kazmi sn. (PEs). Reported from: Quetta s.m.; Bolan Pass, Rechinger 28423; Kurram Valley, Shalizan, Aitchison 140; GiLerr Acency: Bagrot Valley, Conway s.n.; Nakao s.n.; Hunza State, Russell SN. shmir: Puric: Parkochin, Koelz 5993B (micH); Srinagar, R. R. Stewart 11088 (GH); between Kohala and Muzzaffarabad, Kazmi 2518 (pes); Muzza- ffarabad, Kazmi 2535 (PES). Miscellaneous: Kashmir, Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n, (GH); Kashmir (22. 4. 1848), Thomson s.n. (kK). Re- ported from: Gulmarg, R. R. Stewart s.n. ; Ladak, Lamayurun, Osmaston 139. 7. R. bungei Trauty. Acta Hortj Petrop. 9(2): 462. 1886; M. Pop. Fl. URSS 19: 554. 1953: Riedl in Rechinger, Fl. Iranica 48: 93. 1967. Type: Turcomania: Kisil-Arwat, Becker s.n. (LE). Annual, usually simple, erect, but sometimes sparsely branched and decumbent herbs, 8-20(-30) cm. tall. Stem and branches slender, covered on the lower part with patent, on the upper part with subappressed to appressed, thin, white trichomes to 0.7 mm. long. Basal leaves obovate, roundish at apices, entire, attenuated towards the bases, 3—5 mm. long trichomes, arising from tuberculate bases; middle cauline leaves oblong 1971] KAZMI, BORAGINACEAE 117 long, lobes short, suberect. Nutlets 2, yellowish-white, 2-2.5 mm. long, nearly horizontal, stellate-papillose. Stigma exceeding the nutlets by 0.5— 0.7 mm. DIstTRIBUTION: Iran, Afghanistan, West Pakistan, Kashmir, Turkestan, Pamir Alaj, Kara Kum, Kizil Kum, Syr Darya, Caspian Arals, Dsungaro- Tarbagatai, Tien shan. West Pakistan: GiLcir AcENcy: Karakorum range, Baltistan, upper end of Hushe Valley, Chandakore glacier basin at the foot of the Masherbrum, Sha- hircho, 3600 m., Webster & Nasir 5974 (cH-mixed with R. leiocarpa) ; Dirran, upper branch of Bagrot Valley, Conway 332 (kK); Hunza Valley, Zangia Harar, 4000 m., Russell 1068 (Bm). Kashmir: Zanskar, Kargia, 4500 m., Koelz 5582 (cu, MICH, US). Rochelia incana Kar. & Kir. sensu Bunge (in Reliq. Lehmann. 243. 1851; Mém. Acad. Sci. St. Pétersb. Savants Etr. 7: 412. 1852; Boiss. Fl. Orient. 4: 244. 1875) is a misidentification. 8. R. laxa I. M. Johnston, Jour. Arnold Arb. 21: 55. 1940. Type: Kashmir: Zanskar, Bok, 3450 m., camp ground, Sept. 13, 1931, Walter Koelz 2946a (cu-holotype), 29466 (cH-isotype). Annual, erect herb to 20 cm. tall with much branched roots. Stem usually branched near the base, branches long, divaricate; stem and branches covered with thin, erect to suberect, white trichomes, on the lower parts usually arising from white tuberculate bases. Basal leaves ovate, obovate or oblanceolate, obtuse to roundish at apices, entire, at- tenuated towards the bases, 5-20 mm. long, 3~5 mm. broad, both surfaces covered with short, thin, white trichomes arising from white tuberculate bases; middle cauline leaves longer; upper cauline leaves shorter and narrower, covered more densely with comparatively longer, thicker tri- chomes. Inflorescence lax, unilateral, bracteate, bracts minute, ovate- lanceolate. Pedicels in flower lacking to 0.5 mm. long, in fruit to 8 mm. ong, narrowed towards the base, covered with spreading, uncinate tri- chomes. Calyx divided nearly to the base, lobes in flower 1 mm. long, lanceolate, densely pubescent, in fruit oblong-lanceolate, erect to deflexed, 2.5—4 mm. long, 0.50.8 mm. broad, usually shorter or sometimes equalling the nutlets, rarely exceeding them by 0.5-1 mm., covered with white, uncinate trichomes, arising from conspicuous tuberculate bases. Corolla blue, ca. 2 mm. long, tubular, tube equalling the calyx, lobes ovate, sub- erect, 0.5 mm. long. Nutlets 2, white, erect or slightly tilted, tuberculate, 3-3.3 mm. long. Style exceeding the nutlets by 0.2-0.5 mm. DistrisuTion: West Pakistan, Kashmir, northwest India. West Pakistan: Cu1rraL STATE: Lutko, 2550 m., Bowes Lyon 789 (pm). Git- Gir AcENcy: Baltistan, Marpu Nullah, 3600-4000 m., Duthie 11839 (x); Naltar, 3300 m., R. R. Stewart 26336, 26339 (BM). Kashmir: Mitsahoi, Ladak Road, 3150 m., R. R. Stewart 10002a (cH); Rach-. 118 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 52 ogba, Rupshu, 4020 m., Koelz 2104 (cH); Bok, Zanskar, 3450 m., Koelz 2946a, 2946b (cH-type). I, M. Johnston (1. c.) considered Rochelia laxa closely related to R. macrocalyx and distinguished it from the former species by its much more loosely branched habit, elongate inflorescence, deflexed spreading rather than ascending pedicels, and its coarser obtusish calyx lobes, which only slightly exceed the nutlets. Rochelia laxa appears to me to be very closely related to R. peduncularis rather than to R. macrocalyx. Rochelia macro- calyx is unique among all the species of the genus in having unguiculate sepals, each with a very strong midrib. The calyx of R. laxa is quite different and resembles very much that of R. peduncularis in its non- unguiculate, lanceolate and erect calyx lobes. From R. peduncularis it differs in pedicels lacking or to only 0.5 mm. long in flower and much shorter in fruit, at the minimum not twice the length of the calyx. Calyx lobes in R. axa at the fruiting stage are 2.54 mm. long, while in R. ped- uncularis they are 4-6 mm. long. In the former species the calyx lobes hardly exceed the nutlets in length, while in the latter they usually exceed the nutlets by 2~4 mm 19. Asperugo L. Sp. Pl. 198. 1753; Gen. Pl. ed. 5. 67. 1754. TYPE SPECIES: A. procumbens L. Annual procumbent herb. Flowers small, axillary, subsessile, solitary or in pairs. Calyx in flower divided to the base, much enlarged in fruit, lobes joined below the middle, compressed, Corolla blue, subcylindrical; faucal appendages distinct, semiorbicular, Stamens included. Gynobase columnar, with 4 elevations. Style short. Nutlets 4, ovoid, strongly flattened with the small circular areola above the middle and clearly to one side of the medial line. One species, European, widely disseminated as a weed. A. procumbens L. Sp. Pl. 138. 1753; DC. Prodr. 10: 146. 1846; Boiss. Fl. Orient. 4: 275, 1875; C. B. Clarke in Hook. f, Fl. Brit. India 4: 167. 1883; Brand, Pflanzenr. IV. 252 (Heft 97): 23. 1931: M. pig URSS 19: 530. 1953; Ried] in Rechinger, Fl. Iranica 48: Type: Habitat in Europae ruderatis pinguibus, Herb. No. 189.1 (LINN). PP Set ee Icon. Fl. Germ. 18: #, 126. 1858; Brand, 1. c. fig. 1, Annual, procumbent herb. Stems many, to branched; stem and branches with very variable, oblong-lanceolate to 1971] KAZMI, BORAGINACEAE 119 tuberculate bases. Flowers small, subsessile, solitary or in pairs. Pedicels in fruit 2~7 mm. long, usually reflexed. Calyx divided nearly to the base in flower, ca. 1.5 mm. long, enlarged in fruit to 15 mm., lobes joined below the middle, with few scattered spinules on the outside, glabrous inside. Corolla blue, 1.5—2 mm. long, lobes rounded, suberect; faucal appendages whitish. Nutlets ovoid, compressed, brownish, minutely verruculose, ca. 2.5 mm. long. DistrIBuTION: Asia and Europe. West Pakistan: Arrocx Dist.: Campbellpore, Zftakhar s.n. (PEU). CHITRAL STATE: Brumboret Gol, 2000 m., Bowes Lyon 650 (BM); Ziarat, Lowari Pass, 2400 m., Stainton 2340 (pm). Gitcrr AGENCY: Shingo Valley, Duthie s.n. (E); Kowatr Dist.: 4 miles from Kohat, on Kohat-Bannu Road, Kazmi s.n. SN, Kashmir: Chinari, Kazmi 526a (pes); Muzzaffarabad, near the bridge, Kazmi sm, (PES); Purig, Tangola, Koelz 6090 (us). Reported from: Poonch, R. R. Stewart s.n.; Murpur, R. R. Stewart s.n. 20. Omphalodes Moench, Meth. 419. 1794. Type species: O. linifolia (L.) Moench. Annual or perennial strigose or villous herbs. Leaves alternate or some- times the lower opposite. Inflorescence of simple or branched racemes. Calyx divided to the base. Corolla funnelform, campanulate or salverform, lobes patent; faucal appendages, distinct, semilunar. Anthers ovate, in- cluded; filaments short. Nutlets depressed-spherical to ovoid, externally umbilicate, dentate to glochidiate, margins usually incurved, attached longitudinally to and exceeding the gynobase. Species about 24, distributed in Eurasia and Mexico. O. heterophylla Rech. f. & H. Riedl in Oesterr. Bot. Zeitschr. 110: 532. 1963; Riedl in Rechinger, Fl. Iranica 48: 98. 1967. Type: West Pakistan: Swat: Himalaya bor. occid. Kalam, ca. 35° 30° N, 72° 30’ E, Substr. granite, ca. 2200 m., 22. 8. 1962, K. H. Rechinger 19442 (w-holotype). Perennial, tufted, prostrate to rarely ascending herb. Stems numerous, simple to much branched, covered with white spreading trichomes, 7-35 cm. long. Basal leaves lanceolate, petiolate, petioles nearly 10 mm. long, 120 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 52 lamina 30-55 mm. long, 9-18 mm. broad, acute, one nerved, covered on both surfaces with rigid, appressed, white trichomes, arising from distinct tuberculate bases; cauline leaves much smaller than the basal leaves, oblong to oblong-lanceolate, sessile to shortly petiolate, 10-25 mm. long, 3-8 mm. broad; upper cauline leaves, sessile, auriculate. Inflorescence lax, short in flower, elongated in fruit, bracteate, lower bracts 10-18 mm. long, acute, the upper ones 2-10 mm. long, lanceolate, indistinct in the upper- most flowers. Pedicels 2-4 mm. long in flower, recurved, 4-9 mm. long, in fruit. Calyx divided to the base, lobes 2—2.5 mm. long in flower, to 4 mm. long in fruit, patent, incurved at apex. Corolla campanulate, blue, 3 mm. long, slightly exceeding the calyx, corolla tube equalling or slightly shorter than the calyx, lobes suborbicular, patent, 1~1.5 mm. long, faucal appendages much broader than long, incurved. Anthers minute, filaments scarcely distinct, inserted at the middle of the corolla tube. Style 1- 1.5(-2) mm. long, stigma capitate or indistinctly bifid. Nutlets ovoid, subacute, 2 mm. long, margins swollen, dorsal face more or less plane, sparsely tuberculate-aculeate, ventrally densely tuberculate, carinate below and sometimes also above the areola, areola orbicular at base, narrow at apex, ‘sii % length of the nutlet ventrally. Gynobase short, narrow, conical. DISTRIBUTION: West Pakistan. West Pakistan: CHirrAL State: Drosh, Beorai, 4000 m., Bowes Lyon 198 (BM, E). Swat State: Kalam, 2200 m., Rechinger 19442 (w-type). 21. Paracaryum (DC.) Boiss. Diagn. Pl. Orient. 11: 128. 1849. Omphalodes Moench, Sect. Paracaryum DC. Prodr. 10: 159. 184 6. Omphalodes Moench, Sect. Mathia DC. Prodr. 10: 169, 1846, pro parte. Tyre species: P. rugulosum (A.DC.) Boiss. (lectotype species). Annual, biennial or Basal leaves many, sessile. Inflorescence several bracts. Cal perennial, erect, decumbent or prostrate herbs. usually with long petioles; cauline leaves usually short in flower, later elongated, ebracteate to bearing ( yx divided to the base, lobes more or less enlarged in fruit. Corolla blue to various shades of purple-brown, infundibuliform, subcampanulate or cylindrical, lobes usually short; faucal appendages present. Anthers usually included, rarely with exserted apices, elliptical to linear; filaments distinct. Style usually shorter than the calyx; stigma capitate. Nutlets 4, dorsal middle area smooth, aculeolate or glochidiate, often carinate, margins winged, wings inflexed, entire to lacerate, bearing glochidia, membranous, attached ventrally throughout their length to the Narrow, conical gynobase Species about 15, distributed from northeast Africa to south and south- east Russia, Arabia, Iran, Afchani : ene northwest India. . » 4lgnanistan, West Pakistan, Kashmir a 1971] KAZMI, BORAGINACEAE 121 KEY TO THE SPECIES q. Species Selppemt nutlets 5-7 mm. lon b m. long; pedicels in fruit ca. 4 mm. long; calyx lobes 4 mm. long, oe cieee not enlarged in fruit; corolla cylindrical-campanu- a ee er ee rate Fe eh os te ess 1. P. rugulosum, Nutlets 6-7 mm. long: pedicels 7-12 mm. long in fruit; calyx lobes ca. 4 mm. long in flower, enlarged in fruit to 6 mm., lanceolate: corolla infundibuli COTE sae Adee ie Fare AB ee 2:+f, pla tycalyx. Species annual; nutlets 2.5~-4(—5) mm. long. c. Nutlets ovoid, 4-5 mm. long; basal leaves distinctly petiolate. ........ ee ee eee ee Pee oes Ce ere nea. 3. P. calathicarpum. c. Nutlets round, 2.5-4 mm. in diameter; basal leaves ee petiolate. TORN ce 8) RE ORS CEI 914 23 I ep ae 4. P. intermedium. o » 1, P. rugulosum (DC.) Boiss. Diagn. Pl. Orient. 11: 129. 1848; FI. Orient. 4: 256. 1875; Aitch. Bot. Afgh. Delim. Comm. 89. 1888; Brand, Pflanzenr. ee 252 (Heft 78): 48. 1921; Riedl in Rechinger, FI. Tranica 48:99, 1967. Omphalodes rugulosa DC. Prodr. 10: 160. 1848. Paracaryum rubriflorum Stocks in Jour. Bot. Hook. 4: 175. 1852, Type: In monte Horeb Arabiae Petreae, Aucher s.n. (G). Icon.: Brand, 1. c. fig. 5, A-F. 1921 Perennial, erect herb, 10-40 cm. tall. Stems solitary or few, branched, branches long; stems and branches densely covered with thick, usually retrorse, subpatent, white to grayish trichomes to 2 mm. long, arising from tuberculate bases. Basal and some of the lower cauline leaves petiolate, linear-lanceolate to oblong-lanceolate, acute, margins entire, (including petioles) 40-170 mm. long, 2-15 mm. broad, densely covered on both surfaces with trichomes like those of the stem; cauline leaves sessile, shorter and narrower. Pedicels to 2 mm. long in flower, to 6 mm. long, densely pubescent, recurved in fruit. Calyx divided to the base, lobes in flower ca, 4 mm. long and 1 mm. broad, oblong, densely pubescent with retrorse, subpatent trichomes, to 5 mm. long and 2 mm. broad in fruit. ovate to ovate-lanceolate, acutish, subpatent, rarely reflexed. Corolla purple-violet to bluish-violet, cylindrical-campanulate, 4-5 mm. long, tube more or less equalling the limbs, limbs divided to the middle into more or less patent lobes; faucal appendages present. Style 2-3 mm. long. Nutlets to 6 mm. long and 5 mm. broad, ovate, broadly winged at the margins, wings incurved, rugulose, irregularly dentate on the inner margin, dorsally sometimes ruberculates dorsal middle area smooth or bearing few vertical appendages, appendages at apex glochidiate. DistripuTIon: Egypt, Arabia, Iran, West Pakistan. West Pakistan: QueTTA Dist.: Quetta, Dick-Peddie sade (micH); Gwal, 2000 m., 24. 4. 1888, Lace s.n. (E); Zarghun, 2200 m., Lace 3731 (E). Miscellaneous: Baluchistan, Herb, Ind. Or. Hook. f. & Thoms., Stocks s.n, (GH-isotype of P. r briflorum Stocks), Reported from: Quetta, Khanozai, Duthie (?) 18924 on, 122 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 2. P. platycalyx Riedl, Oesterr. Bot. Zeitschr. 110: 535. 1963; in Kgie & Rechinger, Biol. Skr. 13(4): 205. 1963; in Rechinger, Fl. Iranica 48: 102. 1967. Type: Afghanistan: Herat, Jija, 900 m., Kgie 4404 (w-holotype, c- isotype). Icon.: Riedl, 1. c. fig. 156. 1963. Perennial, erect to ascendent, 15-25 cm. tall. Stems many, branched, covered densely with up to 2 mm. long, thick, patent, white trichomes arising from minute tuberculate bases. Basal leaves petiolate, (including petioles) 70-130 mm. long, 7-10 mm. broad, petioles usually longer than the lamina, lamina lanceolate, tapering at both ends, acutish, margins plane, covered on both surfaces with subpatent, more or less retrorse trichomes, like those of the stem; cauline leaves sessile, sometimes more or less semiamplexicaul, gradually reduced in size upwards. Pedicels 2 mm. long in flower, to 12 mm. long in fruit, more or less recurved. Calyx divided to the base, lobes oblong-ovate, densely hairy, in flower ca. 4 mm. long, 1 mm. broad, enlarged in fruit to 6.5 mm. long and 2.5 mm. broad, lanceolate, spreading. Corolla blue to purplish-blue, infundibuliform, 6—7 (-8) mm. long, tube longer than limb, lobes broad, rounded; faucal ap- pendages broader than long. Style 2-3 mm. long. Nutlets up to 6(—7) mm. long, 5—6 mm. broad, ovate, broadly winged at the margins, wings incurved, rugulose, irregularly dentate on the inner margins, middle dorsal area indistinctly carinate, smooth or bearing few vertical appendages, append- ages glochidiate at their apices. DISTRIBUTION: Afghanistan, West Pakistan. Paracaryum platycalysx is very closely related to P. rugulosum in the form of its leaves, flowering parts, and nutlets and can only be differen- tiated from the former s 5. P. calathicarpum (Stocks) Boiss. FI. Orient. 4: 254. 1875; Brand, Pflanzenr. IV. 252 (Heft 78): 48. 1921; Riedl in Rechinger, Fi. Tranica 48: 104, 1967, Echinospermum calathicarpum Stocks in Jour. Bot. Hook. 4: 175. 1852. 1971] KAZMI, BORAGINACEAE 123 Type: Baluchistan superior, Stocks 1003 («-holotype, G-isotype). Icon.: Riedl in Kgie & Rechinger, Biol. Skr. 13(4): 204. fig. 154, 155. 1963; in Rechinger, Fl. Iranica, l. c. ¢. 19. fig. 2. 1967. Annual, erect, up to 25 cm. tall. Stem usually solitary, much branched, branches long, divaricate; stem and branches covered with thin, crisped, subappressed, antrorse trichomes, to 0.5 mm. long, arising from minute tuberculate bases. Basal leaves oblong-spathulate, petiolate, (including petioles) to 50 mm. long and 8-10 mm. broad, apices obtuse to roundish, gradually narrowed towards the base into a short petiole, loosely covered on both surfaces with very short, crisped, patent to subappressed white trichomes, arising from large tuberculate bases; lower cauline leaves very short petiolate, the upper sessile, gradually reduced in size. Pedicels in flower lacking to 0.5 mm. long, in fruit to 4 mm. long, stout, pubescent, more or less reflexed. Calyx divided to the base, lobes oblong, acute, pu- bescent, 1-1.5 mm. long in flower, enlarged in fruit to 2.5-3 mm. long, spreading. Corolla blue, campanulate, 1.5—-2 mm. long, limb shorter than the tube, lobes minute. Style ca. 0.5 mm. long, pubescent. Nutlets ovoid, 5 mm. long, 3.5—-4 mm. broad, margins broadly winged, wings inflexed, laterally blistered, blisters in many rows, those of the outer row slightly elongate to partly mammillate, dorsally rugulose, tuberculate, inner mar- gins irregularly dentate, middle dorsal area tuberculate to vertically ap- pendiculate, appendages glochidiate at apex; one pair of opposite nutlets usually with narrower marginal wings. Distrisution: Afghanistan, West Pakistan. West Pakistan: MAKRAN Dist.: 70 km. northeast of Panjgur, 1100 m., Kazmi 12716 (PEs); Rechinger 28221 (w); Nag, between Panjgur and Surab, 1300 m., Kazmi 1278a (pes); Lamond 632 (£); Rechinger 28260a (c, w). Quetta Dist.: 20 xm. west of Quetta, Kazmi 1589b (PES); Rechinger 29081 (w). Miscella- inger 24844 (w); infra Mach versus Ab-e~Gum, Rechinger 23484 (w); P anjgur, 1000 m., Rechinger 28177(w). 4. P. intermedium (Fresen.) Lipsky, Acta Horti Petrop. 26(2): 487. 1910; Brand, Pflanzenr. IV. 252 (Heft 78): 47. 1921; M. Pop. Fl. URSS 19: 602. 1953; Riedl in Rechinger, FI. Iranica 48: 105. 1967. Cynoglossum intermedium Fresen. in Mus. Senckenb, 1: 169. 1834. Omphalodes intermedia Decne. Ann. Sci. Nat. Paris 2: 255, 256. 1834. Paracaryum die sa (DC.) Boiss. Diagn. Pl. Orient. 11: 129, 1849; Fl. Orient. 4: 255, 18 Omphalodes Sabicnae DC. Prodr. 10: 159. 1846. TYPE: not indicated. Icon.: Riedl, 1. c. 48: ¢. 19. fig. 4. 1967. Annual, erect to ascending herb, 5—30 cm. tall. Stems solitary to many, branched near the base, branches long, delicate, more or less spreading, 124 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 covered with crisped patent to subappressed white trichomes to 0.5 mm. long, arising from minute tuberculate bases. Basal leaves usually obtuse sometimes acutish, entire, 20-40 mm. long, 3-6 mm. broad, attenuated towards the base into a very short to indistinct petiole, covered on both surfaces with thick, patent, white trichomes to 1 mm. long, arising from large, white, tuberculate bases; cauline leaves sessile, linear-oblong, grad- ually reduced in size upwards. Inflorescence long, lax, usually bracteate in the lower part, bracts 1-2. Pedicels very short in flower, elongated to 15 mm. in fruit, pubescent, suberect to horizontal or sometimes reflexed. Calyx divided nearly to the base, lobes 1-1.5 mm. long in flower, 2—3.5 mm. long in fruit, linear to oblong-ovate, obtuse to acute. Corolla blue, campanulate to cylindric-campanulate, 1.5-2.5 mm. long, limbs subequal to little shorter than the tube; faucal appendages semilunar to sub- quadrate. Anthers attached to the middle of the corolla tube. Nutlets roundish, 2.5-4 mm. in diameter, broadly winged at the margins, wings inflexed, leaving a small to large aperture in the middle, more or less longitudinally rugulose, sometimes with few tubercles or densely minutely papillate, inner margin forming the aperture lobed, irregularly dentate or pubescent, middle dorsal area aculeate or vertically appendiculate, ap- pendages glochidiate. DIsTRIBUTION OF SPECIES: Egypt, Arabia, Iran, Afghanistan, West Pakistan, Caspian Aral, Pamir Alaj, Kizil Kum, Tien Shan. KEY TO THE VARIETIES a. Inflexed marginal wings of the nutlets densely white papillate on the upper gt c uc Oa ee eee ee ee C. apillosum. a Inflexed marginal wings of the nutlets glabrous or rarely inconspicuously papillate. b. Aperture formed by the inflexed marginal wings of the nutlets very small, rico Ret ga a re . var. stellatum. b. Aperture formed by the inflexed marginal wings of the nutlets large, NO ee aes 4a. var. intermedium. 4a. Var. intermedium. Basal leaves linear-spathulate; Pedicels 5—15 mm. long in fruit. C panulate, limb subequalling the co the nutlets dorsally glabrous, margin forming the central aper lobed. DISTRIBUTION OF VAR.: as that of species Kew ee nt ky Dist.: 8 miles northwest of Harnai, 900-1200 m., (eay? Finda reek echinger 29556 (W); near Ziarat, 2400 m., Kazmi 1656b Hide ara ar 2145 (BM, E). Miscellaneous: Baluchistan, Herb. Ind. Or. ef oms., Stocks s.n. (GH-as Omphalodes micrantha), Reported from: 1971] KAZMI, BORAGINACEAE 125 (w); Karat Srate: Inter Kolpur et Mach, 900-1800 m., Rechinger 28432, 28437 (w). 4b. Var. stellatum (H. Riedl) Kazmi, comb. nov. Paracaryum stellatum H. Riedl in Rechinger, Fl. Iranica 48: 104. 1967. Type: West Pakistan: Kalat, inter Kolpur et Mach, 900 m., K. H. Rechinger 28434 (w). Icon.: Riedl, l. c. 48: ¢. 19. fig. 3. 1967. Basal leaves oblong to linear-oblong; cauline leaves linear to linear- lanceolate, acutish. Pedicels to 6 mm. long in fruit. Calyx lobes linear, acutish. Corolla campanulate, limbs subequal to the corolla tube. Inflexed margin of the nutlets, dorsally glabrous or rarely indistinctly papillose, margin forming the central aperture dentate, teeth triangular, central aper- ture stellate. DISTRIBUTION OF VAR.: West Pakistan, Afghanistan. West Pakistan: KaLat State: between Kolpur and Mach, 900 m., Rechinger 28438 (w-holotype); Kazmi 1331d (PEs). Quetta Dist.: Inter Qila Abdullah et Sheila Bagh, 1600-1900 m., Rechinger 29033 (w); Saranan, 8-9 miles north of Quetta, Kazmi 1501a (PES); Spin Karez, Kazmi 16156 (PEs). 4c. Var. papillosum (DC.) Kazmi, comb. nov. Omphalodes papillosa DC. Prodr. 10: 159. 1846. Paracaryum papillosum (DC.) Giirke in Engl. & Prantl, Nat. Pflanzenfam. IV. 3a: 105. 1893: Riedl in Rechinger, Fl. Iranica 48: 105. 1967. Type: Ad montem Sinai, Schimper 123 (c-holotype, cH-isotype). Basal leaves oblong to linear-oblong; cauline leaves linear to linear- lanceolate, acutish. Pedicels to 10 mm. long in fruit. Calyx lobes more or less linear, acutish. Corolla cylindrical-campanulate, broad at the mouth, limb much shorter than the corolla tube. Inflexed marginal wings of the _nutlets dorsally densely and minutely papillose, forming the central aper- ture at the margins, subentire, lobed or irregularly dentate, aperture broad. DIsTRIBUTION oF VAR.: as that of species. West Pakistan: Maxran Dist.: Near Nag, Kazmi 1258b (pes). Quetta Disv.: between Wam Tangi and Akhtari, Kazmi 1763 (pes); Wam Tangi, Munro s.n. (kK); Harnai, Lace 2699 (G); Ziarat, Santapau 6452 (GH). Reported from: Inter Kolpur and Mach, Rechinger 28433. Paracaryum intermedium is very variable in the form of its leaves, in the length of pedicels, and in the margin and surface of the inflexed wings of the nutlets. Because the distinguishing characters are not very constant 126 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 and since intermediate forms are found, I consider P. stellatum and P. papillosum only variations of P. intermedium. 22. Mattiastrum (Boiss.) Brand, Repert. Sp. Nov. 14: 150. 1915. Paracaryum Boiss. Sect. 2. Mattiastrum Boiss. Diagn. Pl. Orient. 11: 130. 1849, TYPE SPECIES: not indicated. Usually perennial or biennial, rarely annual herbs. Inflorescence termi- nal to axillary, ebracteate or rarely bearing 1—2 bracts. Calyx usually divided to the base, lobes 5, not at all to slightly enlarged in fruit. Corolla subcylindrical, to infundibuliform or hypocrateriform, limb obviously dis- tinct from the corolla tube, faucal appendages 5, conspicuously developed inside the corolla throat. Anthers included. Nutlets usually winged at the margins, sometimes the wings of the pair of two opposite nutlets reduced, wings plane, to rarely slightly incurved, never inflexed as in Paracaryum, middle dorsal area not covered by the wings, wings dentate to appendicu- late at the margins, appendages glochidiate, nutlets attached throughout their length to the narrow columnar gynobase. Species about 30, distributed in southwestern Asia. KEY TO THE SPECIES a. Calyx lobes 3-5 mm. long in flower; corolla 10-12 mm. oe dae is ee er ee ete hie OS es ke ko cae cs . M. asperum. a. Calyx lobes 1-2 mm. long in flower; corolla 2-4 mm. long. b. Basal leaves oe petioles) 100-140 mm. long, 30-45 mm. broad, ere Oy ere a ine ve . thomsonti. Basal leaves Pictoding petioles) to 90 mm. long and 15 mm. broad, oblong-lanceolate - oblong-spathulate. c. Nutlets 5-8 m d. Plants soreaing pubescent; pedicels to 1 mm. long in flower, to (-8) mm. long in fruit, slender; corolla blue; “ae 1.5-2 mm = a ee ns ete eee RS ys ee . M. ower d. Plants — pubescent; pedicels lacking in ria in fruit be sy ong, thick; corolla bluish-white to pink; style 0.5-1 mm pO eee eat oN AO 2. M. himalayensis. c hoe 2-4 mm. long. e. Plants erect to ascendent; stem stout, thick; basal leaves many; marginal wings of the nutlets plane at the edges, dentate, teeth blunt not glochidiate at apex............._ 5. M. tibeticum. - Plants erect to prostrate, shi weak, slender; basal leaves 1-2 oF missing; marginal wings of the nutlets recurved at the edges, dentate, teeth produced, glochidiate at apex. 6. M. bunget ia") 1. M. asperum (Stocks) Brand, Repert. Sp. Nov. 14: 153. 1915; Pflan- zenr. IV. 252 (Heft 78): 61. 1921; Riedl in Rechinger, Fl. Iranica 48: 114. 1967. 1971] KAZMI, BORAGINACEAE 127 Paracaryum cag Stocks in Jour. Bot. Hook. 4: 175. 1852; Boiss. Fl. Orient. 4: 261. Mattiastrum peel Rech. f. Ann. Naturh. Mus. ear 55: 14. 1947; Riedl in Rechinger, Fl. Iranica 48: 114. 1967, syn. no Type: Baluchistan superior, Stocks 906 (xk). Icon.: Riedl in Kgie & Rechinger, Biol. Skr. 13(4): 213. fig. 164. 1963 (under M. honigbergeri). Shrubby perennial, 20-60 cm. tall. Stems many, branched, branches long; stem and branches covered with crisped, patent, thick, white tri- chomes to 2 mm. long, arising from tuberculate bases. Basal leaves petio- late, (including petioles) to 100 mm. long, 12 mm. broad, oblong-ovate, obtuse, attenuated towards the base, both surfaces covered densely with short trichomes to 1 mm. long intermixed with thick trichomes to 2.5 mm. long, all arising from prominent tuberculate bases and more or less spreading; middle cauline leaves linear-oblong, sessile, little shorter than the basal ones; upper cauline leaves much shorter and narrower. Inflores- cence terminal or axillary, lax, much elongated in fruit. Pedicels 1-2 mm. long in flower, elongated to 6 mm. in fruit, usually erect or sometimes reflexed. Calyx divided to the base, lobes linear-oblong, acute, densely pubescent, 3-4(—5) mm. long in flower, slightly enlarged in fruit. Corolla purple-violet to blue, cylindrical-infundibuliform, 6-12 mm. long, limb more or less equalling the corolla tube, lobes short; faucal appendages semilunar. Nutlets (6—)8-10 mm. in diameter, broadly winged at the margins, wings dentate, teeth triangular, usually glochidiate at apex, dorsal middle area vertically appendiculate, appendages glochidiate. DIstRIBUTION: West Pakistan, Afghanistan. Pakistan: QuETTA Dist.: Urak, 2000 m., Santapau 6731 (cH); R. R. Sisari 691 (micH); Ziarat, Sandeman Tangi, 2 miles from Ziarat, 2500 m., Kazmi 1660 (PEs); Quetta to Ziarat at top of Pass, 2500 m., Lamond 1868 (£); Rechinger 29056 (w); Ziarat, Swiss Zool. Mission (May, 1953), sm. (BM); between Qila Abdullah and Sheila Bagh, Kazmi 1533 (PES); Bra i to Chaman, Khojak Pass, above Sheila Bagh, 2000-2300 m., Lamond 1068 (£); Rechinger 29056 (w); Ziarat, Santapau 6435 (cH). Miscellaneous: Baluchistan, in Herb. Ind. Or. Hook. f. & Thoms., Stocks s.n. (GH-probable duplicate of Stocks 906). Reported from: Quetta, Tak, Rechinger 28402 (w); Gwal versus Surkh, Kakar in valley Lora, Lace s.n.; Chiltan inter Dulai et Kanak, 1600 m., Rech- inger 29119 (w); Coeshtc, 2100-2400 m., Munro(?); Kawas Tangi, 1800 m., Munro(?); Zarghun, 1500 m., Lace 3731 in part (£). The only distinguishing character cited between Mattiastrum asperum and M. honigbergeri is the length of corolla which, in the former species is said to be 10 to 12 mm. and in the latter 6 to 8 mm. The specimens I examined, show a variation in the length of corolla between 6 to 12 mm. and M. honigbergeri represents the variation with shorter corollas. 2. M. himalayense (Klotzsch) Brand, Repert. Sp. Nov. 14: 156. 1916; 128 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Pflanzenr. IV. 252 (Heft 78): 65. 1921; Riedl in Rechinger, FI. Iranica 48: 116. 1967. Mattia himalayensis Klotzsch in Bot. Ergebn. Reis. Prinz Waldemar Bot. 94. t. 64. 1862. Paracaryum himalayense (Klotzsch) C. B. Clarke in Hook. f. Fl. Brit. India 4: 161. 1883; Giirke in Engl. & Prantl, Nat. Pflanzenfam. IV. 3a: 105. fig. 42, L-O 1893; M. Pop. in Fl. URSS 19: 597. 1953. Typr: Himalaya, Hoffmeister s.n. (HEID). Icon.: Riedl in Kgie & Rechinger, Biol. Skr. 13(4): 212. fig. 162, 163. 1963; Klotzsch, 1]. c. ¢. 64. 1862; Giirke, 1. c. fig. 42, L-O. 1893; Brand, l. c. fig. 10, L-O. 1921; M. Pop. 1. c. #. 30. fig. 1. 1953. Biennial to perennial, erect, to 60 cm. tall. Stems solitary to many, branched, branches long, more or less divaricate; stem and branches dense- ly covered with subpatent to subappressed white trichomes 0.5—1 mm. long, not arising from tuberculate bases, sometimes intermixed with thicker tri- chomes to 1 mm. long, arising from prominent tuberculate bases. Basal leaves (including petioles) 50-90 mm. long, 5-15 mm. broad, oblong, ob- long-lanceolate or lance-spathulate, obtuse, margins sometimes slightly revolute, densely covered on both surfaces with variable, very thin to quite thick, appressed to subpatent, white trichomes 0.5—1 mm. long, arising from tuberculate bases or not; cauline leaves sessile, lanceolate, much reduced in size. Inflorescence terminal and axillary, much contracted in flower, elongated in fruit, lax. Pedicel absent in flower, in fruit stout, hairy, erect or horizontal, rarely reflexed, usually 1-2 mm. long. Calyx divided nearly to the base, in flower 1.5-2 mm. long, in fruit elongated to 3 mm., ovate to oblong, obtuse, hairy. Corolla bluish-white to pink, 2—4 mm. long, cylindric-campanulate, tube slightly longer than the limb; faucal appendages trapeziform. Nutlets suborbicular, 5—7(—8) mm. in diameter, marginal wings broad, plane or slightly curved at the bluish, broadly dentate margins, teeth triangular, bearing glochidia at apex. Stigma 0.5- 1 mm. long. DISTRIBUTION OF SPECIES: Afghanistan, Pamir Alaj, West Pakistan, Kashmir, northwest India. 2a. Var. himalayense Plants biennial; stem solita Z ry, slender, less branched, branches not much spreading, West Pakistan: Gitcrr AcENcy: Baltistan, Chorbat, 3000 m., Hunter-Winston 10246 (K). Miscellaneous: Himalaya, Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (K-as Omphalodes no. 4). Kashmir: Reported from: Ladak, 4600 m., Meebold 4236. 2b. Var. fallax (Rech. f, & H. Riedl) Kazmi, comb. nov. Mattiastrum himalayense subsp. fallax Rech. f. & H. Riedl in Rechinger, Fl. Tranica 48: 117. 1967 1971] KAZMI, BORAGINACEAE 129 Type: Afghanistan: Bareki Chaidan, on the way to Band-e-Amir, 2800 m., Hedge & Wendelbo 4719 (Bc). Icon.: Riedl, 1]. c. ¢. 22. 1967. Plants perennial, stems many, stout, much branched, branches divaricate. West Pakistan: Gitcit AGENCY: Baltistan: Shagarthang Valley, 3000 m., Duthie 12, 140 (K). Kashmir: ZASKAR: Seni, Phe, 3600 m., Koelz 5821 (GH, MICH, US); Kargia, 4300 m., Koelz 5470 (GH, MIcH, US); LaDAK: Rabila, Koelz 6153 (MIcuH, Us). Riedl (1. c. 1967) recognized the perennial, many stemmed form as subspecies fallax, considering it to be confined to Afghanistan or possibly extending to Pamir Alaj and thus representing the western area of dis- tribution of the species Mattiastrum himalayense. The specimens I ex- amined show that this form extends to West Pakistan, Kashmir, and even to northwest India, [Kangra Dist.: Spiti Valley, Dankhar, 4000 m., Bhagwan Singh 256 (micH)| and covers the same area as the typical form. Therefore, I consider both taxa only varieties of M. himalayense. The diameter of the nutlets in variety fallax has been given by Riedl as 7 to 8 mm., which is so only rarely. Usually the nutlets are the same size in both varieties. 3. M. howardii ? Kazmi, sp. nov. Type: West Pakistan: Gilgit Agency, Dist. Astore, 7800 ft., 25. 7. 1892, J. F. Duthie s.n. (pM-holotype, £-isotype). Herba annua vel biennis (?), erecta ad 35 cm. alta. Caulis plerumque solitarius, interdum 2-3, striatus, ramosus, rami ad 20 cm. longi, plus- minusve divaricati; caules ramique pilis albis patentibus rigidis, ad basim manifeste tuberculatis. vestiti. Folia basalia petiolata, ad 40-60 mm. longa (petiolis inclusis) et 5-10 mm. lata, oblongo-ovata, obtusa vel subrotun- data ad apices, ad basim attenuata, supra infraque dense pilosa, pili tenues albi; petioli, praecipue ad margines, pilis ad 1.5 mm. longis; folia caulina media pauca, sessilia, lanceolata, acuta, basim apicemque atten- uata, quam folia basalia angustiora brevioraque, pilis longioribus sparse vestita, pili ad basim tuberculati. Inflorescentia ad anthesin brevis, in fructu perelongata, laxa, interdum 1—2-bracteata. Pedicellus ad anthesin + 1 mm. longus, hirsutus, in fructu perelongatus et usque ad 6(—8) mm. longus, tenuis, erectus, raro deflexus. Calyx ad basim divisus, lobi ad anthesin 1. 5-2 mm. longi, oblongi, in fructu accrescentes ad 3-4 mm. longi et 1 mm. lati, erecti. Corolla caerulea, ca. 2 mm. longa, cylindrico-cam- panulata, tubus quam limbus longior, lobi 0.7-8 mm. longi, ovati; ap- pendices faucales suborbiculares. Antherae 0.6 mm. longae, oblongae, ad medium tubi corollae insertae. Nuculae subrotundatae, 6-7 mm. diam ad margines alatae, alae ca. 2 mm. latae, ad margines complanatae, late *T am g this species in honor of Professor Richard A. Howard, Director of the Arnold. prorat of Harvard University. 130 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 dentatae, dentes ad apicem glochidiati; pars media dorsalis verticale ap- pendiculata, appendices apice glochidiatae, paginae laterales ventralesque minute tuberculatae. Stylus 1.5—2 mm. longus. DISTRIBUTION: West Pakistan. West Pakistan: Giicir AcENcy: Astore, 2500 m., Duthie s.n, (BM, E); Gurikot to Godai, on Gilgit road, 2650 m., R. R. Stewart 18930 (cH); Gurikot to Das Kirini, 2500-2700 m., R. R. Stewart 22979 (cx). The new species, Mattiastrum howardii, is closely related to Mattia- strum himalayense, from which it differs in the loose, always spreading, uniform trichomes on the stem, branches, and leaves; in the subsessile to short (1 mm.) pedicellate flowers, the slender, erect or rarely slightly reflexed pedicels elongating to 6(-8) mm. in fruit; and the blue corolla with style 1.5-2 mm. long. In M. Aimalayense the trichomes on the stems, branches, and leaves are dense, subappressed, and of different lengths, the flowers always sessile, the fruits subsessile or short pedicellate, the thick, stout pedicels never exceeding 2 mm. in length, the bluish-white to pink corolla and the style 0.5~1 mm. long. The specimens cited above are annual, not perennial. 4. M. thomsonii (C. B. Clarke) Kazmi, comb. nov. Paracaryum thomsonii C. B. Clarke in Hook. f. Fl. Brit. India 4: 161. 1883; Brand, Pflanzenr. IV. 252 (Heft 78): 48. 1921. Type: Alpine western Tibet: Nubra, Baltal, Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (K-as Echinospermum no. 2). Robust perennial with roots and lower part of the stems woody, up to 50 (—100?) cm. tall. Stems usually many, erect, much branched, branches spreading, stems and branches densely covered with patent, white tri- chomes 0.5-1 mm. long, arising from minute inconspicuous tuberculate bases. Basal leaves many, forming rosettes, petiolate, (including petioles) 100-140 mm. long, 30-45 mm. broad, ovate to oblong-ovate, coriaceous, prominent, white, tuberculate bases, petioles with longer trichomes on the margins towards the base; lower cauline leaves more or less like the basal 1971] KAZMI, BORAGINACEAE 131 lobes 1.25 mm. long and broad, sometimes broader; faucal appendages subquadrate, emarginate. Anthers attached to the middle of the corolla tube. Nutlets 5-6 mm. long, 4-5 mm. broad, broadly ovate, winged at the margins, wings 1-1.5 mm. broad, plane at the margins to very slightly recurved, broadly dentate, teeth triangular, bearing glochidia at apex, dorsal middle area vertically appendiculate, appendages glochidiate. Style up to 1 mm. long. DIsTRIBUTION: Kashmir, Tibet. Kashmir: Lapak: near Kharbu, 3000 m., R. R. Stewart 21106 (cH); Olding near Indus river, 2800 m., R. R. Stewart 21032 (cu), 5. M. tibeticum (C. B. Clarke) Brand, Repert. Sp. Nov. 14: 156. 1915; Pflanzenr. IV. 252 (Heft 78): 66. 1921. Paracaryum tibeticum C. B. Clarke in Hook. f. Fl. Brit. India 4: 162. 1883. P. tibeticum var. schlagintweitii Brand, Pflanzenr. IV. 252 (Heft 78): 66. 1921, syn. nov. Type: Ladakh, 12000-14000 ped., Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n, (K-holotype, BM-, GH-isotypes, as Omphalodes no. 6). Annual, erect to ascending herb, to 35 cm. tall. Stems solitary or few, more or less dichotomously branched, branches long, suberect to spread- ing, covered densely on the lower parts with spreading trichomes to 1 mm. long, arising from minute tuberculate bases and on the upper part with short, antrorsely appressed trichomes to 0.5 mm. long, apparently not arising from tuberculate bases. Basal leaves oblong to oblong-ovate, grad- ually narrowed towards the base into a short petiole, to 6 cm. long and 5-7 mm. broad, covered on both surfaces with short white trichomes arising from prominent white tuberculate bases; lower cauline leaves oblong-lanceolate with shorter petioles; upper cauline leaves sessile, much shorter and narrower. Inflorescence terminal, lax, elongated in fruit. Pedi- cels erect, hairy, lacking or to 0.5 mm. long in flower, 2-3 mm. long in fruit, rarely those of the lowermost fruits elongated to 6 mm., reflexed. Calyx divided to the base, lobes in flower erect, 1-1.5 mm. long, spread- ing or reflexed in fruit, 2.5 mm. long. Corolla light to deep blue, cylindri- cal-campanulate, ca. 2 mm. long, limb subequal to the corolla tube, lobes roundish, spreading; faucal appendages broader than long. Anthers at- tached to the middle of corolla tube. Fruit orange-red. Nutlets ovoid, 3-4 mm. long, 2.5-3.5 mm. broad, winged at the margin, wings ca. 1 mm. broad, plane at the margins, distantly minutely dentate, teeth at apex not glochidiate, middle dorsal area tuberculate to vertically appendiculate, appendages glochidiate at their apices. Style 0.5—1 mm. long. Distripution: West Pakistan, Kashmir, west Tibet. West Pakistan: Gitcir AGENCY: Baltistan, Indus River Valley, 10 miles east of Skardu, 2600 m., Webster & Nasir 5857 (GH); Skardu, 2600 m., R. R. 132 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Stewart 20403 (GH); from Hushe to Sospor, Schlagintweit 6006 (c-type of var. schlagintweitit). K ir: LADAKH: Mulbekh, Rabila, Koelz 6155, 6129a (micH); Marsho, Koelz 6385 (MicH); Lamayuru, 3800 m., Osmaston 143 (K); Ladakh, Thomson $.n, (GH, BM, K-as Omphalodes no. 6). Brand (1. c., 1921) recognized the plants with much longer pedicels (Schlagintweit 1006) as var. schlagintweitii. The length of fruiting ped- icels is quite variable and the specimen cited above represents the varia- tion with the longest pedicel. This has also been mentioned by Clarke (1. c., 1883) in the original description of P. tibeticum as “lower pedicels in fruit occasionally 14 in.” 6. M. bungei (Boiss.) Rech. f. & H. Riedl in Kgie & Rechinger, Biol. Skr. 13(4): 209. 1963; Ried] in Rechinger, Fl. Iranica 48: 121. 1967. Echinospermum bungei Boiss. Fl. Orient. 4: 252. 1875. Paracaryum bornmiilleri Brand, Repert. Sp. Nov. 14: 149. 1915. Paracaryum bungei (Boiss.) Brand, Pflanzenr. IV. 252 (Heft 78): 47. 1921; M. Pop. in Fl. URSS 19: 605. 1953. Type: Persia borealis, prope Shahrud (28 Maio 1868), Bunge s.n. (G- holotype, GH-isotype). Icon.: Riedl, 1. c. fig. 167, 168. 1963, under M. tibeticum; M. Pop. lic. t. 29. fig. 3. 1953, under Paracaryum bungei. Annual, erect, ascending or sometimes prostrate herb, Stems solitary to few, to 25 cm. long, usually branched near the base, branches long; stem and branches covered on the lower part with spreading, on the upper part with antrorsely appressed to subappressed, white trichomes to 1 mm. long, arising from prominent tuberculate bases. Basal leaves 1-2, narrowly oblong to narrowly lanceolate, obtuse, short petiolate, 10-15 mm. long and 1.52 mm. broad; lower cauline leaves lanceolate to subspathulate, subsessile or very short petiolate, obtuse to acutish at apex, narrowed towards the base, 10-13(-18) mm. long, 2-3(-6) mm. broad, covered on both surfaces with short, thin, white trichomes arising from tuberculate bases; upper cauline leaves shorter and narrower. Inflorescence lax, much elongated in fruit. Pedicels hairy, lacking in flower, elongated to 6 mm. in fruit. Calyx divided to the base, lobes lanceolate, hairy, ca. 1.5 mm. long in flower, 2 area vertically 0.5 mm. long DISTRIBUTION: Iran, Afghanistan, West Pakistan, Pamir Alaj, Kizil 1971] KAZMI, BORAGINACEAE 133 West Pakistan: KALAtT Dist.: Basima, between Panjgur and Surab, 1350 m. Kazmi s.n. (PEs); Rechinger 28281 (w); between Surab and Kalat, 10 kms. from Kalat, Lamond 681 (£); Shalku, Jafri & Akbar 2936 (E). MAKRAN Dist.: between Panjgur and Surab, near Nag, 1300 m., Kazmi 1258 (GH, PES); Rech- inger 28260 (w); Panjgur, Irrigation Scheme, sandy river bed below Dams, 1000 m., Kazmi 12086 (pes); Lamond 604 (£). Quetta Dist.: between Padag and Nushki, 1200 m., Lamond 169 (gE); Rechinger 27335 (w). ZHos Dist.: Fort Sandeman to Wana, 26 km. from Fort Sandeman, between Zhob and Sambaza, 1500 m., Kazmi 1920a (pes); Lamond 1755 (E); Rechinger 29844a w). Reported from: Chagai, 60 km. E. Dalbundin, ad radices M. Rus Kuh, Rechinger 27296 (w). The characters, which differentiate Mattiastrum bungei from its very closely related species M. tibeticum, are the marginal wings of the nutlets curved along the side in M. bungei, those of M. tibeticum being flat. The margins of the wings in both species are dentate, but the teeth are blunt and do not bear glochidia at the apex in M. tibeticum, whereas in M. bungei the teeth are slightly produced and do bear glochidia at the apex. The plants of M. tibeticum are stouter and erect as compared to the weak stems, and branches of M. bungei. The former usually have many basal leaves, the latter have only one or two. Brand (1. c. 1921) and Riedl (1. c. 1963 & 1967) do not mention the length of the pedicels of M. bungei. In the isotype (Shahrud, Bunge (GH)) the pedicels of the lower fruits are as long as 15 mm., a condition I did not find in the specimens cited above, only a few of them having pedicels to 6 mm. in length. 23. Solenanthus Ledeb. Icon. Pl. Fl. Rossica f. 26. 1829; Fl. Alt. 1: 193. 1829. TYPE SPECIES: Solenanthus circinnatus Ledeb. Perennial or biennial, villous or tomentose herbs. Leaves alternate, In- florescence terminal or axillary, short thyrsoid, later usually elongated into loose panicles, ebracteate. Flowers pedicellate, pedicels short, later elongated. Calyx divided to the base, lobes hairy, scarcely enlarged in fruit. Corolla cylindrical to funnelform; faucal appendages present. Stamens much exserted from the corolla; anthers subglobose to ovoid. Nutlets ovoid to subglobose, appendiculate at the margins, on the dorsal face or all around, appendages glochidiate, coherent to free from the style. Species 10 to 14, distributed in southern Europe, central and southwest sia. KEY TO THE SPECIES fe Basal leaves truncate to cordate at the base, abruptly narrowed into the petioles; cauline leaves semiamplexicaul; anthers ca. 0.5 mm. Ong. ........ St 9, bet 6.00 aie Kins ws bd wy SONAL ORES Kee iG ee Ge ea De tee 1. S. circinnatus. a. Basal leaves neither truncate nor cordate at the base, gradually narrowed into the petioles; cauline leaves sessile; anthers ca. 1 mm. long. . . 2. S. stamineus. 134 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 1. S. circinnatus Ledeb. Fl. Alt. 1: 194. 1829, Icon. Pl. Fl. Rossica t. 26. 1829; DC. Prodr. 10: 164. 1846; Boiss. Fl. Orient. 4: 270. 1875; C. B. Clarke in Hook. f. Fl. Brit. India 4: 160. 1883; Aitch. Bot. Afgh. Delim. Comm. 88. 1888; Brand, Pflanzenr. IV. 252 (Heft 78): 157. 1921; M. Pop. in Fl. URSS 19: 642. 1953; Riedl in Rechinger, Fl. Iranica 48: 131. 1967. Solenanthus amplifolius Boiss. Diagn. Pl. Orient. 11: 126. 1849; Fl. Orient. : 270. 1879. S. rumicifolius Boiss. Dec. Pl. Nov. Orient. 2: 9. 1875. S. stamineus (Desf.) Wettst. var. cuneatifolius Bornm. Beih, Bot. Centralbl. 20b: 193. 1906. Type: In regionibus orientalibus ad fl. Irtysch. C. A. Meyer s.n. (LE). Icon.: Ledeb. Icon. Pl. Fl. Rossica t. 26. 1829; 'M.:Pop, l.'e. 229%) 1953; Riedl, 1. c. t. 46. fig. 1, 1967 Perennial, erect herb, 30-120 cm. tall. Stem simple, striate, fistulous, covered loosely to densely with soft, crisped white trichomes. Basal leaves, petiolate, (including petioles) 10-60 cm, long, (2—)5-10(—14) cm. broad, elliptic-oblong to oblong-lanceolate, obtuse to acutish, at the base usually cordate, sometimes truncate, loosely or densely pubescent on both surfaces, on the lower surface sometimes with some scattered trichomes arising from minute tuberculate bases, upper surfaces rarely glabrous; cauline leaves much reduced in size, usually sessile, cordate at the base, semi- amplexicaul, loosely or densely pubescent on both surfaces. Inflorescence congested, capitate in young plants, later thyrsoid or rarely loose panicu- late, ebracteate. Pedicels short. Calyx densely pubescent, divided to the base, lobes oblong-ovate to oblong-lanceolate, obtuse to subacute, 3-5 mm. long in flower, not much enlarged in fruit. Corolla blue, 4.5—-6 mm. long, cylindric-infundibuliform, lobes triangular, subacute, erect, ca. 1 mm. long; faucal appendages oblong, developed below the middle of the corolla tube. Anthers ovate, 0.5 mm. long; filaments at the base not at all to conspicu- ously dilated, ca. 5 mm. long, much exserted from the corolla. Nutlets ovoid, obtuse, 6-7 mm. long, dorsally irregularly appendiculate, append- ages glochidiate. DistriBuTion: Iraq, Iran, Afghanistan, West Pakistan, Kashmir; Cau- casus, Turkestan, Pamir Alaj, Siberia, Tien Shan, Dsungaro-Tarbagatai. West Pakistan: CHITRAL STATE: Rumbur, Bowes Lyon 663 (BM, £); Ziarat, Lowari Pass, 2600 m., Stainton 2339 (pm); Agram, Arkari Gol, west Tirichmir, 3150 m., Stainton 2634 (BM, E); Shigar Arirat, 2850 m., Bowes Lyon 758 (BM). Reported from: Chitral, Barum Gol, Shokor Shal, 3500 m., Wendelbo s.n. mir: Lapak: Dras, 3500 m., Ludlow & Sherriff 8319 (cu); Zanskar, Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (cH). Reported from: Pangee, Watt s.n.; Ladakh in Chilungtal, Meebold 4242. 2. S. stamineus (Desf.) Wettst, Denkschr. Akad. Wiss. Wien 50(2): 88. 1885; Brand, Pflanzenr. IV. 252 (Heft 78): 155. 1921; M. Pop. 1971] KAZMI, BORAGINACEAE 135 in Fl. URSS 19: 652. 1953; Riedl in Rechinger, Fl. Iranica 48: 133. 1967. Cynoglossum stamineum Desf. Ann. Mus. Hist. Nat. aa to 431. 1807. Mattia staminea Roem. & Schult. Syst. Veg. 4: 82 & 764. Moltkia libanotica Zucc. Abh. Acad. Wiss. Miinchen 3: ae ene Solenanthus tournefortii DC. Prodr. 10: 164. 1846. Solenanthus conglobatus DC. Ibid. 166 Solenanthus strictissimus Brand, Repert. Sp. Nov. 13: 546. 1915; Pflanzenr. IV. 252 (Heft 78): 156. 1921; Riedl in Rechinger, Fl. Iranica 48: 133. 1967, SYN, NOV. Type: Cappadocia, Tournefort s.n. (Pp). Icon.: Brand, |. c. fig. 217. 1921. Perennial, erect herb, to 70 cm. tall. Stem simple, striate, fistulous, covered loosely or densely with soft, patent or retrorsely subappressed white trichomes. Basal leaves petiolate, (including petioles) to 45 cm. long and 3.5(—6) cm. broad, lanceolate to oblong-lanceolate, acute to sub- acute, at the base narrowed into a long petiole, 3—5-nerved, nerves prom- inent below and covered on the upper surface loosely, on the lower surface densely with short, crisped, white trichomes, the nerves with dense, subap- pressed trichomes; middle cauline leaves sessile, shorter and narrower; upper cauline leaves shorter and broader. Inflorescence terminal or in the axils of the upper leaves, congested in short, cylindrical thyrses when young, later elongated, ebracteate. Pedicels lacking in flower, up to 2 mm long in fruit. Calyx divided to the base, lobes 2-4(—5) mm. long, ca. 1 mm. broad, oblong-lanceolate, obtuse, erect, pubescent. Corolla dark red to rarely yellowish, 5—-6(—7) mm. long, cylindrical, lobes 0.5 mm. long, 1 mm. broad at the base, triangular to broad-ovate, erect; faucal append- ages subquadrate, emarginate, developed at the middle of the corolla tube. Filaments up to 5 mm. long, slightly dilated at the base, inserted at the middle of the corolla tube; anthers 1 mm. long, ovate, much ex- serted from the corolla. Nutlets ca. 10 mm. long, ovoid, at the margins densely, and at the dorsal middle area sparsely appendiculate, appendages glochidiate at apex. DIstrIBUTION: Greece, Turkey, She Lebanon, Iraq, Iran, Caucasus, Afghanistan, West Pakistan, Pamir West Pakistan: CH1TRAL STaTE: Serer, 3300 m., Bowes Lyon 791 (BM, E); Agram, Arkari Gol, west of Tirichmir, 3330 m., Stainton 2627 (BM). HAZARA Dist.: Kagan ad Kamalban forests, Kazemi sn. (PES); Kagan Valley, Narang, 3. 6. 192 , Champion s.n. (cH). Gitcit AcENcy: Plants of Gilgit Expedition nS eg = 5980 Griffith, Kew Distrib.), Giles s.n. (E-type of S. strictissimus). Brand (1. c. 1915, 1921) distinguishes Solenanthus strictissimus from S. stamineus by its linear, basal leaves to 10 cm. long and by its corolla and calyx lobes 1 mm. shorter. The primary basal leaves in S. stamineus are also linear and short, they dry and fall away as soon as the secondary 136 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 basal leaves appear. The secondary basal leaves, which are found on most collections are much longer and broader than the primary leaves. Solenanthus strictissimus is based on the only collection of Giles from Gilgit Agency and no other collection of this species has been made since. It appears that one of the early leaves, which is shorter and narrower, left attached to the type collection, led Brand to recognize it as a distinct species. As there are no differences in the form and indument of the cauline leaves and the structure and measurements of the floral parts, I consider S. strictissimus to be a synonym of S. stamineus. | To be continued | 1971] CULLEN, THE GENUS NEILLIA 137 THE GENUS NEILLIA (ROSACEAE) IN MAINLAND ASIA AND IN CULTIVATION J. CULLEN MY ATTENTION WAS DRAWN to the genus Neillia by an attempt to identify three specimens growing at the University of Liverpool Botanic Gardens, Ness, Neston, Cheshire. The standard works for the identifica- tion of garden shrubs all seemed unsatisfactory as far as this genus was concerned. This led me to borrow herbarium material for comparison and study, and to write to various gardens, asking for specimens of all their Neillias for inclusion in a taxonomic survey of the genus. Comparison of the herbarium material with the relevant taxonomic literature (particularly Vidal’s revision of the genus, in Adansonia II. 3: 1963) revealed that many specimens had never been named, and led me to a reappraisal of the species, as represented by both herbarium and gar- den material. The genus Neillia is the type genus of the small tribe Neillieae of the subfamily Spiraeoideae of the Rosaceae. Its closest allies are Stephan- andra and Physocarpus, both of which contain species in cultivation. The distinctions between these genera are not great, and errors appear in most of the accounts. The most satisfactory key to their identification is that provided by Vidal (0. cit.), on which the following one is based. 1. Carpels 5-3 (-1); follicles dehiscent along both sutures. .... Physocarpus. 1. Carpels 1-2(-5); follicles dehiscent along the ventral suture only. 2. Style terminal; follicle 2-10-seeded; testa smooth and SUIT oo isa cc Defined as in this key, the genus consists of shrubs (one species suf- frutescent) of arching habit, with lobed, serrate leaves, and racemes or panicles of white or pink flowers. A flowering bush has a strong super- ficial resemblance to a Ribes, and, in gardens, Neillias tend to be grown in mixed shrubberies and semiwild areas. Vidal, treating the whole area of € genus, recognizes 13 species; I have seen material only from the main- land of Asia (the genus occurs also in Java and Sumatra), and in this area I recognize ten species, which correspond to twelve of those maintained by Vidal. Most of the characters used in the taxonomic treatment of the genus are straightforward, and have been described by Vidal. In addition, how- ever, I have found the mode of branching to be very useful in dividing the genus into two parts, and as the branching is unusual, and to my piwiction has not been described before, I give some explanatory notes ere. 138 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 MODE OF BRANCHING The following observations were made on mature bushes of Neillia thi- betica and N. sinensis var. ribesioides growing at Ness, supplemented by studies of herbarium material. Mature bushes of these two species consist of a large number of old, woody, brown-barked stems, leafless at the base, arising from ground level. These old stems, which may reach 15 feet (45 dm.) or more, are usually markedly zigzag, especially in the upper part, and their apices are apparently dead, as each ends in a short, dry extension without ob- vious buds. Alternate leaf scars are present on these stems, and in their axils are multiple, superposed buds, of which generally three or four are visible, the uppermost being some distance (often about 1 cm.) above the actual leaf scar. The plants are deciduous, and in winter completely leafless. As spring advances buds borne near the apices of the old shoots begin to develop; in all cases the uppermost bud of each axil develops first. These growths develop into the short shoots bearing the inflorescences, and quite early in the spring the young flower buds may be seen terminating the short, leafy shoots. By about the end of May these shoots reach their maximum extension and the flowers open. Each flowering shoot is markedly perulate at the base, with hard, persistent bud-scales; bears from three to seven leaves, and is terminated by the raceme. The leaves on these shoots of limited growth have only one bud in each axil. As the flowers mature, and the fruits develop, one of the lower buds from the axil which produced the flowering shoot begins to develop, as do buds in the lower parts of the woody stems. These buds develop into vegetative shoots of unlimited growth, which extend rapidly, and again, have multiple, superposed buds in their leaf axils. After the release of the seeds, the whole of the flowering shoot is often shed, leaving the vegetative shoot which developed from the same axil, and those from below, to persist through the winter. In Neillia sinensis var. ribesioides the leaves on the flowering shoots are smaller and more deeply lobed than those on the vegetative shoots, whereas in NV. thibetica there is little difference between leaves from the two types of shoot. Thus, in both these species there is a shoot dimorphism, with the earlier flowering shoots, which are of limited growth, being formed from the uppermost buds of each group and terminating in an inflorescence, with each leaf having one bud in its axil. The later vegetative shoots, are formed from the lower buds of a group, are of unlimited growth, with each leaf having multiple, superposed buds in its axil. As far as can be judged from herbarium material (which is usually collected ay the flowering stage) a similar situation is found in all the other species of the genus with the exception of Neillia thyrsiflora, N- serratisepala, and probably N. fallax. In these species the dimorphism be- tween vegetative and flowering shoots is not so marked, and the leaves 1971] CULLEN, THE GENUS NEILLIA 139 on the long flowering shoots bear multiple buds in their axils. The flow- ering shoots are not perulate at the base, and in general, are more leafy. The exact behavior of these species is in doubt, as I have not seen living plants of any of them, but they clearly form a distinct group, easily sep- arable from the other species. Multiple superposed buds occur in the one species of Stephanandra I have seen (S. tanakae Franch.) but are absent in at least two of the species of Physocarpus (P. opulifolius (L.) Maxim. and P. amurensis Maxim.). A NOTE ON TERMINOLOGY The flowers of Neillia are of the type generally known as perigynous, i.e., with sepals, petals, and stamens borne on the top of a tubular struc- ture. The nature of this structure has been much debated, and various names have been given to it. For instance, Rehder (Man. Cult. Trees & Shrubs, 1927) refers to it as “calyx tube,” whereas Vidal (op. cit.) calls it “hypanthium.” Both of these terms have morphological implica- tions which are not certainly valid in the case of V eillia, and I have there- fore used the neutral term “floral tube.” GROUPING OF THE SPECIES In a genus of this size there seems relatively little point in recognizing subgenera or sections. The ten species treated here, do however, fall into four more or less clearly defined groups: 1) Neillia thyrsiflora group. This consists of certainly two and pos- sibly three species. NV, thyrsiflora and N. serratisepala are closely related, and clearly belong together. The Javanese and Sumatran JV. fallax Blume, which I have not seen, probably also belongs here, as it has been frequent- ly confused with N. thyrsiflora, which also occurs on both Java and Su- matra. The group is characterized by its branching (see above), its strigose floral tube, and its white flowers. Neillia affinis group. Again consisting of two species, the poly- morphic NV. affinis and the closely allied and vicariant NV. rubiflora. The usually pink flowers have a globose-campanulate floral tube as broad as, or broader than long. The carpels contain three or more ovules. ah 3) Neillia lobata group. This is a rather variable group containing three species, NV. lobata, N. uekii, and N. gracilis. The one common character is the occurrence of only two ovules per carpel. The flowers are pink or whitish, and the floral tube is short and cylindric-campanulate. 4) Neillia sinensis group. Another group of three species, including the polymorphic NV. sinensis, N. thibetica, and N. sparsiflora. It is charac- terized by a long cylindrical floral tube, carpels with three or more ovules, and usually pink flowers. This group is taxonomically the most complex in the genus (except for NV. sparsiflora, which is a perfectly distinct species ) and the present account of it differs considerably from that of Vidal. 140 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 As originally described, Neillia sinensis and N. thibetica differed in a number of characters as shown below: NV. sinensis N. thibetica Floral tube long, glabrous outside shorter, pubescent outside Pedicels long short Inflorescence _ lax, few-flowered dense, usually many-flowered As more and more specimens were collected, these distinctions became blurred, and the identification of individual specimens became more dif- ficult. In attempting to deal with this situation more taxa were described, either at the specific or varietal level. These included NV. longiracemosa Hemsl. (1892), N. glandulocalyx Lévl. (1914-15), N. ribesioides Rehder (1913), N. sinensis var. caudata Rehder (1913), N. villosa W. W. Sm. (1917), and NV. kypomalaca Rehder (1932) (N. sinensis var. hypomalaca (Rehd.) Hand.-Mazz., 1933). No attempt was made to evaluate these taxa until Vidal’s paper of 1963 (Handel-Mazzetti’s treatment in Symbolae Sinicae being largely compila- tion). Vidal recognizes the following taxa: NV. glandulocalyx Lévl., N. sinensis Oliv. (with vars. sinensis, ribesioides (Rehd.) Vidal, and hypo- malaca (Rehd.) Hand.-Mazz.), N. villosa W. W. Sm., and N. thibetica Bur. & Franch. (NV. longiracemosa Hemsl.) with vars. thibetica, duclouxit Vidal, and caudata (Rehd.) Vidal. In discriminating between the four species recognized, Vidal lays great stress on the number of ovules per carpel: 4 or 5 in Neéillia glandulocalyx and N. sinensis, 5 to 8 in N. thi- betica, and 5 or 6 in N. villosa. This interpretation considerably alters the original distinctions, both N. sinensis and thibetica here including forms with hairy or glabrous floral tubes. Other characters used by Vidal in discriminating between the taxa include indumentum, leaf-shape, stipule and well-developed flowers; changes after fertilization render fruiting specimens unsuitable), Each specimen was scored for the following char- scatter-diagram form. This analysis shows that in pair of completely correlated characters by which one may divide the a) Floral tube glabrous or with . a few patent trich = e long; pedicels (2-)3F aaa Maen ae 10-17(-23) flowers per inflorescence. 1971] CULLEN, THE GENUS NEILLIA 141 10.« B oO dj 00 =) Bg oO 8 ie a 3 # ~ a ng 4 Oo Sm a = a a a] 6s A ar (e) N a A pe Ye a . © A a A ow" s 5 . ’ 7 ide Pedicel length 4 a 4 # o 3 ss 3 ] = oO 3 Oo of = ra @ a a a 2 a ag Oo a = = go 3 24 a " a= Da a g oo 7. ge OQ. 1 5 Pedicel length Figures 1 (above) and 2 (below). Variation in the Neillia sinensis/N. thibetica Stroup. A floral tube pubescent; [ floral tube with few patent trichome S; @ floral tube glabrous. (All measurements in mm.) On graphs for Perig. Zone” read “Floral tube”. 142 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 b) Floral tube finely pubescent, up to 6 mm. long; pedicels 0.6-2(-3) mm. long; (19—) 24-60 flowers per inflorescence. Two facts are immediately apparent: 1) ovule number is in no way correlated with any of the other characters (cf. Ficure 4); 2) the two taxa defined above correspond closely with the original conceptions of Neillia sinensis and N, thibetica. These correlations break down in the case of only a few specimens, indicating perhaps, occasional hybridization. As a result of these studies, I recognize only two species in this group, Neillia sinensis and N. thibetica. Neillia thibetica appears as a very uni- form taxon, showing little variation. Of the varieties recognized by Vidal, var. caudata goes with N. sinensis (as originally described by Rehder), having long, glabrous, long-pedicelled flowers in few-flowered inflorescences. Varietas duclouxit is a problem. Vidal describes it as having a glabrous floral tube (“. . . calycis tubo glabro’’), which would indicate its rela- tionship to N. sinensis; but the one specimen I have seen cited by him as var. duclouxii (Wang 23084) has a finely pubescent floral tube and differs in no essential way from typical NV. thibetica. As opposed to Neillia thibetica, N. sinensis is polymorphic, varying in flower size, leaf shape, and indumentum of all parts. The specimens can easily be separated into two groups: the sparsely hairy or glabrous (re- ferring to leaves, pedicels, and floral tubes) and the densely hairy. Such a division is somewhat subjective as there is no clear discontinuity in the variation, but individual specimens are usually easily assignable to one or the other group. What is remarkable is that variations in other charac- malaca (Rehd.) Hand.-Mazz., villosa (W. W. Sm.) Cullen, and rosea Cullen. Additional details are given in the taxonomic revision. TAXONOMY Neillia D. Don, Prod. FI. Nepal. 228. 1825. Suffrutescent herb or, more usually, arching shrubs up to 6 m., usually hairy in at least some part. Shoots usually dimorphic. Leaves alternate, bracteate, pedicellate, with a cylindric to campanulate floral tube (hypan- 1971] CULLEN, THE GENUS NEILLIA 143 A o- A A * 3 £ A ® a : o MPM): BBR Batt tig, “Thy @ 7% fe} 67 A N a A 3 - v e v + f se v : 3 - + v ® rol 2 + + . , . . ~ 4 8 No. of ovules ‘| | a | = 9 a8 @ c - @ 6 oa a o @ = S = 6s a a Sg | a . = a = 5s - Sky oom a a % oe e A A a» * 5 ae T n scnacmene + T T + 7 ¥ 60 30 No. of flowers ) V floral tube pubescent; A floral tube glabrous. Ficure 4 (below). Variation in the NW. sinensis/N. thibetica group. A floral tube pubescent; gg floral tube glabrous. (All measurements in mm.) On graphs tr Perigynous zone” read “floral tube’. thium or calyx tube of authors), glabrous, sparsely patent-pilose, pubes- cent or strigose outside, pilose within below the insertion of the stamens. Nectar secreted near the base of the floral tube. Sepals 5, pubescent on 144 JOURNAL OF THE ARNOLD ARBORETUM [voL, 52 the inner surface. Petals 5, white or pink, very shortly clawed, erose-ciliate on the margins, sometimes pubescent on the inner surface. Stamens (10—) 15-30, in two or three series, inflexed in bud. Carpels 1(—2- 5), free when more than one, ovules 2~10 per carpel. Fruit a follicle or group of follicles opening along the ventral suture, enclosed in the persistent, ac- crescent floral tube, which often develops an additional indumentum of long, spreading, capitate-glandular bristles after fertilization. Seeds sev- eral, shining, pale brown. Lectotype species: Neillia thyrsiflora D. Don. 1. Leaves on the flowering shoots with multiple, superposed buds in their axils; floral tube strigose; petals white. 2, Ovary velutinows; ovules 3<4. ........6.......>... 2. N. serratisepala. Z, iis glabrous or hairy at the apex only, rarely ete pubescent ovules co CR ee ne . N. thyrsiflora. 1 . Leaves on the flowering shoots with one bud in their sci Pa, tube glabrous or variously hairy but not strigose; petals usually p ‘| hei tube globular-campanulate at anthesis, as bad as or broader than ong. 4, ete 1-5, pubescent all over the surface, narrowing abruptly into Cpudl eke DRS ce A Uae rr . affinis. 4. cies 1, pilose only at apex, tapering gradually ja the style. iS ne ee ete WR cries pees 8) oe 947, 6 N. rubiflora. Floral tube cylindric or cylindric-campanulate, always he than broad. 5. Short shoots with an indumentum of < gee ale haits, (4 126 Be are We ee Tae 8. N. sparsiflora. Be age shoots without capitate-glandular hairs 6. Suffrutescent herb, up to 30 cm. hi gh, wiih creeping, rhizomatous shoots (ovules 2 Oe ee S: N. gracilis. WwW Ow Oo wm 4 o n B42 =a au {=} gs > mS 7 2 q: i} nel aa 4 © =] a my oO wn = = So ° c = re) ~ ia") oO a SS ga Be ee N ~) as rx) = ° c wm ‘ Branches and habeicines axis with an indumentum of sparse, finally deciduous, tuberculate- based, stellate hairs (ovules 2). 6. N. uekit. ~~ cy ty =] = oO n * =) Qu =a Eb oO wn ® =| a om iS) ral ete a “08, ist = tod ° c n °o 4. jars = 2 oe 3 a =a oO ys 8. Bark of old shoots puberalent; leaves lobed iS - 3 of the width of lama: ovules 7... ............. lobata. 8. Bark of old shoots as. leaves less Prise pases ovules 3 or m 9 ; sleds tube wabeons or with sparse patent hairs near e base; flowers 10-17(-23) per inflorescence. .....- Norte) a rane I aaa rr 9. N. sinensis. Floral tube finely pubescent all over me surface; flow- ers (19-)24-60 per inflorescence. 0. N. thibetica. 2 1. Neillia thyrsiflora D. Don, Prodr. Fl. Nepal. 228. 1825; Vidal, Adansonia IT. 3: 152. pl. 1, figures 36.1963. Frqure 5: 17, 18 N. virgata Wall., nomen nudum, 3 Ovary glabrous or sparsel y pilose at the apex and al th oes te anicles usually much branched. ‘i ie etapa nana : ape OR ait ied ieee ts 3) oa. thyrsifiora. 1971] CULLEN, THE GENUS NEILLIA 145 Ovary pubescent over most of the surface; panicle usually little branched. ...... Pe tedliye ps Ha pur eu Fe street kro caw «bas wrken demees OS var. tunkinensis. Sikkim. Regio temp., 2000-2600 m., Hooker (A, £); Gassing to Ratong river, 2 Oct. 1882, Anderson 505 (£); Sureil, 1800 m., 29 Aug. 1912, Cave (A); Mung- poo, 1300 m., 28 July 1909, Cooper 556 (A). Bhutan. Chukha Tunpu, 1500 m., 4 July 1914, Cooper 1229 (E); Tongsa, 2600 m., 11 Sept. 1914, Cooper 2069 (E); Tilageng, Punabha, 2100 m., 18 Aug. 1914, Cooper 2814 (£); sine loc., Griffith (E). Nepal. Sine loc., E.J.C. (ce); Gurke Bandhara to Charikot, 2000 m.,10 Oct. 1960, Banerji 1210 (A), Burma. Laktang, 1500 m., 21 June 1919, Kingdon-Ward 3241 (£). Tibet. Raiotdong to Santok, 800-2000 m., 27 June 1903, Younghusband A). China, YUNNAN: Szemao forest, 1500 m., Henry 12275 (A, £); S. of Red river, Henry 13653 (A, ©); sine loc., 1939-40, Wang & Liu 81820 (a, E); Kiukiang valley (Taron), Monting, 1350 m., 10 Sept. 1939, Yii 20199 (a, £); ibid., Chien- gen, 1650 m., Vii 19409 (£); above Kan San Chai, W. of the Mekong, en route from Pingpo to Yungchang and Tengyueh, Salween watershed, Oct. 1922, Rock 7056 (A); Ping Pien Hsien, 1300 m., 7 July 1934, Tsai 60717 (A); ibid., 1400 m., 15 June 1934, Tsai 60207 (a); ibid., 7 June 1934, Tsai 62383 (A); Lu-se, 1750 m., 10 Feb. 1934, Tsai 56895 (A); ibid., 1750 m., 1 Mar. 1934, Tsai 56350 (a); Dzung-duei, Cham-pu-lung, 2000 m., Oct. 1935, Wang 67076 (A) and 66940 (a); Che-tse-lo, 3200 m., Tsai 58432 (aA). Kwancsr: On Tak, 21 Oct. 1935, Ko 55787 (A). In addition to the areas mentioned above, Vidal (op. cit., 153) records this taxon from Sumatra. Var. tunkinensis (Vidal) Vidal, Adansonia IT. 3: 153. 1963. N. tunkinensis Vidal, Not. Syst. Paris. 13(4): 202. 1948. ? N. tunkinensis var. bibracteolata Vidal, loc. cit. Burma. N. Triangle (Hkinkum), 1300 m., 10 July 1953, Kingdon-Ward 21135 (a). China, YUNNAN: sine loc., Wang & Liu 83503 (A, E); sine loc., Forrest 29897 (=); Mengtze, 1800 m., Henry 9419 (a, £); ibid., S. mts., 1500 m., Henry 9419A (E); Mienning, Poshan, 2700 m., 9 Oct. 1938, Vii 17974 (a, £); Salwin- Kiukiang divide, Newahlung, 2200 m., 8 July 1938, Vii 19228 (a, k). Vidal records this taxon from Assam, North Vietnam, Java, and Su- 146 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 matra. It appears to be intermediate between Neillia thyrsiflora var. thyrsiflora and N. serratisepala, the only distinct character separating them being the number of ovules (8-10 in both varieties of V. thyrsiflora, 3-4 in N. serratisepala). 1 have followed Vidal in treating them as separate species, but there are grounds for considering N. serratisepala as merely another variety of NV. thyrsiflora. 2. Neillia serratisepala Li, Jour. Arnold Arb. 25: 300. 1944; Vidal, Adansonia IT. 3: 151. 1963. China. YUNNAN: Shang Pa, 2000 m., 31 Oct. 1934, Tsai 59158 (a, holotype). This species is known only from the type collection; for further observa- tions see under Neillia thyrsiflora. 3. Neillia affinis Hemsl. ney Linn. Soc. 29: 304. 1892; Vidal, Adan- sonia IT. 3: 155. 196 A polymorphic species, aie into four varieties: 1. Carpels 3- : free but contiguous, forming a globose ovary. .. Var. polygyna. 1. Carpels 1 2; Sepals as long as or longer than the floral tube, 4.2~-6 mm., much exceed- ta a rr ee et? Vat: longise pala, 2. Sepals shorter than the floral tube, up to 4 mm., not or scarcely ex- ceeding petals. 3. Raceme elongate, the flowers 10 or more, spaced along the ers Je oe EOS Be) as scale, "olan re a rrr afinis. 3. Raceme contracted, the 5-10 flowers fascicled at the end i td axis. Pree Gee ee aes kd IE oye, Var. paucifiora. Var. affinis Vidal, Adansonia II. 3: 156,156. plod, ‘fig: 11. Aisin uRE 5: 15. (A, a Kengma, Chuichay ko, 2500 m., 10 Aug. 1938, Vii 17207 (A, E); Fuchuan Range, W. of Mekong-Salwin divide and W. of Wei Hsi, 3800 m., 4-5’, June 1932, Rock 22729 (a, £); ibid., July- -Sept. 1932, Rock 23275 (A, E); inter fluvios Landsang-djiang (Mekong) et Lu- -djiang (Salween ) ca, 28°, in valle a jugo Si-la, vicum Tsekuo descendente, 2300-3050 m., 15 June 1916, Handel-Mazzetti 8884 E); Tengyueh, Howell 200 (E); sine oc. Forrest 30023 (£); Yi Liang Hsien, 1400 m., 27 May 1932, Tsai 52055 veh gener 2500 m., 31 or (A, E, oe Wu-shan, 2600 m., Jul i i , E, is ; .. July 1908, Wilson 3559 (a); Hsiao hsiang-ling, 2000 m., 21 May 1922, Smith 1865 (a); sine loc., 2600 m., July ; Nanchuan Hsien, 1928, Fang 935 (a). SIKANG: Chi-na- tung, Tsa-wa-tung, 3000 m., Aug. 1 5, Wang 65196 (a). N. Burma. Moku-ji pass, 3300 m., 28 July 1920, R, pba 1757 (£), Var. polygyna Cardot ex Vidal, Adansonia IT. 3: 156. pl. 1, fig. 12. 1963. 1971] CULLEN, THE GENUS NEILLIA 147 China, YUNNAN: Wei-hsi-hsien, Yeh-chih, 3200 m., mountain slope, Aug. 1935, Wang 68154 (a); on the Kari pass, lat. 28°N , 4000 m., June 1917, Forrest 14002 (A, E, wsy). Also recorded from eastern Tibet by Vidal. Var. pauciflora (Rehd.) Vidal, Adansonia II. 3: 156. pl. 1, fig. 3. 1963. N. pauciflora Rehd. Pl. Wils. 1: 437. 1913. China. YUNNAN: Mengtze, N. mt., 2300 m., Henry 10231A (A, holotype); hills to the E. of Tengyueh, 25°N, 2000 m., May 1912, Forrest 7854 (a, ©); Shweli- Salween divide, 25° 6’ N, 98° 50’ E, 2600 m., Forrest 29567 (£). SzECHUAN: Mt. Omei, 2300 m., 27 July 1939, Sun & Chang 1234 (A); ibid., Yen-wang-po, 2400 m., 18 Aug. 1940, Tu 272 (a). Var. longisepala Cullen, var. nov. A var. affinis sepalis majoribus, 4.2—-6 mm. longis, conspicuis differt. China. YUNNAN: Shweli-Salween divide, 25° 20’ N, 3000 m., shrub of 4-5’, fls. creamy yellow, in scrub and thickets in side valleys, Aug. 1918, Forrest 17605 (£, holotype; A, wsy, isotypes). SzECHUAN: Ta-hsiang-ling, 2700 m., 28 May 1922, Smith 2043 (a); Mt. Omi, June 1904, Wilson 4886 (A, NA). SIKANG: Tsarong, on Doker-la, Mekong-Salwin divide, 28° 20’ N, Forrest 14342 (a, E, wsy); Doker-la, 3600-4000 m., July 1913, Kingdon-Ward (£). Burma. Hpimaw village, road to Feng-shiu-ling, 2000-2300 m., forests, 22 May 1914, Kingdon- Ward 1584 (kr). As is shown by the number of varieties, Neillia affinis is a rather poly- morphic species. The variations, however, do not seem to be separable geographically. Var. longisepala is very distinct in its long conspicuous sepals; it may also differ from the other varieties in having stipules which are more toothed. The latter character is one which can only be deter- mined subjectively, and it is usually of no value in determining speci- mens in fruit, where the stipules have usually been lost. The inflorescence in NV. affinis is often a panicle, sometimes richly branched; this has caused some confusion with NV. thyrsiflora, a perfectly distinct species. 4. Neillia rubiflora D. Don, Prodr. Fl. Nepal. 228. 1825; Vidal, Adan- sonia IT. 3: 155. pl. 1, fig. 9, 10. 1963. FicurE 5: 16. India. E, Himalaya, Sandakpha, 4000 m., 21 July 1913, Rohmoo Lepcha 766 (E); Darjeeling district, 2300 m., May 1908, Smith (a). E. Pakistan. Bengal, Chittagong division, nr. Soylu, 3000 m., Cowan 79 (£). Sikkim. Reg. temp., 2600-3000 m., Hooker (a, E); ibid., Schneider (a); Zemu and Lonakh valleys, Zemu Chu, 3000 m., 6 Sept. 1947, Cave 97/47 (£); Sureil, 7 Sept. 1921, Wilson (A). Bhutan. Tobrang (Trashi Yangse Chu), 2300 m., 6 July 1948, Ludlow, Sherri & Hicks 20495 (e); Lao (Trashi Yangse Chu), 3000 m., 5 July 1949, Ludlow, Sherrif & Hicks 20475 (e); Shabjethang, Bumthang Chu, 3000-3200 m., 1 July 1949, Ludlow, Sherrifi & Hicks 19288 (£); Gyasa Dzong, Mo Chu, Peep 12 June 1949, Ludlow, Sherriff & Hicks 16520 (gE). Nepal. Sine loc., J.C. (E) 148 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 pea aay. sp ap PH CN = 8 | 1 11 Figure 5. Leaves of Neillia species. 1); 2. N. sinensis va Forrest 30475); 4. N, caudata (Cavalerie 2461): r. ¥ 1. N. sinensis var. sinensis Mote: r. sinensis (Maire s.n.); 3. N. sinensis var. hypoma i 5. N. sinensis vat. N. sinensis var. villosa (Maire 2330); 7. N: ); 12. N. gracilis (Maire s.n, ); 13. N. uekii (Russell 90); 14. Ai lobata (Schneider 3558); 15, N. affinis var. affinis (Rock 22729); 16. N. ru flora (Ludlow et al. 19288); 17. N. thyrsiflora (Vii 20199); 18. N. thyrsiflora (Henry 12275) ~ > 1971] CULLEN, THE GENUS NEILLIA 149 Neillia rubiflora is closely related to NV. a finis. It is also rather similar to NV. thyrsiflora, which grows in the same area and with which it has been confused. It differs from N. thyrsiflora in the lack of multiple buds on the flowering shoots, in its less-branched (usually racemose) inflorescence, its adpressed-sericeous floral tube, and its entire stipules. 5. Neillia gracilis Franch. Pl. Delav. 1: 202. 1889; Vidal, Adansonia II. 3: 157. pl. 2, fig. 2. 1963. FicureE 5: 12. Slopes of Likiang snow range, 4000 m., June 1922, Rock 4422 (A, ©); eastern flank of the Lichiang Range, lat. 27° 25’ N, 3000-3300 m., June 1910, Forrest 6033 (£); eastern flank of the Lichiang Range, lat. 27° 12’ N, 3000 m., May 1906, Forrest 2244 (A, £); at base of range to NW. of Ho Shui, 3300-4000 m., June 1906, Forrest 1871 (£); paturages des mont derriére Tong Tchouan, 2600 m., Maire 1100 (£); ibid., 1914, Maire 559 (E); 27° 30’ N, 100° 45’ E, 2800 m., 25 May 1921, Kingdon-Ward 3986 (E); Likiang snow range, 28 June 1939, Ching 30281 (a); Likiang-hsien, 2800 m., July 1935, Wang 70883 (a); Chungtien valley, 3500 m., 9 July 1939, Feng 1936 (A); in cacuminis montis Tche-tchang prope Tong-tchan, 3 June 1882, Delavay 249 (A, isolectotype); mountains S. of Lichiang nr. Hochin and Chinho, 25-28 May 1922, Rock 4079 (a); Lichiang range, 31 July 1914, Schneider 2071 (A); sine loc., 1937, Yii 15282 (aA). SzE- CHUAN: In declivibus graminosis regionis temperatae jugi Da6érlbi medii inter Yenyuen et Yungning, supra vic. Hungga, 3500 m., 13 June 1914, Handel- Mazzetti 2960 (£); inter Hunka et Woholo, 3000 m., 13 June 1914, Schneider 1506 (A, E). This species, Neillia gracilis, is the most distinct species of the genus, being suffrutescent and rhizomatous. In inflorescence and flower characters it resembles NV. affinis var. pauci flora but differs in having only two ovules per carpel, a character it shares with NV. lobata and N. uekii. Forrest 1871 is a very robust variant, with the inflorescence forming a small panicle. In other respects this collection matches the rest of the material; it may perhaps represent a distinct variety. 6. N. uekii Nakai, Bot. Mag. Tokyo 26: 3. 1912; Vidal, Adansonia II. 3: 158. 1963. FicureE 5: 13. N. millsii Dunn, Kew Bull. 1912: 108. 1912. Korea. Prov. N. Heian, around Matan (Unsan), 28 June 1917, Wilson 8714 (A); N. Heian, Kakai, 5 Aug. 1918, Wilson 10587 (A); Prov. Pyengan, Yengben, 30 June 1937, Mrs. R. K. Smith (A); Prov. Pyengan, in dumosis Pyekdan, 12 June 1914, Nakai 1774 (a); Kangkai, 6 Oct. 1909, Mills 107 (x, holotype; A, isotype of N. millsii); in petrosis Pyongyang, June 1901, Faurie 85 (a); in pe- trosis vallium Coreae mediae, July 1906, Faurie 316 (A, E). Neillia uekii is the only species of the genus known from Korea, where it is endemic. It is similar to V . lobata and N. gracilis in having only two ovules per carpel, but differs from all the species of the genus in having a Sparse indumentum of tuberculate-based, stellate hairs on the branches. in the region of the inflorescence. These hairs are ultimately deciduous, 150 JOURNAL OF THE ARNOLD ARBORETUM [voL, 52 but the tuberculate bases remain attached to the fruiting axis and are easily recognized. According to Vidal (loc. cit.), the name N. uekii was published in January 1912, whereas NV. millsii was published in March of the same year. 7. Neillia lobata (Rehd.) Vidal, Adansonia II. 3: 157. pl. 2, fig. 5. 1963. IGURE 5: 14, N, longeracemosa Hemsl. var. lobata Rehd., Jour. Arnold Arb. 1: 257. 1920. hina. SZECHUAN: Inter Oti et Yungsin-hsien prope Quentin, ca. 2900 m., 4 June 1914, Schneider 3558 (a, holotype; £, isotype); inter Tu-yung-pu et Yen- yuan-hsien, ca. 3000 m., 12 May 1914, Schneider 4151 (A); in regionis temperatis inter oppidum Yenyuen et flumen Yalong, 27° 31’, fruticetis circa vicum Tsin- tienpu, ca. 2750 m., 12 May 1914, Handel-Mazzetti 2218 (E). Neillia lobata is allied to N. uekii, but differs in possessing more deeply lobed leaves, finely tomentose bark on the old wood, and in lacking the characteristic tuberculate-based stellate hairs of N. uekii. The Handel- Mazzetti specimen cited above is a very poor one, and some doubt must remain as to whether it actually is V. obata. 8. Neillia sparsiflora Rehd., Jour. Arnold Arb. 1: 237. 1924; Vidal, Adansonia II. 3: 157. pl. 2, fig. 1. 1963. Ficure 5: 11. China. YuNNAN: Circa Pe-yen-tsin, Kou-ty, 1917, Ten 462 (a, holotype; £, isotype) ; ibid., 1918, Ten 531 (A, E). A very distinct species, Neillia Sparsiflora is characterized by the pres- ence of long capitate-glandular bristles on the short shoots and petioles. It is closely related to NV. sinensis, and is known from only three specl- mens: the two cited above, and a further one cited by Vidal (Yunnan, Djo Kou La prés Pin Tchouan, vii 1907 , Ducloux 5348, P). 9. Neillia sinensis Oliv. in Hook. Ic. Pl. 16: ¢. 1540. 1886. As mentioned above (p. 142), this species and Neillia thibetica form an intergrading complex of forms difficult to separate into distinct taxa. However, six varieties are recognizable in N. sinensis: _ . Floral tube glabrous; leaves glabrous or pilose only on the veins beneath; short shoots glabrous or sparsely pilose. caudate (pedicels 2-3 mm., glabrous; floral tube 1.6—2.5 times longer than broad ar. caudata. oblong, oblong-ovate, or lanceolate, not cordate, entire or with 1 or 2 small teeth; leaves relatively smaller, often lobed, usually not caudate, 3. Floral tube less than 6.5 mm, tube 2-3 times longer than bro 3. Floral tube more than i] 2 me OC Mo) ¢ ® — oO vw long (pedicels 2-5 mm., glabrous; floral ad; flowers 15-19 in each inflorescence). toate ee ee var. ribesioides. 6.5 mm. long (pedicels 3.5—7 mm., glabrous; 1971] CULLEN, THE GENUS NEILLIA 151 floral tube (2—)2.5—4 times longer than broad; flowers 10-17 in each IRON i iia Ph cr een ee var. sinensis, 1. Floral tube usually with sparse patent hairs near the base; leaves pubescent to pilose on the surface beneath, usually densely so; short shoots pubes- cent to pilose. 4. Floral tube 5.5—6 mm.; pedicels up to2mm............. var. rosea, 4. Floral tube 6.5 mm. or more; pedicels exceeding 2 mm. 5. Stipules ovate-cordate; pedicels less than 3 mm. ..... var. villosa, 5. Stipules oblong, not cordate; pedicels exceeding 3mm. .......... var. hypomalaca. Var. sinensis Vidal, Adansonia II. 3: 160. pl. 2, figs. 7, 8. 1963. FicureE 5: 1, 2. China, SHeNnsi: Langchow, Kuan Shan, 2000 m., 3 July 1922, Hers 2343 (a); Tai-pe-shan, 1910, Purdom 1 (a, ©); ibid., Purdom 467 (a); Tu-kia-po, 4 June 1897, Giraldi (A). Kwancst: Hang-on-yuen, woods, 10 June 1936, Chung 81758 ( PEH: W. Hupeh, May & Sept. 1907, Wilson 86 (a, ©); Hsing-shan- hsien, 1300-2000 m., June & Sept. 1907, Wilson 189 (A, £); sine loc. Henr J554A (A); sine loc. Henry 5554B (); N. Tan-schian, Mar. 1912, Silvestri 4589 (a); Paokang, 10-20’, Wilson 701 (a, ©). SzECHUAN: Sine loc. Henry 5695 (a). Yunnan: Wei-hsi plain, McLaren D27 (a, £); Wei-hsi-hsien, Yeh-chih, 2400 m., July 1935, Wang 67945 (a). Hunan: Yiinschan prope urbem Wukang, 600-1300 m., 6 June 1918, Handel-Mazzetti 765 (a); ibid., Handel-Mazzetti 12010 (cr). Kansu: About Satanee and near Di-er-kan, Farrer & Purdom 54 (£); ad fl. Tchi- luco, 18 June 1885, Potanin (A). Neillia sinensis var. sinensis is easily recognized by its long, narrow, glabrous floral tube, its few-flowered inflorescences and its long, glabrous or sparsely hairy pedicels. An extreme form is represented by Wilson 189, in which all parts are glabrous. Var. caudata Rehd., Pl. Wils. 1: 436. 1913. FicureE 5: 4, 5. N. glandulocalyx Lévl., Fl. Kouy Tcheou 348. 1914-15; Vidal, Adansonia II. 3: 160. pl. 2, fig. 6. 1963. N. thibetica var. caudata (Rehd.) Vidal, Adansonia 3: 164. pl. 3, figs. 5, 6. 1963. China. YUNNAN: Mengtze, E. mt., 2000 m., Henry 9669 (a, holotype); Yong- shan-hsien, 2100 m., 4 June 1932, Tsai 50992 (a); Chen-hsiung-hsien, 1850 m., 3 July 1932, Tsai 52706 (a); sine loc., Tsai 51196 bis (A). SzecHUAN: Nanchuan- hsien, 2100-2500 m., 27 May 1928, Fang 1057 (A, E). KwertcHow: Route de Pin Fa 4 Tou Yun, 25 July 1905, Cavalerie & Fortunat 2461 (£, holotype and isotype of N. glandulocalyx; a, isotype); Top-kying, Ten-shan, Sunyi, 400 m., 4 June 1930, Tsiang 5283 (a); Tuyun, Ma-tsoon-ling, 500 m., 11 July 1930, Tsiang 5744 (A); She-won-san, Haufeng, 4 July 1936, Teng 90524 (a). Kwancst: Yang wu, Ling-chuan district, 21-30 July 1937, Tsiang 27882 (a); Kon-Tung, Min-shan, N. Luchen, 1000 m., 18 June 1928, Ching 6107 (A); Loh-hoh-tsien, Lin-yuin-hsien, Lao-shan slope, 1450 m., 20 May 1933, Steward & Cheo 450 (a); Sze-hoh-tou, Lin-yuin-hsien, 12 Apr. 1933, Steward & Cheo 172 (A); Ling- wan district, Lau 28719 (a). 152 JOURNAL OF THE ARNOLD ARBORETUM [voL, 52 As explained above I prefer to treat var. caudata under Neillia sinensis (as it was originally described by Rehder), rather than under NV. thibetica, where Vidal places it. It is a variable taxon, particularly in the lobing of the leaves, which may vary from quite deeply lobed, as in the type, to completely unlobed, as in some leaves of Fang 1057. Even on the same shoot, considerable variation in the degree of lobing may be found. The number of flowers per raceme is also variable. The type of Neillia glandulocalyx (Cavalerie & Fortunat 2461) is a very poor fruiting specimen; however, in the Edinburgh herbarium there is a sheet labelled: “Flora of Yunnan, Yunnan-sen district [all this printed] Cavalerie 2461.” This is a much better specimen, and appears to be a mislabelled duplicate of Cavalerie & Fortunat 2461 (Mr. A. Lauener in- forms me that other, similar mislabellings are found in the Léveillé her- barium). On the basis of this specimen, which has both flowers and fruits, I am convinced that Neillia glandulocalyx and N. sinensis var. caudata belong to the same taxon; the distinctions between them noted by Vidal (number of ovules, leaf shape) are variable and are not, in my opinion, reliable for the separation of distinct species. Var. ribesioides (Rehd.) Vidal, Adansonia II. 3: 161. 1963. FicureE 5: 8. N. ribesioides Rehd. Pl. Wils. 1: 435. 1913. China. YUNNAN: Chao-tung-hsien, 2000 m., 19 May 1932, Tsai 50929 (A). SZECHUAN: Pan-lan-shan, 2300-3000 m., June and Sept. 1908, Wilson 2382 (A, holotype; see below); Tchen-keou-tin, Farges 494 (g); Cheng-kou-hsien, 3 May 1932, Fang 10315 (a). The type sheet of this taxon poses a problem; I have seen three sheets labelled Wilson 2382, two at the Arnold Arboretum and one at Edinburgh. The holotype sheet (a) has two shoots on it. One of these, which has flowers with glabrous floral tubes, is marked “typical form,” thus agreeing with a note with Rehder’s original description: “Our specimen contains 2 branches, one with a glabrous calyx tube, which I consider the typical form...” The other shoot is, in general, rather hairier, the flowers have finely pubescent floral tubes, and there are more flowers in the in- florescence. All the specimens on the other two isotype sheets (A and E) match this second, hairier shoot on the holotype sheet. I regard this hairier shoot as being Neillia thibetica, and it is interesting to note that the pre- ceding number, Wilson 2381, is clearly N. thibetica; this matches the hairy shoots on Wilson 2382 very well. Is it possible that these specimens have become mixed at some Stage, and that most of the plants labelled 2382 are really part of 2381? The other specimens cited above match the typical Wilson 2382 very well and are characterized by their short, glabrous floral tubes, glabrous pedicels, and small, rather deeply cut leaves, somewhat similar to those of the common British hawthorn (C7ré- taegus monogyna Jacq.). 1971] CULLEN, THE GENUS NEILLIA 153 Var. villosa (W. W. Sm.) Cullen, comb. et stat. nov. FIcurRE 5: 6. N. villosa W. W. Sm. Notes Bot. Gard. Edinburgh 10: 53. 1917; Vidal, Adan- sonia II. 3: 161. pl. 2, figs. 9-11. 1963. China. YUNNAN: Vicinity of Yunnan-sen, Maire 2330 (gE, holotype and sev- eral isotypes). This taxon is essentially a densely hairy analogue of Neillia sinensis var. caudata; it has similar stipules and leaves and identical flowers. The indumentum of the leaves and short shoots is very dense, as it is also in the variety which follows. Var. rosea Cullen, var. nov. A Neillia sinensis var. sinensis foliis subtus dense crispule pilosis, tubis floralibus brevissimis (5.6 mm.) differt. China. YUNNAN: Vicinity of Yunnan-sen, Maire 1120 (£, holotype and sev- eral isotypes); Kin-tchouy-chan, Lou-pou, coteaux, May—June 1922, Maire (ek). As Neillia sinensis var. villosa is the hairy analogue of var. caudata and var. hypomalaca that of var. sinensis, so var. rosea is that of var. ribe- sioides. It is characterized by its small, lobed leaves, very densely crisply pilose beneath, and its short, glabrous floral tubes. Var. hypomalaca (Rehd.) Hand.-Mazz. Symb. Sin. 7: 449. 1933; Vidal, Adansonia IT. 3: 161. 1963. Figure 5: 3, N. hypomalaca Rehd., Jour. Arnold Arb. 13: 337. 1932. China. YUNNAN: Litiping range, Mekong-Yangtze divide, east of Wei-hsi, June 1923, Rock 9171 (a, holotype; E, isotype); Tong-tchouan, collines, 2990 m., Maire 429 (a, paratype); Du-long, halliers des mont, 2990 m., Maire 529 (aA, paratype) ; sine loc., 2900 m., Maire 41 (a, paratype); Kangpu, Wei hsien, 1932 m., ravine, July 1935, Wang 64159 (a); NW. Likiang, Lu-tien, 3 June 1939, Ching 20483 (A); Pan Kiao Ho prés Se’tsung, 17 May 1904, J. B. Lo (Ducloux 633—E); Liu-ti-ping, 27° 12’ N, 3300 m., June 1917, Forrest 13883 (A, E, WSY); Wei-hsi area, Forrest 30475 (E); Mo-tsou, 600 m., June 1906, Maire 387 (E); Tsekou, May 1912, Mombeig 11 (£); San-o-chou, 2600 m., June 1913, Maire 1108 (); Kin-tchou-chan, 2900 m., May 1913, Maire (); Ma-long, 2800 m., May 1913, Maire 310A). This variety is the hairy analogue of var. sinensis. 10. Neillia thibetica Bur. & Franch., Jour. Bot. Morot 5: 45. 1891; Vidal, Adansonia II. 3: 163. pl. 3, figs. 1, 2. 1963, excl. vars. FicureE 5: 9, 10. N. longiracemosa Hemsl. Jour. Linn. Soc. 29: 304. 1892. China. Szecuuan: neighborhood of Tachienlu, 3000-4500 m., Pratt 730 (K, Syntype of N. longiracemosa); ibid., July and Sept. 1908, Wilson 974 (A, E); ibid., 1924, Cunningham 529 (e); ibid., 2600 m., Oct. 1910, Wilson 4220a (A); Pao-hsung-hsien, 2400 m., 20 June 1936, Chu 2827 (kg); ibid., 16 Aug. 1936, 154 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 Chu 3611 (£); Mupin, June and Sept. 1908, Wilson 916 (a, ©); ibid., 1600-2600 m., Oct. 1910, Wilson 4220 (a); Ta-hsiang-ling, Ching-chi-hsien, June 1907, Wilson 2381 (A, ©); ibid., Ta Kwan, 2400 m., 7 Aug. 1939, Chiao 1633 (a); Ma- pien-hsien, 29 May 1931, Wang 23084 (A); sine loc., 1300-3000 m., Wilson 3558 (A); Opien Hsien, 1500 m., 14 May 1932, Vii 740 (A); Ta Hsian Ling, 2000- 2500 m., 28 May 1922, Smith 2078 (a). As explained above, I have taken a different view of Neillia thibetica than Vidal. Of the specimens cited above, the most divergent is Chu 2827, which approaches N. sinensis in its long, narrow floral tube and rather long pedicels. However, in other features it is characteristic of NV. thibetica, and it does not seem useful to treat it as a distinct variety. The following specimens, because of their condition or the unsuitable stage at which they were collected, cannot be accurately identified. Pos- sible names are suggested for some of them in parentheses at the end of the citations. India. Khasia, Maw (a, probably N. rubiflora); Manipur, Ching Sow, 2600 m., Apr. 1882, Watt 6532 (rE); E. Bengal, Hb. Griffith 2115 (a, probably N. thyrsiflora var. thyrsiflora). China. YUNNAN: Ping Pien Hsien, 1400 m., 17 Apr. 1934, Tsai 60254 (A, perhaps N. thyrsiflora); ibid., 27 June 1934, Tsai 60483 (a); Wei-hsi-hsien, 2600 m., 16 Sept. 1934, Tsai 57936 (A); between the Yangtze and Mekong, 2850 m., 29 Aug. 1915, Handel-Mazzetti 7867 (r, N. sinensis var.); Atuntze, Doker-la, 3100 m., 4 Nov. 1937, Ya 7855 (z); hills to the E. of Tengyueh, 2000 m., May 1912, Forrest 7584 (E, N. affinis var.), SzecHuAN: Dailangshan, E. of Ningyueh, 2350 m., 24 Apr. 1914, Handel-Mazzetti 1644 (£); sine loc., Henry 5585 (x); Tien-chuan-hsien, 2300 m., 18 May 1936, Chu (£). SHENSI: Mt. Tou-pei-san, 20 July 1897, Giraldi (a, N. sinensis var.); Lungchan, Li-kia- po, 1800 m., 4 July 1922, Hers 2370 (A). Hunan: Yang-shan, Changning Hsien, 740 m., 13 July 1935, Fan & Li 242 (a, N. sinensis var.). NEILLIA IN CULTIVATION Through the kindness of the owners and/or curators of many gardens, I have had the opportunity to survey the Neillia material in cultivation in Britain. Specimens of all the material sent to me are now in the her- barium at Liverpool University Botanic Gardens (Livu-Bc). A few specimens of cultivated plants were also included among the herbarium material studied. All this together shows that the following taxa are, OF ve been, in cultivation: WN, thyrsiflora, N. affinis var. affinis and vat. pauciflora, N. uekii, N. sinensis var. sinensis and var. ribesioides, and, N. thibetica, A search through the reley of Neillia reputed to be in cultivati racemosa (= thibetica) is y The following simplified ke taxa: vated y will serve for the identification of the culti- 1971] CULLEN, THE GENUS NEILLIA 155 1. Leaves on flowering shoots with multiple, a far buds; floral tube strigose; inflorescence usually a much branched PNG: : sain, PROT ee, Baa ria apices ook a al bpd aoe’ aid lai'w xeieihvur Ooo se casiae yrsiflora, . Leaves on flowering shoots with single seat floral tube ee or gla- brous, not strigose; inflorescence usually a me. 2. Floral tube campanulate, as broad as or sites than long. 3. Raceme elongate, the flowers 10 or more, spaced ne the axis. ...... BUNS toh irs xs AN ee ee ge gry che aa nis var. affinis, cA sae contracted, the 5-10 flowers one at “the ends of the MSS rs eters fe ieee sgt MEADE. N. affinis var. pauciflora. 2. Floral ie cylindric or cylindric-campanulate, longer than broa 4. Axis of inflorescence with tuberculate-based, stellate hairs; ovules #: SE eA ER pee mee ee UE pn MRE PRE AU oA er te ot Sas ee . uekii. 4. te of inflorescence glabrous or pubescent with simple hairs; ovules —_ a Floral tube glabrous or with sparse patent hairs near the base; flowers 9-17(—21) in each racem 6. Floral tube less than is 5 mm. long. . N. sinensis var. ribesioides. 6. Floral tube 6.5-10 mm. long. ...... N. sinensis var, sinensis. 5. Floral tube finely pres: * pubescent: flowers oe in each WACOM ros haces eh pe Oa Eee . thibetica. Neillia thyrsiflora D. Don I have seen no recent material of this species, which according to Rehder was introduced in 1850. I have seen the following specimens: beam towards W. end, 2 July and 27 Aug. 1937, Anon. (wsy). Germany. Bot- anischer Garten der For. stakademie, Muenden, Hanover, orig. Haage and Schmidt, Erfurt (sub Spiraea), 1873, H. Zabel 86 (BH). The Wisley plants have presumably both died, as neither was included in the living material sent to me from there. I have not seen any wild material under Farrer 54 (see first specimen cited), but Farrer and Purdom 4 is N. sinensis. Neillia affinis Hemsl. var. affinis Introduced in 1908, according to Rehder, but not widely distributed in gardens. I have seen living material from Kew, Cambridge, and Wisley; the Cambridge specimen was obtained from Kew (fide J. S. L. Gilmour in litt.) and was grown as NV, thibetica; the Wisley specimen was grown as N. = ase ah Britain. R. H. S. Gardens, Wisley, 30 May 1950, F. G. Meyer 3330 (sn); ibid, May ek Anon, (BH). Neillia affinis Hemsl. var. pauciflora (Rehd.) Vidal Another taxon apparently not grown today, but represented in the Wisley herbarium by the following: 156 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Great Britain. R. H. S. Gardens, Wisley, from the W. end of 7 acres; 25 June 1941, Anon. (wsy). This plant, which was grown as NV. longiracemosa, and later identified as N\. thibetica is of unknown origin; it may well have been grown from seed collected by Forrest. Neillia uekii Nakai Grown extensively at the Arnold Arboretum, introduced in 1906. I have seen living material from the Royal Botanic Garden, Edinburgh, and the following specimens: United States. Arnold Arboretum, Jamaica Plain, Mass., 1 June 1926, S. F. Blake (Na-48213); ibid., 12 July 1918, Schneider (NA—48212); ibid., 21 June 1918, Schneider (NA-48211); ibid. 4 June 1936, P. Russell 90 (wa-11465); ibid., 2 Sept. 1921, G. M. Merrill (BH). Japan. Kyoto city, 20 iv 1954, Murot 39 (A); ibid., 2 Aug. 1954, Muroi 764 (a). Neillia sinensis Oliv. var. sinensis Introduced in 1901, this is the largest-flowered of all the species, though the smaller number of flowers makes it a less showy bush than good forms of NV. thibetica. I have seen living material from the Royal Botanic Gar- dens, Kew, and the Royal Botanic Garden, Edinburgh, and the following specimens: Great Britain. R. G. B. Edinburgh, from seed of Wilson 189, Edinburgh C642 (e); ibid., from seed of Wilson 4220, Edinburgh C87 (£). United States. Arnold Arboretum, Jamaica Plain, Mass., 5 June 1926, S. F. Blake (Na—48208); ibid., 9 June 1920, Anon. 6797-1 (BH): ibid., 20080-A, 22 June 1936, E. Nilbending (BH); ibid, A. A. 20080, 3 June 1936, P. Russell 48 (NA-11433); ibid., 3 Aug. 1921, G. M. Merrill (sx); garden of B. Y. Morrison, Piney Branch Road, Takoma Park, Md., 13 May 1944, B. Y. Morrison (Na-161974, with photo- graph, Neg. 87983; BH): The Hickories, Cazenovia, N.Y., 20 July 1939, 5-6’, S. R. Hafely (NA~188174; BH); Bennington, Vermont, 4 June 1937, E. T. & H. N. Moldenke 9696 (BH). N. sinensis Oliv. var. ribesioides (Rehd.) Vidal The only cultivated s at Liverpool University it is a large, spreading sent to me from elsewh other taxon. pecimen of this taxon that I have seen is growing Botanic Gardens, Ness. Its origin is unknown, but bush, obviously of considerable age. All specimens ere as N. ribesioides have proved to belong to some N. thibetica Bur. & Franch. of the Botanical Magazine. 1971] CULLEN, THE GENUS NEILLIA 157 I have seen living material from the following gardens: R. B. G. Kew, R. B. G. Edinburgh, Chelsea Physic Garden, University of Exeter, Keil- lour Castle, University of Liverpool, University of London, R. H. S. Wis- ley, Messrs. Hilliers’, Winchester; and the following specimen: Netherlands. The Hague, Zuiderpark, 22 June 1956, B. K. Boom 32240 (BH). Several species of Physocarpus (which at one time was included in Neil- lia) may still be grown in some gardens as NV eillia; the distinctions be- tween the two genera are given in the introduction. INDEX TO NUMBERED SPECIMENS CITED Species are numbered 1-10, as in the revision; varieties are indicated by a, b, c, etc., in the order in which they are given in the revision. Anderson 505: 1a. Henry 5554A: 9a; 5554B: 9a; 5585: Banerji 1210: 1a. indet.; 5695: 9a; 8968: 3a; 9419 Cavalerie 2461: 9b. and 9419A: 1b; 9669: ie 10231A: Cavalerie & Fortunat 2461: 9b. Sc: 12273) tae 2302: Cave 56: 1a; 97/47: 4. Hers 2343: 9a; 2370: ae 2965: Chiao 1633: 10. indet Ching mee 9b; 20483: 9f; 30281: 5; Howell ‘200: 81758: Kingdon-Ward 1584: 3d; 3241: 1a; Chow ae rae 3086: 5% 21735* ib. Chu 2827: 10; 3611: 10. Ko 55787: la. Chun 3795: 9a. Lace 2347: la. Cooper Ste la; 1229: 1a; 2069: 1a; Lau 28719: 2814: Ludlow, Sherriff & Hicks 16520: 4; Cowan = of 19288: 4; 20495: Caosingham sees 10, Maire 41: 9f; 310: of ; 387: 9f; 429: Delavay <5. OF < $20 OE 9595S} 2100; 55 F108: pateeeihg 21043: la. Of; 1120: 9e; aie d. Ducloux 633: 9. M cLaren Mask Fan & Li 242: indet Mills 107: 6. Fang 935: 3a; 1057: 9b; 4694: 3a; Mombeig 11: 9f; 78: 9f. 10315: 9c. Nakai 1774: Farges 494: 9c. Pratt 347: 3a; 730: 10. Farrer 1757: 3a. Purdom 1: 9a; 467: 9a. Farrer & Purdom 54: 9a. Relzada 19034: 1a. Faurie 85: 6; 316: 6. Rock 4079: 5; 4422: 5; 7056: la; Feng 1936: 5. 9171: 9f; wig Se 23275: 3a. Forrest 554: 9f:; 1871: 5: 2244: 5: Rohmoo Lepcha 7 6033: 5; 7584: indet.; 7854: 3c; Schneider 1506: 5, ey ge 3 So i 17605: 3d; 29567: 3c; 29897: 1b; Silvestri 4589: 9a. 30023: 3a: 30475: Of. Smith 1865: 3a; 2043: 3d; 2078: 10. Gallatly 214: 1a. Steward & Cheo 172: 9b; 450: 9b. Griffith 2115: indet. Sun & Chang 1234: 3c af eaetn 765: 9a; 1644: in- Ten 462: 8; 8. €t.; 2218: 7; 2960: 5: 7867: in- Teng 90524: 9 : det.; 8884: 3a; 12010: 9a. Tsai 50929: 9c; 50949: 3a; 50992: 158 JOURNAL OF THE ARNOLD ARBORETUM [voL, 52 9b; 51041: 10; 51196: 9b; 52055: Wang & Liu 81820: 1a; 83503: 1b. 3a; 52706: 9b; 56350: 1a; 56895: Watt 6532: indet. 1) Oeraasiag oi) ay 1a; Wilson 86: 9a; 189: 9a; 701: 9a; has eure . ee, ee : ; : 16: 10; 916a: 3; 974: 10; 23am det.; 60438: indet.; 60717: 1a; 920° 10; 9400 , ‘ ; 3a; 2381: 10; 2382: 9c andaim sa niet lees 3558: 10; 3559: 3a; 3560: 3a; Tsang 27882: 9b. : 9530 Tsiang 5744: 9b. 4220: 10; 4220a: 10; 4886: 3d; Wang 22992: 3a; 23084: 10; 64159: Yu 740: 10; 941: 3a; 7855: ae 9f; 65196: 3a; 66940: la: 67076: 15282: 5; 17115: 3a; 17297: 3a; la; 67945: 9a; 68154: 3b; 70883: 17974: 1b; 19228: 1b; 19409: la; 5. 20199: la. ACKNOWLEDGMENTS I would like to thank the curators of the following herbaria for the loan of material: Royal Botanic Garden, Edinburgh (£); Royal Botanic Gar- dens, Kew (k); Royal Horticultural Society, Wisley (wsy); Arnold Ar- boretum (A); U.S. National Arboretum, Washington (Na); L. H. Bailey Hortorium, Cornell University, Ithaca (pu). I would also like to thank the owners or curators of many gardens (too numerous to mention individually) for their help in finding cultivated ma- terial; and finally, Dr. J. E. Vidal (Paris) for sending me a copy of his paper on Neillia; Mr. A. Lauener (Edinburgh) for help with Chinese place names and localities; and Mrs. K. Purcell (Ness) for preparing floral dissections. LIVERPOOL UNIVERSITY Botanic GARDENS Ness, NEsTon, CHESHIRE, L64 4AY ENGLAND 1971] ELIAS, GENERA OF FAGACEAE 159 THE GENERA OF FAGACEAE IN THE SOUTHEASTERN UNITED STATES? Tuomas S. Exias FAGACEAE Dumortier, Anal. Fam. 11, 12. 1829, “Fagineae,” nom. cons. (BEECH FamIty) “Prepared for a generic flora of the southeastern United States, a project of the Arnold Arboretum and the Gray Herbarium of Harvard University made possible through the support of the National Science Foundation (Grant GB-6459X, principal investigator, Carroll E. Wood, Jr.). This treatment follows the format established in first paper of the series (Jour. Arnold Arb. 39: 296-346. 1958). The area covered area, with additional information in brackets. References which I have not seen are marked by an asterisk. : I am grateful to Dr. Wood for his thorough review of the manuscript and his humerous suggestions and additions in the course of this study. Dr. E. C. Abbe has indly read the manuscript and has offered a number of useful and much-appreciated comments, Mrs. Nancy Dunkly has helped greatly in checking the bibliography and typing the manuscript. The illustrations are by Virginia Savage. 160 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 A family of eight genera and 400-500 species distributed throughout the temperate regions of the world, in mountainous tropical areas from Central America south to Colombia and the West Indies, and in Old World tropical areas, except tropical and southern Africa. Five genera are found in the continental United States, with Fagus L., Castanea L., and Quercus L. in our range. Chrysolepis Hjelmqv. is restricted to the Pacific coast region of the United States, and the chiefly Southeast Asian genus Lithocarpus Blume has a single species in California and Oregon. The Fagaceae and the closely related Betulaceae (including Corylaceae) have generally been considered as constituting the order Fagales (Wett-_ stein, Tippo, Rendle, Hutchinson). Current opinion among phylogenists is that the Fagales are derived from hamamelidaceous stock, which (fide Takhtajan) may serve as a link between the Magnoliales and the Amenti- ferae. Bessey, however, placed the Fagaceae with the Betulaceae and Juglandaceae and considered all three to be derived from a sapindaceous line. The Fagaceae have also been considered to have been derived from epigynous rosaceous stock (Berridge, 1914), but this view has not been accepted by later taxonomists. Hjelmqvist (1948), followed by Takhta- jan (1969), preferred to place the Fagaceae and Betulaceae in separate but closely related orders. Cronquist (1968), in addition to including the Betulaceae and Fagaceae in the order Fagales, added the Balanopaceae. The affinities of this last family appear to be with the Fagales, but it should be studied further to determine its relationships more exactly. Using evidence primarily from the carpellate inflorescences and fruit, Forman (1966) has postulated the lines of evolution of the cupules in the Fagaceae and has established intergeneric and specific relationships based on his findings. The basic structure of the carpellate inflorescence is considered to be a three-flowered dichasium (Ficure 1). In the least specialized condition, each of the flowers and subsequent fruits is suf rounded by several lobed involucres, In C hrysolepis the three trigonous fruits are enclosed by five free outer valves and two inner ones. From this condition it is supposed that the ancestral type had fruits that were trigonous (versus irregular or rounded) in section and that each of the three fruits of the basic dichasium was surrounded by three free valves; a condition that has not been found in any of the extant Fagaceae. Pro- gressive specialization resulted in reduction in the number of involucral lobes and, later, fusion of th found in Quercus, the C. Lithocarpus. Specialization abortion) of the basic thr radi from trigonous to + fe occurred concurrently with the reduction and fusion of the involucté- Bas lee are illustrated diagrammatically in Ficure 1 and = a ene at each of the three generally recognized subfamilies pe as ; Schwarz, 1936; Forman, 1964), which can be — 1971] ELIAS, GENERA OF FAGACEAE 161 Inflorescence of aments of dichasial clusters or of solitary flowers. Stigmas terminal, minute, punctiform; styles cylindrical; staminate flowers often with a rudimentary gynoecium; stamens usually 12: anthers tinute, dorsifized or versatile. =. 7. 3 oe. Castaneoideae. Stigmas on the inner surface of the generally flattened styles; staminate flowers without a rudimentary gynoecium; stamens usually 6; anthers Meri May ce RRO, ey ides udies, Jiutdee. Saeed Quercoideae. The subfamily Fagoideae includes only Fagus and Nothofagus. In Fagus the staminate flowers are in many-flowered heads, the carpellate dichasia lack the central flower of the basic three-flowered dichasium, and the styles are elongate and tapering, while in Nothofagus Blume the staminate flowers typically are in sessile to short-pedunculate one- to three- flowered dichasia, the central flower is always present in the carpellate dichasia, and the styles are short. That fruiting structures apparently have not evolved as rapidly as the floral parts, may be seen in the persistence of both free involucral valves and trigonous fruits. These tendencies are demonstrated in Nothofagus, some species of which exhibit the less special- ized features, while others have more advanced characters, and many clearly show transitional features (cf. Ficure 1). Fagus has 3(4)-colpor- ate, suboblate-subprolate pollen grains, and Nothofagus has (4—)6(-7)- aperturate, oblate pollen grains. Of the four genera in the Castaneoideae, only Castanea Mill. occurs in our range. The Old World Castanopsis Spach (approximately 110 species) is separated from Castanea mainly on floral and fruiting characters (see Castanea). Recently segregated from Castanopsis, Chrysolepis Hjelmav. is distinguished by free involucral lobes and fruits (three per involucre) separated from each other by inner valves. Lithocarpus Blume (Pasania Oersted), including ca. 100 species distributed mainly in Malesia and the Pacific Islands and L. densiflorus (Hook. & Arn.) Rehd. in southwestern Oregon and in California, is distinguished by having each carpellate flower of a dichasium surrounded by its own involucre, which never shows vertical divisions or lobation. Langdon (1949) suggested several evolutionary trends in the floral structures in this subfamily, including a noticeable reduction in the inflorescence, apparently from a primary many-flowered dichasial type, and a reduction in floral members. There are also various transitional stages between hypogyny and epigyny that should be studied anatomically. Forman (1966) concluded that reduction of the basic three-flowered Carpellate dichasium to a single flower took place by suppression of the lateral flowers and that reduction by fusion of the four-valved cupule, first to a two-valved one and then to a single cup with no signs of valves, has occurred. The reduction of the individual flower-cupules within a dicha- sium is seen in Lithocarpus as eventually leading to a single flower and, Consequently, a single fruit that closely resembles that of Quercus. This Convergent evolution has led many workers to think that Lithocarpus and 162 JOURNAL OF THE ARNOLD ARBORETUM [voL, 52 FAGACEAE Castaneoideae Quercoideae Fagoideae 0@> N. cornuta N. resinosa Castanopsis . 6 (Fisen group) : Lithocarpus : ® : i Quercus N. pullei ee : : : oop : : a : oo : Castanea : N. gunnii : i Castanopsis = N. antarctica © N. pumilio re) | AF) : O) Lithocarpus : Trigonobalanus : N. brassii : : A» do ee Sr : r)s : v4) t Se : pee : Cestemls % N. alessandrii A % a bh he : ace rf ” : ri wa _ a : ) O > ° scapans = Oe ceses 4 r ) > is aS. a \ matin, NUE Fagus Lithocarpus Nothofagus WA ri Chrysolepis 3 Hypothetical ancestor : = ee scheme Peay gr evaadl a relationships oe _based cre, flow involucre. No Ss been very variatio e Fagaceae, Based sith on Fo pai (1966) and ‘th Steenis (1953). Quercus are eo aoa = these workers have failed to consider the total biology of the O gen he fede in he | number of valves by lateral fusion is best see? 1971] ELIAS, GENERA OF FAGACEAE 163 in Castanea and Castanopsis. A four-valved cupule, usually with three fruits is considered to have undergone fusion of the lateral valves to form a two-valved cupule, and, along with this, the number of flowers has been reduced to one per cupule. Further reduction can be seen in the “fissa- group” of Castanopsis, in which the two-valved cupule has given way to a completely fused, one-flowered cupule (cf. Ficure 1). The subfamily Quercoideae includes only Quercus (300-350 species) and the recently described Trigonobalanus Forman (2 species: T. doi- changensis (A. Camus) Forman, of northern Thailand, and T. verticillata Forman, of Borneo and Celebes). Evolutionary tendencies in the Quercoideae are toward a reduction of the carpellate clusters of flowers to either a pair of flowers or a solitary one; toward a change in shape and reduction of the number of fruits per cupule; and toward a reduction by fusion of the cupular valves to form a single valveless cupule. Forman (1964) considered Trigonobalanus to be less advanced than Quercus and perhaps close to the ancestral stock of the Fagaceae, an idea based primarily on the carpellate flowers, i.e., the flowers borne in clusters of 3(—7) as compared with the presumably derived condition of the paired or solitary carpellate flowers in Quercus. In addition, the 1-3 (—7) trigonous fruits of Trigonobalanus are borne in a 3—5-valved cupule. In Quercus, a reduction of the flowers from 3 to 1 apparently has occurred, along with a change in fruit shape from triangular to round. Reduction of the cupule valves and their probable fusion to form the single valveless cupule of Quercus are further specializations. Fagaceous remains (largely leaves and leaf fragments) have been recog- nized in the Upper Cretaceous, and these indicate that the family was once far more widespread than the extant members are. The Fagaceae are of considerable economic importance. Members of this family, because of their abundance and very suitable wood, are one of the most important sources of hardwood timber. The wood, valued for its tough, hard, strong, yet attractive appearance, is used in various phases of construction, furniture making, and in quality tools. Quercus Suber L., cork oak, native to the Mediterranean region is the source of commercial cork, Edible nuts and, to a limited degree, lumber are obtained from Castanea. Species of Fagus are also used as lumber sources, and the beech- nuts are highly sought after by various animals, both domestic and wild. REFERENCES: ARMSTRONG, J. M., & A. N. Wyte. A new basic chromosome number in the family Fagaceae. Nature 205: 1340, 1341. 1965. [m = 13 in Nothofagus Menziesii, N. solandrii var. clifortioides, & N. fusca.]| Battey, I. W. Notes on the wood structure of the Betulaceae and Fagaceae. Forestry Quart. 8: 178-185. 1910. Battton, H. Castanéacées. Hist. Pl. 6: 217-259. 1877. (English ed. 6: 220-262. 188 0. BENson, M. Contributions to the embryology of the Amentiferae.— Part I. Trans. Linn. Soc. Bot. 3: 409-424. pls. 67-72. 1894. 164 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 BentHamM, G., & J. D. Hooker. Cupuliferae. Gen. Pl. 3: 402-410, 18853. [In- cludes tribes Betuleae, Coryleae, & Quercineae. | Berriwce, E. M. The structure of the flower of the Fagaceae, and its bearing on the affinities of the group. Ann. Bot. 28: 509-526. 1914. Berry, E. W. Tree ancestors. vi + 270 pp. Baltimore. 1923. [Fagus, 119-128; Quercus, 129-140; Castanea, 140-145.] Brett, D. W. The inflorescence of Fagus and Castanea and the evolution of the cupules of the Fagaceae. New Phytol. 63: 96-117. 1964. CeLakovsky, L. Uber die Cupula von Fagus und Castanea. Jahrb. Wiss. Bot. 21: 128-162. pl. 5. 1890. Couper, R. A. Southern Hemisphere Mesozoic and Tertiary Podocarpaceae and Fagaceae and their paleogeographic significance. Proc. Roy. Soc. London B. 152: 491-500. 1960. [Includes maps. | Davis, D. E., & N. D. Davis. Guide and key to Alabama trees. viii + 136 pp. Wm. C. Brown Co., Dubuque, Iowa. 1965. [Castanea 42, 53; Fagus, 51; Quercus, 88-99. | EtcHier, A. W. Cupuliferae. Bliithendiagramme 2: 20-32. 1878. Erptman, G. On the pollen morphology of Trigonobalanus (Fagaceae). Bot. Not. 120: 324-333. 1967. Forman, L. L. Trigonobalanus, a new genus of Fagaceae, with notes on the classification of the family. Kew Bull. 17: 381-396. 1964. On the evolution of cupules in the Fagaceae. /bid. 18: 385-419. 1966a. —. Generic delimitation in the Castaneoideae. Jbid, 421-426. 1966b. D. F. Curter. Additional notes on Trigonobalanus Forman (Faga- ceae). Ibid. 21: 331-334. 1967. [Trigonobalanus in Malaya & Sarawak; further notes on the anatomy of the vegetative organs. FoweELts, H. A. Silvics of forest trees of the United States. U. S. Dep. Ast. Forest Serv. Agr. Handb. 271. 1965. [Fagus, 171-180; Quercus, 557-640; distribution maps by E. L. Lirt ez, Jr.] Groom, P. The medullary rays of Fagaceae. Ann. Bot. 26: 1124, 1125. 1912. Hyetmgvist, H. Studies on the floral morphology and phylogeny of the Amentt- ere Bot. Not. Suppl. 2(1): 1-171. 1948. [Fagaceae, 77-121; Chrysolepts, sia Gy ————. Some notes on the endosperm and embryo development in Fagales and related orders. Bot. Not. 110: 173-195. 1957. . i es on some names and combinations within the Amentiferae. / bid. 113: 373-380. 1960. | Chrysolepis, 377. ] HUTCHINSON, J. The genera of flowering plants. vol. 2. xi + 659 pp. Clarendon Press, Oxford. 1967. [Fagaceae, 126-132.] JaRETzKy, R. Zur Zytologie der Fagales. Planta 10: 120-137. 1930. Jaynes, R. A., ed. Handbook of North American nut trees. vii + 421 PP- Northern Nut Growers Association. Knoxville, Tenn. 1969. [Quercus, ai 342; Fagus, 342, 343; Castanea, 264-286. Kuprianova, L. A. The palynology of the Amentiferae. (In Russian.) 2 14 PP. 48 pls. Bot. Inst. Komarov. Akad. Nauk SSSR. Moskava-Leningrad. 1965. =e H., & R. K. Goprrey. Trees of northern Florida. xxxiv + 311 PP. : a Florida. 1962. [Castanea, 57-59; Fagus, 60-62; Quercus, = eae L. M. Ontogenetic and anatomical studies of the flower and fruit of the Fagaceae and Juglandaceae. Bot. Gaz. 101: 301-327. 1939. - The comparative morpholo Notho- : gy of the Fagaceae. I. The genus fagus. Ibid. 108: 350-371. 1947. 6 s 1971] ELIAS, GENERA OF FAGACEAE 165 LEBEDENKO, L. A. The ontogeny of root and stem wood in certain representa- tive of Fagales. (In Russian.) Dokl. Akad. Nauk SSSR 127: 213-216. 1959.* Lirtte, E. L., Jr. Check list Kg native and naturalized trees of the United States (including Alaska). U. S. Dep. Agr. Agr. Handb. 41. 472 pp. 1953. [Cas- tanea, 92-94; Fagus, ay 184; Quercus, 311-361. Lussock, J. A contribution to our knowledge of seedlings. vol. 2. 646 pp. New York. 1892. [Cupuliferae, 526-541; includes genera from Fagaceae & Martin, A. C. The comparative internal morphology of seeds. Am. Midl. Nat. 36: 513-660. 1946. [Fagaceae, 638, 646. Poote, A. L. The development of Nothofagus seed (including a preliminary account of the embryogeny, etc.). Trans. Roy. Soc. New Zealand 80: 207— 21251952; PRANTL, =a Beitrage zur Kenntnis der Cupuliferen. Bot. Jahrb. 8: 321-334. 188 —. a gaceae. Nat. Pflanzenfam. III. 1: 47-58. 1888. Reupber, A. Bibliography of cultivated trees and shrubs. xl + 825 pp. Jamaica Plain, Mass. 1949. [Fagaceae, 112-135.] SARGENT, C, S. Cupuliferae. Silva N. Am. 8: 1-190; 9: 1-30. 1895, ———. Manual of the trees of North America (exclusive of Mexico). ed. 2. xxvi + 910 pp. Boston & New York. 1922. [Fagaceae, 227-308; includes map. ScHwarz, O. Entwurf zu einem natiirlichen System der i und die Gattung Quercus L. Notizbl. Bot. Gart. Berlin 13: 1-22. 1936. Sears, P. B. Postglacial migration of five forest genera. bas Jour. Bot. 29: 684-691. 1942. [Includes Quercus & Fagus. | SHmmajr, K. Anatomical studies on the wood of the Japanese Pasania, Castanea and Castanopsis. (With a key to the 22 Japanese representative species of the Fagaceae.) Bull. Tokyo Univ. Forests 55: 81-99. 1959.* STEENIS, C. G. G. J. vAN. Results of the eit Expeditions: Papuan Notho- fagus. Jour. Arnold Arb. 34: 301-374. 1 Additional note on Nothofagus. ee: 35: 266-267. 1954. TAvLor, L. A. Plants used as curatives by certain southeastern tribes. Part I. Plants used medicinally. xi + 88 pp. Botanical Museum, Harvard Uni- versity. 1940. [Castanea pumila, 16; ee borealis var. maxima = Q. rubra, 16, 17; Q. marilandica, 17; Q. rubra = Q. falcata, 17, 18; Q. stellata, Tippo, 0. Comparative anatomy of the Moraceae and their presumed allies. Bot. Gaz. 100: 1-99. 1938. [Fagaceae, 38-42. ] Vines, R. A. Tr rees, shrubs and woody vines of the southwest. xii + 1104 pp. Austin, Texas. 1960. [Fagaceae, 147-202. WALKER, D., & G. WitTMANN. Notes on the pollen morphology of Nothofagus Bl. subsection Bipartitae Steen. Pollen Spores 7: 457-464. 1965. West, E., & L. ARNOLD. The native trees of Florida. xx + 212 pp. Gainesville, Florida. 1946. [Fagaceae, 33-53. ] Wovenouse, R. P. Hayfever plants. xx + 245 pp. Chronica Botanica, Wal- m, Mass. 1945. [Fagus, 75, 76; Quercus, 76-81.] Us8; Dev. Acr. Forest Serv. Woody-plant seed manual. U. S. Dep. Agr. Forest Serv. Misc. Publ. 654. 1948. [ Castanea, 112; Fagus, 174, 175; Quer- cus, 297-304.] 166 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 KEY TO THE GENERA OF FAGACEAE IN THE SOUTHEASTERN UNITED STATES General characters: Trees or shrubs with alternate, simple leaves with small usually caducous stipules ; flowers unisexual; staminate flowers in erect or pendu- lous aments or heads, the perianth 4-8-lobed; carpellate flowers 1-4 mm an m- , the ovary inferior, 3—6-loculate with 1-2 ovules ber locule; involucre becoming indurated in fruit, lobed or entire; fruit usually @ mut, 1-seeded by abortion; seed without endosperm. A. Staminate flowers in den se many-flowered, pendulous heads; the fruit triangu- lar; germination epige ous; bark smooth; buds lanceolate, ti pendulous aments; carpellate flowers borne separately, the ovary 3(4~5 -loculate, the styles generally 3, stigmatic on the inner surface; involucre cupuliform with numerous usually 1m- bricate scales, partially inclosing the single fruit. 3.Q Subfamily FAGOIDEAE [ Oersted } Fagus Linnaeus, Sp. Pl, 2: 997. 1753; Gen. Pl. ed. 5. 432. 1754. Deciduous trees 09 habitats in tem i , airy bracts, the outer bracts often bright red, longer than flowers, the j ate, villous, with 4 or 5 linear-lanceolate acute nth lobes, the tube tri € and adnate to the 3-locular ovary; styles 3, angulat 1971] ELIAS, GENERA OF FAGACEAE 167 io _ Ficure 2. Fagus. a-l, F. grandifolia: a, flowering branchlet, x %; b, staminate inflorescence, X 2: c, staminate flower, X 6; d, carpellate inflorescence, X 3; 6; g, fruiting branchlet, x 14; h, involucre and two fruits, X 2; i, apace bracts unfolding to show mature fruits, 114; j, fruit in cross section to show cotyledons, X 2; k, seedling, X %4; 1, winter twig showing bud, X 114. 168 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 52 slender, recurved, pilose, exserted, becoming stigmatic toward the apex; stamens wanting; ovules 2 in each locule, suspended from apex of locule, amphitropous, the micropyle superior. Involucre (in fruit) becoming woody, pedunculate, ovoid to subglobose, covered with [sometimes glan- dular] stout, recurved prickles or tubercles, inclosing 2—4 fruits, ultimately separating into 4 valves. Fruits ovoid, unequally triangular, acute or winged at the angles, concave and longitudinally ridged on the sides, chestnut-brown and lustrous at maturity, rostrate (remnants of style); pericarp thin, of two closely united integuments, the inner membranaceous. Seed dark chestnut-brown, solitary, without endosperm; cotyledons thick and fleshy, plano-convex, plicate and + united, sweet, oily; radicle minute, superior; germination epigeal. LecrorypE sPEctEs: F. sylvatica L.; see Britton, N. Am. Trees 271. 1908. (Classical Latin name derived from Greek phagein, to eat, in reference to the edible nuts.) — BEECH. A genus of eight to ten species, represented in the southeastern United States by Fagus grandifolia Ehrh., 2n = 24,2 (F. ferruginea Ait., F. americana Sweet), which occurs from Cape Breton Island and Nova Scotia to Maine, southern Quebec and southern Ontario, northern Mich- igan and eastern Wisconsin, south to the Wabash and Mississippi River valleys in southern Illinois, southeastern Oklahoma, eastern Texas an eastward to northern Florida. (See Map.) Varietas caroliniana (Loud.) Fern. & Rehd., distinguished by its darker leaves, which are often thicker, rounded or subcordate at the base, and smaller toothed, is found from Florida to Tennessee, north to southeastern Massachusetts, Ohio, Indiana, southern Illinois, and Missouri. Fagus mexicana Martinez, known from four localities in the Sierra Madre Oriental in the states of Tamaulipas, Hidalgo, and Puebla, Mexico best treated as F. grandifolia var. mexicana ommon orobanchaceous parasite of beech, Ep t., is scattered throughout the range of F. grandi- in two of the Mexican populations, but is as yet Fagus sylvatica L., 2n = 24, often cultivated in the United States 1s ence that Fagus occurred farther west in the Unite States than the present limits of the genus. Numerous fossilized leaves ™ * Aufderheide (1931 ‘ : i view of later mainte Wee ant haploid number of F. grandifolia as 6, but ” = 6 count must be considered questionable. 1971] ELIAS, GENERA OF FAGACEAE 169 ’ \ Map 1. Present distribution of Fagus grandifolia. Distribution in United States mostly generalized; disjunct populations in Tamaulipas, Hidalgo, and Puebla, Mexico, are var. mexicana. Based mainly on Little in Fowells (1965) and Fox & Sharp (1954). Europe from Miocene and Pliocene deposits have been attributed to the New World F. grandifolia by Tralau (1962). The present disjunct dis- tribution of Fagus probably represents the remnants of a once widespread pan-temperate genus of numerous species. Three series of populations of Fagus grandifolia, each with distinct ecological requirements, along with areas of introgression, were recognized by Camp (1951) but not given formal taxonomic status. The white beech, €st seen in moist river valleys, is found in the southeastern and southern United States, Typically found on the well-drained southern Appalachian slopes on usually stabilized talus, the red beech extends from there north- ward to Canada. The northern gray beech is adapted to the cool, moist area at the southern edge of the northern coniferous forest and occurs southward at higher elevations than the red beech. It commonly occurs in pure stands above 650 m. in New England and Canada. A detailed ecological and/or biosystematic study of the three series is needed to determine not only the exact relationships of the various populations, but the status of var. caroliniana and var. mexicana. Fagus grandifolia and Acer saccharum are the climax species in much of the deciduous forest formation from west of the Allegheny mountains to New York, Ohio, and Wisconsin. The shade-tolerant beech is often found in nearly pure stands in many areas within its range, e.g., higher 170 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 altitudes in the southern Appalachians (Russel, 1953). Fagus in our range is generally more abundant on the moister northern slopes and on soil with pH 4.1 to 6.0. The anemophilous flowers appear in late April or early May, shortly after the leaves begin unfolding. Cross pollination is generally necessary for seed production, plants of most species of Fagus being moderately to largely self-sterile. Individual trees seem to show varying degrees of self- fertility in successive years, but this needs to be carefully studied and documented. Neilson & Schaffalitzky (1954) reported proterandry in F. sylvatica and F, grandifolia varying in degree with the weather, warm dry spring conditions resulting in shorter periods of proterandry. They found the stigmas of F. sylvatica to be receptive to pollen for approximately 10-14 days. Artificial hybrids between F. sylvatica ( 2 ) and F. grandifolia ( é ) were produced, the progeny showing intermediate leaf characters and some tendency toward hybrid vigor. The fruits require only one growing season to mature and are shed when the first heavy frost causes the involucral valves to separate. Germina- tion takes place in early spring to early summer of the following year. Seed production is somewhat sporadic, with good crops every two to three years. Dispersal is mainly by rodents or large birds. Although seed germination and seedling development are common in beech stands, vegeta- tive reproduction by root suckers is often encountered. In open areas, beech trees develop a thick, short trunk with large, low spreading limbs and a broad, rounded crown. The thin, pale bark is very susceptible to fire injury and overexposure to sunlight (sunscalding). A number of forms of Fagus are cultivated, Of these the copper beech, F, sylvatica {. atropunicea (Weston) Domin, is probably most frequently grown as an ornamental in the United States. Several of the Asian species are also cultivated to a limited extent in the Northeast. In New Eng- land and Canada, Fagus grandifolia is susceptible to attack by the beech scale insect, Cryptococcus fagi, which is soon followed by the fungus Nec- tria coccinea var. faginata. Several species of the fungus Fomes occur on the beech in our range. The wood is hard, strong, and very close grained but not very durable. It is used in the manufacture of chairs and tool handles and is perhaps the principal wood used for clothespins (Strausbaugh & Core, 1953). REFERENCES: Under family references see also Berry, 119-128; ForMAN, 1966a; FOwELLS, 171-180; HyELmovist, 82-86; Kurz & Goprrey, 60-62; LANGDON, 303-316; LITTLE, 183, 184; SEars, 689; STEENIS; VINES, 198, 199; West & ARNOLD, 34. AUFDERHEIDE, H. Chromosome numbers in Fagus grandifolia and Quercus vir- gumana. Butler Univ. Bot. Stud. 2: 45-52. 1931 Banputska, H. On the cuticles of some recent and fossil Fagaceae. Jour. Linn. Soc. Bot. 46: 427-441, 1924 BENNINGHOFF, W. S., & A. I. GEBBEN, Phytosociological studies of some beech- 1971] ELIAS, GENERA OF FAGACEAE 171 maple stands in aren Lower Peninsula. Pap. Mich. Acad. Sci. Arts Lett. I. 45: 83-91. 1960. Berry, E. W. Notes on the ancestry of the beech. Pl. World 19: 68-77. 1916. [Includes maps. ] BLINKENBERG, C., H, Brix, M. SCHAFFALITZKY DE MucKADELL, & H. VEpEL. coe i in Fagus. Silvae Genet. 7: 116-122. 1958. [Includes Cie. W. H. A biogeographic and paragenetic free des of the American beech (Fagus). Am. Philos. Soc. Yearb. 1950: 166 951. CLowes, F. A. L. The structure of mycorrhizal sot of Fagus sylvatica. New Phytol. 50: 1-16. figs. 1-7. 1951. Domin, K. On the vaio of the beech. Acad. Tchéque Sci. Bull. Int. 33: 65-75. pls. 1-2. 1932. Du RievTz, G. E. Peoblens of bipolar plant distribution. Acta Phytogeogr. Suec. 13: 215— 282. 1940. Sahay Fagus & Nothofagus. | ETTINGSHAUSEN, C. F. . Uber Tertiare Fagus- pie der siidlichen Hemi- sphare. Sitz-ber. Aiea: Wins: Math.-Naturw. Wien 100(Abt. 1): 114-137. 9 Fox, W. B., & A. J. SHarp. La distribucién de Fagus en México. Soc. Bot. Méx. Bol. 17: 31-33. 1954. [Includes m ap. | FRIESNER, R. C. A preliminary study of growth in the beech, Fagus bprdee by the dendrographic method. Butler Univ. Bot. Stud. 5: 85-94. Fritts, H. C. An analysis of radial growth of beech in a central ong aie during 1954-55. Ecology 39: 705-720. 1958. [F. grandifolia.] & N. Holowaychuk. Some soil factors affecting the Pela of beech in a central Ohio forest. Ohio Jour. Sci. 59: 167-186. 1959. Garrison, H. J. Floral morphology and ontogeny of Fagus grandifolia Ehrh. Ph.D. Thesis, Pennsylvania State University. 1956. [Unpublished.] [See also Diss. Abstr. 17: 2777, 2778. 1957.] Gray, J. Temperate pollen genera in the Eocene (Claiborne) flora, Alabama. Science 132: 808-810, 1960. [ Fagus, 809. | Harey, J. L., & C. C. McCreapy. A note on the effect of sodium azide upon the respiration of beech (Fagus sylvatica) mycorrhizas. New Phytol. 52: 83-85. 1953. ———,;, ——, & J. K. Briertey. The uptake of phosphate by excised mycor- Bal sais roots of the beech. [I.] New Phytol. 49: 388-397. 1950; II. asad tion of phosphorus between host and fungus. /bid. 51: 56-64. pls. 1, 2. 1952; III. The effect of the fungal sheath on the availability of phosphate to the core. Ibid. 342-348; IV. The effect of oxygen concentration upon host and fungus. bid. 52: 124-132. 1953; V. The examination of possible sources of misinterpretation of the quantities of phosphorus passing into the host. Ibid. 53: 92-98. 1954: VI. Active transport of phosphorus from the fungal sheath into the host tissue. Ibid. 240- 0-252; VII. Active transport of P®2 from fungus to host during uptake of phosphate from solution. Jbid. 54: 296-301. 1955; VIII. Translocation of phosphorus in mycorrhizal roots. Ibid. 57: 353-362. 1958. [BRIERLEY coauthor of IV-VIII; McCreapy of =v, VL] ——, ——,, & J. A. Geppes. The salt respiration of excised beech mycor- thizas, ® The development of respiratory response to salts. New Phytol. 53: 427-444. 1954 i J K. Briertey, & D. H. Jennincs. The salt respiration of ? 172 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 excised beech mycorrhizas. II. The relationship between oxygen consump- tion and phosphate absorption. New Phytol. 55: 1-28. 1956. Jackson, B. D. A note on nomenclature. Jour. Bot. London 34: 352, 353. 1910. [Earliest publication of F. americana. | Jounsson, H. Chromosome numbers of twin plants of Quercus Robur and Fagus silvatica. Hereditas 32: 469~472. 1946. Lancpon, L. M. The comparative morphology of the Fagaceae. 1. The genus Nothofagus. Bot. Gaz. 108: 350-371. 1947. Littte, E. L., Jr. Mexican beech, a variety of Fagus grandifolia. Castanea 30: 167-170. 1965. [F. grandifolia var. mexicana. | Locan, L. A. An ecological study of the American beech along the southwestern border of its distribution. Ph.D. thesis, University of Missouri, Columbia, Missouri. [Unpublished.] [See also Diss. Abstr. 20(9): 3487. 1960.] Martinez, M. Una nueva especie forestal (Fagus mexicana sp. nova). Anal. Inst. Biol. Méx. 11: 85~89. 1940 Mrranpa, F., & A. J. SHarp. Characteristics of the vegetation in certain tempe erate regions of eastern Mexico. Ecology 31: 313-333. 1950. [F. mexicana, 325~327. | MyczxKowskI, S. The influence of ecological factors upon the formation Y i habit of beech, Fagus silvatica L. (In Polish; English summary.) A Sect. Dendrol. Soc. B Bot. Pologne 10; 233-251. 1955.* NIELSEN, P. Chr., & M. SCHAFFALITZKY DE eons, Flower observations and ‘controlled pollinations in Fagus. Zeitschr. Forstgen. Forstpflanzenziicht. 3: 6-17. 1954. Porzcer, J. E., & A. N. Luwine. Secondary succession in stands of red maple- sweet gum-beech forests in Ripley County, Indiana. Butler Univ. Bot. Stud. 11; 50-59. 1953. [F. grandifolia. | Rice, H. P. A rough-barked American beech. Jour. Forestry 46: 48. 1948. Rosinson, R., & H. SmirH. Anthocyanins of the leaf of the copper beech (Fagus sylvatica) and the os of the cultivated strawberry (Fragaria virginiana). Nature 175: 634. 1955. Russet, N. H. The beech gaps of the Great Smoky Mountains. Ecology *: 366-374. 1953. SCHAFFALITZKY DE MucKADELL, M. Experiments on developments in Fagus silvatica by means of liethareous grafting. Physiol. Pl. 9: 396-400. 1956. STRAUSBAUGH, P. D., & E. L. Core. Flora of West Virginia. xxxi + 1075 pp. West Virginia Univ. Bull. 1952-64. [F. grandifolia, 298 | = S. Breaking dormancy of beech. Physiol. Pl. 10: 728-731. 1957. [Use f Rindite, consisting of ethylene chlorhydrine, ethylene dichloride and ie tetrachloride, employed for breaking dormancy in F. silvatica. | Tratau, H. Late-Tertiary Fagus species of Europe. (In German; English summary). Bot. Not. 115: 147-176. 1962. [Includes maps. | Trotter, A. The physiognomy of trees and the plasticity of the beech (F agus). (In Ttalian; English summary.) Monti Boschi 4: 339-348. 1953. Warp, R. T. The beech forests of Wisconsin; their phytosociology a relation- ships to forests of the vii without beech. Ecology 39: 444-457. 1958. [F. grandifolia; includes m . Some aspects of the inde habits of the American beech. /bid. 42: 828-832. 1961. [F. grandifolia; includes map.] Wits, A. B. The composition and dynamics of a beech maple climax com- munity. Ecol. Monogr. 6: 319-408. figs. 1-16. 1936. 1971] ELIAS, GENERA OF FAGACEAE 173 Wricut, J. W., R. T. BrncHam, & K. W. Dorman. Genetic variation within geographic ecotypes of forest trees and its role in tree improvement. Jour. Forestry 56: 803-808. 1958. [F. grandifolia.] Subfamily CASTANEOIDEAE Oersted 2. Castanea Miller, Gard. Dict. Abridg. ed. 4. 1: ord. alph. 1754. Large to small, erect to spreading deciduous trees or shrubs of the temperate Northern Hemisphere, with large taproots and furrowed, brown- ish bark; branchlets terete, the terminal buds covered by 2 pairs of slightly imbricate scales, the outer pair lateral, the inner accrescent, becoming oblong-ovate and leaving narrow ringlike scars upon falling. Leaves con- volute in vernation, elliptic to oblong-elliptic, the veins terminating at the tips of the usually coarse serrations, deciduous, petiolate; stipules paired, ovate to linear-lanceolate, acute at apex, scarious to herbaceous, infolding the leaf in bud, usually caducous; leaf-scars semioval, slightly elevated, with an irregular marginal row of small vascular bundle scars. Flowers in + erect catkin-like staminate or androgynous racemes or spikes (false panicles fide Camus) ; staminate inflorescences -—+ erect, vernal (appearing with the unfolding of the leaves), borne in the axils of the lower leaves of the year; androgynous inflorescences usually erect, shorter than the staminate, borne in the axils of later leaves. Staminate flower subtended by a minute ovate bract; perianth deeply divided into 6 ovate lobes, the lobes imbricate in bud, slightly puberulous, pale stramineous, with 6-18 (-20) exserted stamens, filaments filiform, incurved in bud, free; anthers small, ovoid to subglobose, dorsifixed, introrse, 2-locular, with longitudinal dehiscence, pollen usually 3-colporoidate, prolate to subprolate; rudimen- tary ovary occasionally present. Carpellate flowers at base of androgynous inflorescences, sessile, solitary or 2 or 3 within an involucre of closely imbricate, pubescent to tomentose scales and + branched spines; involucre subtended by a bract with two lateral bractlets; perianth urceolate, with 6 short obtuse lobes, pubescent, with 6-8 rudimentary stamens; styles usually 6 (—9), linear, spreading, pubescent below the middle, tipped by the minute punctiform stigmas; ovary incompletely (3—)6(—9)-loculate, each locule with 2 axile, collateral, semianatropous, and descending ovules. Involucre at maturity of fruit subspherical to oblong, pubescent and spiny on the outer surface, with elongated, ridged, branched spines fasciculate between the deciduous scales, inclosing 1-3 (rarely more) fruits, and splitting into 2-4 valves at maturity, the inner surface of the involucre with a lustrous pubescence, Fruits maturing in one season, + ovoid and variously compressed, acute apically and beaked by the persistent peri- anth and styles, usually bright chestnut-brown and lustrous, with a large basal scar. Seed usually solitary, the abortive ovules borne apically ; embryo straight, the cotyledons thick, fleshy, + undulate and ruminate, Inclosing remnants of endosperm between the folds, often sweet; germina- tion hypogeal. Lectotype SPECIES: Fagus Castanea L. = C. sativa Mill.; See Britton, N. Am. Trees 271. 1908. (The ancient Latin name of the 174 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 ¥ We We NA wy 2 Tine Wwe ned Ve a Re a (\ Vo LL, Y,| fe s KS bi) NI _ FicuRE 3. Castanea. a~j, C. pumila: a, fruiting branchlet, X ¥%; b, staminate inflorescence, X 1; c, androgynous inflorescence » X 1’; d, sta minate flower, X 10; e, carpellate flo ower, X 6; f, carpellate flower in semidia grammatic longitudi- nal section showing two of the six a ~*~ 6: g, ovary in semidiagrammatic ate legion X_ 10; h, involucre with fruit, x 1; i, hada ith aborted ovules apex, x 2; j, seedling, showing cal taproot, x 1971] ELIAS, GENERA OF FAGACEAE 175 European chestnut, derived from its Greek name, kastanea or kastanos.) — CHESTNUT. About 12-14 species, widely distributed in eastern North America, southern Europe, northern Africa, southwestern Asia, central and northern China, Korea, and Japan. Characterized by having 6(—9) styles, the ovary 6-9-locular, the carpellate flowers borne at the base of androgynous cat- kins, the fruits maturing in one season, and the leaves deciduous, Castanea is closely related to the Old World Castanopsis (D. Don) Spach, which has three styles, the ovary three-locular, the carpellate flowers on separate catkins, the fruits maturing in two years, and the leaves persistent. Dode (1908) divided Castanea into three sections: CASTANEA (§ Eucas- tanon Dode, BALANOCASTANON, and HypocaAsTANon, based on the num- ber of fruits per involucre, the presence or absence of a persistent beak on the fruits, and their shape. Section HypocAsTANoNn includes only C, Henryi Rehd., of China. Section CAsTANEA, including about six species with C. dentata ( Marsh.) Borkh., 2n = 24, American chestnut, in our range, demonstrates the dis- junction shown by a number of relict Arcto-Tertiary genera between the eastern United States (C. dentata), southern Europe (C. sativa Mill., nm = 24), and southeastern Asia (C. crenata Sieb. & Zucc., 2n = 24 mollissima Blume, 2n = 24, C. Seguinii Dode, and C. Davidii Dode). Section CASTANEA is characterized by involucres each containing three fruits, rarely more, by the middle fruit being as wide as or wider than long, and by the fruits usually being conspicuously beaked. Castanea dentata, formerly a dominant large tree in the eastern decidu- ous forest, ranges west from central Maine to New York, extreme southern Ontario, and southeastern Michigan, south to Ohio, southern Indiana, southern Illinois, western Kentucky, central Tennessee, northern Missis- sippi, Alabama, Georgia and northwestern Florida. Easily recognized by the large, oblong-lanceolate leaves which are coarsely serrate and by the very characteristic fruits, C. dentata can still be found throughout most of its original range in the form of stump sprouts that have been produced persistently for many years in spite of the attacks of the chestnut blight. The sprouts rarely reach flowering age, for usually when about 8-12 years old, they contract the blight and soon die (Neilstaedt, 1953). A few fruit- ing trees still survive in the wild, but most are in cultivation in areas isolated from the natural range of the species (cf. Thompson). The natural replacement of Castanea dentata in the oak-chestnut forest of the eastern United States has been studied by several workers (cf. Keever, 1953; Nelson, 1955; Woods & Shanks, 1957, 1959). Replace- ment is mainly by species that were associated with C. dentata during its dominance. The main replacement species are Q. rubra L., Q. Prinus L., Q. alba L., Acer rubrum L., and Liriodendron tulipifera L. Although other tree species, many of which were subordinate to the dominant species of Quercus and C. dentata, can be found, this association is apparently be- coming an oak-oak association. 176 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Section BALANOCASTANON Dode, restricted to the eastern United States, has one (rarely two) fruit per cupule, the fruit longer than broad and usually beaked. The taxa of this section need careful study to determine their exact status. Castanea ozarkensis Ashe, Ozark chinquapin, a tree with gray, glabrous branchlets, leaves broadly lanceolate to oblong, coarse- ly toothed, minutely pubescent to glabrate beneath, the involucral spines pubescent, and the seed not flattened, is found in woods and on rocky slopes from Mississippi and Louisiana to Arkansas, Oklahoma, and south- ern Missouri. The closely related C. pumila Mill. var. pumila, 2n = 24, Allegheny chinquapin, occurs in dry woods and thickets from New Jersey and eastern Pennsylvania, southwest to Tennessee, Arkansas, and eastern Texas, and east to Florida. The generally shrubby C. pumila can be dis- tinguished from C. ozarkensis by the pubescent branchlets, the leaves oblong, serrate, whitish-downy beneath, the involucral spines pubescent and much longer than in C. ozarkensis. Varietas Ashei Sudw. (C. Ashei (Sudw.) Sudw.) found mainly on the Coastal Plain from northern Florida northward to southeastern Virginia, and westward to eastern Texas and Arkansas, has the clusters of involucral spines remotely arranged, leaving conspicuous open areas on the involucre. A third species, C. alnifolia Nutt. var. alnifolia, trailing chinquapin, found in Georgia, Florida, Ala- bama, and Mississippi, typically occurs in dry, sandy soils. Frequently spreading by underground rootstocks, var. alnifolia is a low, often trailing shrub. Varietas floridana Sarg. (C. floridana (Sarg.) Ashe, C. Margaretta (Ashe) Ashe), Florida chinquapin, is a small tree of upland woods, flat- woods, and hammocks in the Coastal Plain from North Carolina to north- ern Florida and westward to southeastern Louisiana. Kurz & Godfrey (1962) consider both this variety and C. pumila var. Ashei to be distinct species. Castanea paucispina Ashe, a poorly known plant described from eastern Texas (Newton County) and also attributed (fide Camus) to adjoining western Louisiana, is closely related to C. alnifolia but is sepa- rated from it by having a shorter, pubescent fruiting axis and the spines sparsely arranged on the involucre. Castanea < alabamensis Ashe, a hy- brid of C. dentata and C. alnifolia var. floridana, occurs only in northern Florida. Another hybrid, C. x neglecta Dode (C. dentata X C. pumila) is of questionable status. Although the species of Castanea are monoecious, cross pollination ap- parently is necessary for the production of viable seed. Solitary trees are frequently sterile, although the involucres may develop. A plausible ex- planation (which someone, hopefully, will prove or disprove) is that pollination may stimulate (chemically) the development of the involucre and ovary wall but that fertilization occurs only if the gametes are from another plant. Cross pollination is promoted by the staminate catkins maturing earlier than the androgynous catkins of the same plant. Clapper (1954) found that C. mollissima and C. crenata are self-sterile, but arti- ficial hybrids between them show some self-compatibility. Self-sterility in C. mollissima was determined by McKay (1942) to be due to in- compatibility, the pollen tube growing normally in cases of self-pollination, 1971] ELIAS, GENERA OF FAGACEAE 177 but growth ceasing as the tube approached the egg, and no fertilization taking place. The two to three per cent fruit set was attributed to apomixis. In a series of experiments to determine whether Castanea species are wind pollinated or insect pollinated, as suggested by some workers because of the rather conspicuous staminate spikes and the slightly odorous flowers, Clapper (1954) found them to be largely wind pollinated and concluded that this is more efficient than insect pollination and that insects are not essential for pollination. The stigmas were found to be receptive a few days after anthesis and most receptive 10-12 days after the beginning of anthesis. He also concluded that pollination was more effective in the morning than in the afternoon, apparently because of differences in hu- midity. From a series of interspecific crosses, it was determined that early flowering is dominant over late flowering. Chestnut blight, caused by the ascomycete Endothia parasitica (Murr.) Anderson & Anderson, was first discovered in 1904 in the New York Zoo- logical Park. Within 50 years it swept throughout the natural range of Castanea dentata, the species most susceptible to it. Castanea sativa is also proving to have little resistance against infection, but species of sect. BALANOCASTANON, as well as several species of Quercus, Acer, Carya, and Rhus, are susceptible in varying lesser degrees. The Asiatic species of Castanea show considerable resistance, but all serve as host to the blight ungus. Following infection through a wound in the trunk or branches, the spores of Endothia germinate, and the mycelium rapidly spreads through the inner bark and cambium. Upon sporulation, a girdling canker (either appearing swollen or sunken) is usually formed, resulting in the death of the affected branch or the entire tree. Two types of spores are produced: relatively long-lived, one-celled, sticky, conidiospores produced during moist conditions, and larger, short-lived, two-celled ascospores. Conidio- spores are spread by birds, insects, and to a lesser extent by rain, while ascospores are wind dispersed. A very extensive breeding program was started soon after the devastat- ing results of the chestnut blight became obvious. Breeding efforts were centered around Castanea crenata, the Japanese chestnut, and C. mollis- sima, the Chinese chestnut. Several disease-resistant hybrids have re- sulted, some suitable for commercial nut growing, others for ornamental trees. Clapper (1954) and Graves (1960, 1962) have presented techniques and results of some of the major breeding programs. Castanea dentata was formerly an important timber tree in the eastern United States. The wood, brown, soft, light, open-grained, very durable, and easy to work, was widely used both indoors and in outside construc- tion. The wood of Castanea species is high in tannin and was the major tanning agent for leather. Commercial chestnut growing in the United States has had to rely on Asiatic species and hybrids since the virtual elimination of the American chestnut. 178 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 REFERENCES: See also under family references BERRY 140~145; ForMAN; Kurz & Goprrey, 57-59; HJELMQvist; HutcHinson; LANcpon; SARGENT; VINES, 191-202; and West & ARNOLD, 33. ArrETINI, C. Bio-morphological aspects of flowering in young chestnut plants (Castanea sativa Mill.). (In Italian; English summary.) Monti Boschi 8: 323-331. 1957.* Beatriz, R. K., & J. D. Dimer. Fifty years of chestnut blight in America. Jour. Forestry 52: 323-329, 1954, Berry, F. H. Relative resistance of some chestnut species and hybrids inoculated with the blight fungus. North. Nut Growers Assoc. Rep. 51: 46, 47. 1960. [Endothia parasitica. ] Brooks, A. B. Castanea dentata. Castanea 2: 61-67. 1937. [Range of C. dentata; also mentions introduced spp. | Burterick, P. L. The chestnut in North Carolina. Econ. Paper North Carolina Geol. Econ. Surv. 56: 7-10. pis. 452.1925. Camus, A. Les chataigniers. 2 vols.: Texte, 604 pp. 28 jigs.; Planches, 76 pls. + XXXIV pls. Paul Lechevalier, Paris. 1929. [Monograph of Castanea & Castanopsis. | Crapper, R. Chestnut breeding, techniques and results. I. Breeding material and pollination techniques. Jour. Hered. 45: 106-114, 1954; II. Inheritance of characters, breeding for vigor, and mutations. Ibid. 201-208. Dermen, H., & J. D. Ditter. Colchiploidy of chestnuts. Forest Sci. 8: 43-50. 1962. [C. dentata, C. mollissima, & C. crenata. | Ditter, J. D. Is our American chestnut developing blight resistance? North. Nut Growers Assoc. Rep. 47: 29-31. 1956. ~———-. The present status of screening the American chestnut for blight resis- tance. bid. 51: 47-50. 1960.* - Chestnut blight. U. S$. Dep. Agr. Forest Pest Leafl. 94. 7 pp. 1965. [Excellent account of chestnut blight. ] Gravatt, G. F. The chestnut situation in Europe. North. Nut Growers Assoc. Rep. 49: 67, 68. 1958.* Graves, A. H. Relative blight resistance in species and hybrids of Castanea. Phytopathology 40: 1125-1131. 1950. . Hastening flowering in a chestnut seedling. North. Nut Growers Assoc. Rep. 49: 90-94. 1958.* [C. Henryi.] - Some outstanding new chestnut hybrids. I. Bull. Torrey Bot. Club 87: 192-204. 1960; IT. Jbid. 89: 161-1 72. 1962. Graves, H. S. Notes on the rate of growth of red cedar, red oak, and chestnut. Forestry Quart. 3: 349-353. 1905. Hoitm, T. Medicinal plants of North America. 88. Castanea dentata (Marsh.) Borkh. and C. pumila (L.) Mill. Merck’s Rep. 24: 85-87. 1915. It11ck, J. S. Replacement of the chestnut. Jour. Forestry 19: 105-114. 1921. i R. A. Chestnut chromosomes. Forest Sci. 8: 372-377. 1962.* EVER oe) 7 is:] wm oO =] =| ror} ° B ss) °o B, =a ° =] ° th wm j=) 5 La] nm ¢ as) ==) OQ. wn ° Leas] cf = 19°) o “ io] & ~ oO Ss © wn oo 5 oc =" 1971] ELIAS, GENERA OF FAGACEAE 179 Mattoon, W. R. The origin and early development of chestnut sprouts. Fores- try Quart. 7: 34-37. 1909. McCormick, J. F., & R. B. Puarr. A half century of succession in an Appala- chian forest following the chestnut blight. (Abstr.) ASB Bull. 17: 54. 1970. [Replacement of C. dentata by various genera, especially Carya.] McKay, J. W. Self-sterility in the Chinese chestnut (Castanea mollissima). Proc. Am. Soc. Hort. Sci. 41: 156-160. 1942. . Heterosis in chestnuts. (Abstr.) Genetics 33: 116. 1948. [C. mollissima x C. crenata. . Seed and seedling characters as tools in speeding up chestnut breeding. Proc. Am. Soc. Hort. Sci. 75: 322-325. 1960. — & F. H. Berry. Introduction and distribution of Chinese chestnuts in the United States. North. Nut Growers Assoc. Rep. 51: 31-36. 1960.* [C. mollissima. Murriti, W. A. A serious chestnut disease. Jour. N. Y. Bot. Gard. 7: 143-153. 1906. [First report of chestnut blight in U. S.] Nernstaept, H. Tannin as a factor in the resistance of chestnut, Castanea spp., to the chestnut blight fungus, Endothia parasitica (Murr.) A. & A. Phyto- pathology 43: 32-38. 1953. [C. dentata, C. crenata, & C. mollissima.] Netson, R. M. Growth and mortality of chestnut sprouts. Jour. Forestry 30: 872, 873. 1932. NELson, T. C. Chestnut [Castanea dentata] replacement in the southern high- lands. Ecology 36: 352, 353. 1955. [C. dentata in North Carolina. | Ricuarps, E. C. M. A study of reforested chestnut cut-over land. Jour. Fores- try 15: 609-614. 1917. SOLIcNAT, G. Observations sur la biologie du chataignier. Ann. Amel. Pl. 8: 31-58. 1958.* Stout, A. B. Why are chestnuts self-fruitless? Jour. N. Y. Bot. Gard. 27: 154— 158. 1926. THompson, P. W. A unique American chestnut grove. Mich. Academician 1: 175-178. 1969. [Outside natural range of C. dentata, near Frankfort, Michigan. VILKOMERSON, H. Flowering habits of the chestnut. North. Nut Growers Assoc. Rep. 31: 114-116. 1941.* Woops, F. W., & R. E. SHanxs. Replacement of chestnut in the Great Smoky Mountains of Tennessee and North Carolina. Jour. Forestry 55: 847. 1957. [C. dentata. | ails & . Natural replacement of chestnut by other species in the Great Smoky Mountains National Park. Ecology 40: 349-361. 1959. [Cc. dentata. | Subfamily QUERCOIDEAE Oersted 3. Quercus Linnaeus, Sp. Pl. 2: 994. 1753; Gen. Pl. ed. 5. 431. 1754. Evergreen or deciduous trees or shrubs of varied habitats, often with astringent properties, the sap watery, the bark pale and scaly or dark and furrowed, the wood usually hard and close-grained or brittle and either diffuse or ring porous with both uniseriate and multiseriate rays, the pith star-shaped in section and continuous; branchlets slender, lenticellate, often fluted or + prominently 5-angled; winter buds clustered at end of branch- lets, the scales chestnut-brown, closely imbricate in 5 ranks, slightly ac- 180 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 crescent, caducous, leaving ring-like scars upon falling. Leaves usually with involute, convolute or revolute vernation, deciduous or persistent until spring or persistent for 2-4 years, simple, usually distinctly petiolate, the blade entire, variously lobed or dentate, often variable on the same branch, membranaceous to coriaceous, pinnately veined, the prominent primary veins extending to the margins or beyond as bristles, or veins anastomosing within; stipules obovate to lanceolate, scarious, often caducous, or those of the upper leaves occasionally persisting to the next year; leaf-scars slightly elevated, semiorbicular, usually broader than high. Flowers vernal, with or slightly after unfolding of leaves; staminate inflorescences in pendulous, fasciculate aments from axils of leaves of the previous year, from axils of the inner scales of the terminal bud, or from axils of leaves of the current year. Staminate flowers solitary or in clusters of 2 or 3, bracts absent or, if present, caducous; perianth campanulate, 3—6-lobed, usually divided to near the base; stamens (2—)4—12, inserted on a slightly thickened torus, the filaments free, filiform, exserted, the anthers usually ovate-oblong to subglobose, glabrous to rarely pubescent, 2-loculate, ex- trorse, pollen 3-colporate or 3-colporoidate, usually subprolate to prolate. Carpellate flowers solitary or in erect few-flowered spikes from the axils of the leaves of the current year, each flower inclosed by an involucre of many imbricate [or annular| scales, the involucre subtended by a caducous bract and 2 bractlets and later becoming the cupule of the fruit; perianth urceolate, with 6 short lobes, the perianth tube adnate to the incompletely 3(—4—5)-loculate ovary; styles usually 3, short or elongate, erect or incurved, often dilated above, stigmatic on the inner face or at the apex, usually persistent in fruit, stamens wanting [or abortive stamens as many as the lobes]; ovules 2 per locule, collateral, anatropous or semi- anatropous, descending. Fruit a nut (acorn) maturing in one or two grow- Ing seasons, ovoid, subglobose, or turbinate, usually short-pointed at the apex, generally 1-seeded by abortion, marked at the base by a large con- spicuous circular scar; pericarp crustaceous or coriaceous, glabrous or with a pale tomentum on the inner surface, partially to completely inclosed in the accrescent cupule (involucre) ; cupule scales loosely or closely imbri- cate, flattened or variously thickened [rarely united into concentric rings], occasionally with a fringed margin. Seed with the aborted ovules at the base or at the apex, rarely lateral; cotyledons thick, fleshy, usually plano- convex and entire; often undulate on the back; radicle superior, included within the base of the cotyledons, hilum minute, basal or apical, germina- tion hypogeal. x = 12. (Quercus Raf., Cerris Raf., Scolodrys Raf., Macrobalanus (Oersted) Schwarz, Erythrobalanus (Spach) Schwarz.) LecTotyPE species: Q. Robur L.; see Britton, N. Am. Trees 278. 1908. (Classical Latin name for oak.) — Oax. A genus of approximately 500-600 species in the mild-temperate regions of the Northern Hemisphere and in higher altitudes of the tropics, in the New World distributed southward to Mexico, Central America, and Colombia, and in the West Indies. In the Old World the genus is dis- 1971] ELIAS, GENERA OF FAGACEAE 181 tributed from Manchuria, China, and Japan, southward to the East Indies and the Philippine Islands, and to southern India, westward along the Himalayas to western Asia, and through the Mediterranean region into western and northern Europe. It is absent from central and southern Africa, New Guinea, Australia, and the Pacific Islands. Approximately 30 species occur in the southeastern United States. Although numerous infrageneric names of various ranks have been pro- posed for the American oaks by several workers, Trelease (1924) in his comprehensive treatment arranged them in three sections and more than 100 series. Three subgenera are recognized here: ERYTHROBALANUS (Spach) Oersted, restricted to the New World; Quercus (subg. Lepid- obalanus (Endl.) Oersted) of both Old and New Worlds; and Cyctosat- ANOPSIS (Oersted) Schneider, of eastern Asia. The differences between the two subgenera in North America are summarized in the following key. Leaves entire or lobed, the lobes rounded or, if toothed, only mucronate-tipped, never aristate-tipped; styles short, abruptly flaring; acorns annual, cup scales usually prominently thickened basally and loosely appressed apically, acorn wall (pericarp) glabrous within; aborted ovules generally basal; bark usually soft gray and scaly, rarely black and deeply furrowed; wood pale, the larger vessels often plugged by tyloses, the smaller summer vessels thin-walled and WOR ea cs ur wt a ae eee ee we ee ee subg. QUERCUS. Leaves entire or lobed, the lobes not rounded and, when toothed, aristate-tipped or, if entire, aristate-tipped from the veins; styles long, gradually flaring; in; aborted ovules usually apical; bark rather hard, dark and furrowed, scarcely scaly; wood usually reddish, the vessels seldom filled with tyloses, the smaller summer vessels always thick-walled and rounded. .. subg. ERYTHROBALANUS. Subgenus Quercus (Lepidobalanus Endl. ex Oersted; sect. Lepidobal- anus (Oersted) Sargent; subg. Macrobalanus Oersted; Macrobalanus (Oersted) Schwarz; sect. Leucobalanus Engelmann; subg. Leucobalanus (Engelm.) Trelease) is represented in the eastern United States by ap- proximately 16 species, many of which are confined to the Gulf Coastal Plain. The 57 series of subg. QuERcUs proposed by Trelease do not con- tribute significantly to an understanding of relationships and are not con- sidered here, for the sections recognized by Rehder (1949) seem to rep- resent more natural groups in this subgenus. Sections Cerris Dumort., Super [Reichenb.] Spach, GALtiFERA Spach, and QuERCUS ($ Robur Dumort.) are confined to the Old World; our species are members of sects. ILEx and PrINUS. : Section Inex Loudon, characterized by the evergreen habit; leaves usually entire, coriaceous; styles short, dilated toward the rounded apex; involucre with appressed scales; and fruit usually annual, is disjunctly distributed between the southeastern United States, southern Europe (Quercus Ilex L.), and southeastern Asia (Q. phillyraeoides Gray). Nearly restricted to the Atlantic and Gulf coastal plains, Q. virginiana Mill. ex- tends from southeastern Virginia to southern Florida and western Cuba, 182 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 ge , staminate flower prior : , e flower at anthesis, X 10; c, stamen showin longitudinal dehiscence, X 20; d, carpellate flower, X 15. e-g, Q. alba: e, fruit- ing branchlet, xX Bs; f, mature acorn, X 1%; g, mature acorn in semidiagram- matic longitudinal section showing basally aborted ovules, X 114. h, i, Q. macro- ¢arpa: h, leaf, X %4; i, mature acorn, X 1. j, k, Q. Muhlenbergii: j, leaf, X 13 1971] ELIAS, GENERA OF FAGACEAE 183 and west to central and southern Texas and northeastern Mexico. Three varieties are recognized: var. virginiana (Q. minima (Sarg.) Small, Q. pygmaea (Sarg.) Ashe), var. fusiformis (Small) Sarg. (Q. fusiformis Small), and var. maritima (Michx.) Sarg. (Q. maritima (Michx.) Willd., Q. geminata Small, Q. Rolfsii Small, QO. succulenta Small). Section Prinus Loudon, with some 14 species in the United States, is distinguished by the deciduous, variously lobed to dentate leaves; the short styles, broad at the apex; the involucral scales appressed or rarely the upper free and subulate (Q. macrocarpa); and the annual fruit. The species that occur in our range include Q. alba L.; Q. stellata Wang., with var. stellata (QO. Ashei Sterrett, Q. similis Ashe), var. Margaretta (Small) Sarg. (Q. Margaretta Ashe ex Small, Q. Boyntoni Beadle), and var. missis- sippiensis (Ashe) Little (Q. mississippiensis Ashe); Q. lyrata Walt.; Q. macrocarpa Michx.; Q. bicolor Willd. (Q. platanoides (Lam.) Sudw.); Q. oglethorpensis Duncan; Q. Chapmanii Sarg.; Q. Durandii Buckl. (Q. austrina Small, Q. sinuata Walt.); Q. Prinus L. (Q. montana Willd.) ; Q. prinoides Willd.; and Q. Muehlenbergii Engelm. (Q. Castanea Muhl., Q. acuminata (Michx.) Sarg., Q. Alexanderi Britton, Q. Brayi Small). Approximately five species belonging here are found in the western United States, and four (perhaps more) species occur in eastern Asia. Subgenus Quercus has fewer species (ca. 19) in Central America than subg. ERYTHROBALANUS. In both subgenera the Mexican species are poorly known, poorly represented in herbaria, and in need of careful study. Subgenus ERYTHROBALANUS (Spach) Oersted (§ Erythrobalanus Spach, Erythrobalanus (Spach) Schwarz, § Melanobalanus Engelmann) is con- fined to the New World, mainly to the southern and eastern parts of the United States, Mexico and Central America. Trelease (1924) recognized 190 species (plus numerous subspecific taxa), which he distributed among 73 series, many of these with only one to three species! Muller (1942), in a taxonomic survey of the Central American species of Quercus, recog- nized many of Trelease’s series, but he reduced to synonymy more than he recognized and, in addition, reduced many of the species described by Trelease. Since the series seem to be largely artificial, no attempt is made here to account for them. Muller concluded that the difficulties in delimit- ing subgeneric categories in subg. ERYTE aNus lie in the poorly known Mexican species. For convenience, the species discussed here have been grouped mainly on the basis of leaf characters that do not necessarily reflect natural relationships. These relationships need to be worked out carefully. Quercus chrysolepis Liebm., of southwestern Oregon south along the coastal ranges and the Sierra Nevada to northern Lower California and western Nevada, and Q. tomentella Engelm., of the California Channel Islands and Guadalupe Island off Baja California, are somewhat inter- mediate between subg. ERYTHROBALANUS and subg. Quercus. Both have k, mature acorn, X 1%. 1, m, Q. virginiana: 1, leaf, X 1; m, mature acorn, X 1%. 184 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 IGURE 5. Quercus subgenus Erythrobalanus, a-e ilicifolia: a, branchlet with staminate inflorescences, X 1; b, staminate boner, : 12; c, carpe ellate flower, , first year receptive carpellate pst in semidiagr ammatic longitudina” section showing absence of ovules, x 20; mature x 1%. i, rubra: f£, second year carpellate flower in seinidinerapatuntie: longitudinal section 1971] ELIAS, GENERA OF FAGACEAE 185 entire persistent leaves that are involute in bud, usually 6—8 stamens, and abortive ovules that are basal or lateral. Although Q. chrysolepis was placed by Rehder in subg. QuERcUsS, and some of its characters are in- termediate between the subgenera, both species seem best referred to ERYTHROBALANUsS because of the silky tomentum on the inner surface of the acorn wall and the biennial fruits. Four other species, including Quercus myrtifolia Willd., of the south- eastern Coastal Plain (southern South Carolina to southern Florida, west to southern Mississippi), have similar leaf characters but differ (as do the remaining species of the subgenus) in having 4—6 stamens per flower, styles finally recurved, and the abortive ovules apical. Almost grading into the above group are five species of the southeastern United States, commonly known as the willow oaks, which have mainly entire, but deciduous leaves. These are Q. Phellos L., Q. incana Bartr. (Q. cinerea Michx., Q. brevifolia (Lam.) Sarg.), Q. imbricaria Michx., and Q. pumila Walt. Formerly thought to be a distinct species, 0. < Jlaurifolia Michx. (Q. hemisphaerica Bartr., O. hybrida (Chapm.) Small, Q. obtusa (Willd.) Ashe, Q. rhombica Sarg.) has recently been considered by Burk (1963) to be the hybrid of Q. Phellos and Q. nigra. The other species of subg. ERYTHROBALANUS have deciduous, variously lobed leaves (except some Quercus nigra L. and Q. ilicifolia Wang.) that are convolute in bud. An almost complete gradation in lobing can be found from one species to another, from an entire leaf to a slightly three to five lobed (apically) leaf, to those with five shallow lobes, and even- tually to those with five to nine to eleven deep lobes. Entire to three lobed leaves can be found in Q. nigra L. (Q. microcarya Small) and Q. ilicifolia Wang. (Q. nana (Marsh.) Sarg.), while slightly three to five lobed, usually obovate leaves are generally evident in Q. marilandica Muenchh. (Q. neo-ashei Bush), Q. arkansana Sarg., Q. laevis Walt. (Q. Catesbaei Michx.), and Q. georgiana Curtis. The leaves of Quercus falcata Michx., including var. falcata (Q. triloba Michx., Q. elongata Willd., Q. digitata (Marsh.) Sudw.) and var. pagodifolia Ell. (Q. Pagoda Raf., Q. pagodifolia (Ell.) Ashe, Q. leucophylla (Ashe) Ashe) show a condition that continues to the more deeply lobed leaves of Q. coccinea Muenchh. ; Q. velutina Lam. (Q. missouriensis (Sarg.) Ashe); Q. palustris Muenchh. ; Q. ellipsoidalis Hill; Q. Shumardii Buckl., with var. Shumardi (Q. Schneckii Britton) and var. texana (Buckl.) Ashe (Q. texana Buckl.) 3 Q. Nuttallii Palmer; and Q. rubra L. (Q. borealis Michx. f., Q. maxima (Marsh) Ashe). In the United States, the entire-leaved species of Quer- cus occur mainly in the southern United States, and there appears to be a trend toward the more lobed condition in the northern portions of the range of the genus. : : : Muller (1942) recognized 27 species as belonging to this subgenus in showing ovule formation, X 8; g, second year carpellate flower in semidiagram- matic cross section, X 25; h, mature embryo showing apically aborted ovules, X 1%. i, j, Q. imbricaria: i, mature leaf, X 72; j, mature acorn, X 1 2. k-n, Q. nigra: k, 1, m, variation in mature leaves, X ¥%- n, mature acorn, X 1%. 186 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Central America (from Chiapas, Mexico, through Panama). Quercus Humboldti Bonpl., the southernmost of the American species, is found from Panama southward into the Colombian Andes. The taxonomy of the species of subg. ERYTHROBALANUS in Mexico is very poorly known, per- haps due to the paucity of fruiting collections. Hybridization among species of Quercus has been known and discussed for many years. Numerous hybrids have been described, but only rela- tively recently has the critical study of the true nature of the putative hybrids been begun. Natural, usually fertile, hybrids between recognized species of Quercus are common, but the total number of hybrid combina- tions is uncertain, for several species (e.g., Q. rubra) are polytypic, and some recognized forms may prove to be only variants of a polytypic species. Palmer (1948) listed what he considered to be 75 certain hybrids and approximately 15 dubious ones. Muller (1955) questioned the actual existence of so many hybrids, for many were based on minor variations that probably fall within the variability of the supposed parental species. Unfortunately, many hybrids described between 1900 and 1950 were based primarily on a few specimens with no attempt to provide quantita- tive data supporting the putative hybrids. It has been clearly shown that the species that are compatible and produce fertile hybrids are the more closely related ones. Although all species of Quercus counted have 2n = 24, natural hybridization between species of the subgenera QuERcUS and ERYTHROBALANUS is unknown and not expected to be found, although Pjatnitzky (1946) was able to cross experimentally Q. Robur with Q. borealis maxima [= Q. rubra] and Q. macranthera with Q. borealis maxima and produce vigorous seedlings. Hybrid individuals or populations are usually found in areas where the ranges of two species overlap or where they once overlapped but now do not. Muller (1953) demonstrated that Q. Havardii Rydb. (confined to deep sandy soils) and Q. Mohriana Buckl. (on limestone) will only come into contact where erosion of the sandy areas exposes the limestone. Hybrids are readily produced but are confined to soils composed of @ mixture of sand and limestone. Several other studies have also shown that hybrids are better suited to “hybrid” habitats than to the edaphic condi- tions of either parent. Hybrids between two allopatric species whose ranges now do not overlap, Q. macrocarpa Michx. and Q. Gambelii Nutt., were interpreted by Maze (1968) to have occurred during a period in the past when the two species were partially sympatric. Pleistocene migration was believed responsible for the sympatry. Some species of Quercus in our range are sympatric with others of the same subgenus but have not yet been shown to hybridize with them. Among these are Q. pumila, Q. myrtifolia, 0. Chapmanii, Q. Nuttallit, and Q. oglethorpensis. Whether the apparent absence of hybrids is due to insufficient study or whether these species are incompatible with others remains to be seen, but the former appears to be more likely. Flowering is vernal, the flowers appearing shortly before, with, or short- ly after the new leaves. The staminate inflorescences usually develop from 1971] ELIAS, GENERA OF FAGACEAE 187 axillary buds of leaves of the previous season. Development of the stam- inate inflorescences of Q. alba is initiated in early summer (late June); the inflorescence initials overwinter and then complete development the following spring. In contrast, the solitary or paired carpellate flowers are borne in the axils of the leaves of the current season. In Q. alba the carpellate flower primordia do not appear until late summer or early fall, overwinter, and mature the following spring as the terminal bud expands. The lightweight pollen, produced in enormous amounts and a major contributor to pollen allergies, is dispersed by wind. Sharp & Chisman (1961) reported that the mature anthers of Q. alba open or close, depend- ing upon the relative humidity, high humidity and rain often delaying or reducing pollen shedding and dispersal and, consequently, lowering acorn production. The staminate inflorescences are sensitive in varying degrees to freezing temperatures and prolonged dry winds. Nearly complete pollen dissemination requires as little as a few hours in some of the shrubby oaks and as long as several days in the larger tree species. Although not sufficiently studied, individuals of the various species of Quercus are cross- pollinated, for the plants are partially or wholly self-sterile. In subg. QuERcus, pollen-tube development and subsequent fertilization usually occur soon after pollination. In Q. alba, and presumably other species, pollination occurs before the ovules are completely developed. Syngamy may be delayed for a month or more in Q, alba (Turkel e¢ al., 1955) depending on the time of maturation of the ovule. Following fer- tilization, one of the six ovules (the other five abort, basally) develops along with the other parts of the carpellate flower into a mature acorn in one season. Pollination and pollen-tube development also occur in the spring in subg. ERYTHROBALANUS, but fertilization in most species does not occur until the following spring. At the time of pollination, the ovules have not differentiated, and the entire ovary is rather undeveloped. By late spring, the ovule and placenta develop meagerly, but most of the development takes place the next spring. Fertilization of one of the ovules (the other five abort, usually apically) occurs approximately a year after pollination and the acorn matures in the fall of the second year. Acorns of subg. Quercus generally require little or no dormancy and germinate soon after falling, if suitable environmental factors are present. The acorns of most species (temperate) of the subg. ERYTHROBALANUS appear to require a period of dormancy before germinating the following spring. Germination in both subgenera is hypogeal, the fleshy cotyledons remaining inside the pericarp. Polyembryonic acorns with two or very rarely three embryos per fruit, are known in Q. Prinus, Q. alba, Q. rubra, and Q. Robur and probably occur in many more species of Quercus. The embryogeny of Q. Robur and Q. velutina follows the onagrad type but with very irregular cleavage. The bitegmic ovule is crassinucellar. The formation of short branches in the pollen tubes of Q. Robur is noteworthy. Oaks are the most abundant trees over most of the eastern United States and in parts of the southern forest. The most extensive association is the oak-hickory (Quercus-Carya) forest found through much of the 188 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 Piedmont and Inner Coastal Plain of the Atlantic States, the Gulf Coastal Plain, and westward to eastern Texas and north throughout most of the Mississippi River Valley (Shelford, 1963). Various combinations of oak and hickory species occur throughout the association, the most frequent species being Q. alba L., Q. rubra L., Q. velutina Lam., Q. stellata Wang., Q. marilandica Muenchh., and Carya cordiformis (Wang.) Koch, C. ovata (Mill.) Koch, and C. laciniosa (Michx. f.) Loud. Braun (1950), restricting the oak-hickory forest to the Interior Highlands, centered mainly around the Ozark and Ouachita mountain region, preferred to label the remaining area as the oak-pine forest, emphasizing the sometimes transitory, some- times persistent nature of the pines in the southeastern United States. Bourdeau (1954) attempted to determine why certain species of Quercus are found in some oak-hickory associations but not in others. He found that Q. alba, Q. velutina, and Q. rubra, generally restricted to mesic upland areas, are eliminated from drier, more rocky areas because their seedlings are not very drought resistant. They are, however, more shade tolerant than Q. marilandica and Q. stellata, which are restricted to the thin, rocky soils with periods of drought and considerable exposure. Other associations include the Liriodendron-Quercus forest found in the northeastern United States, the maritime live-oak forest usually found in a narrow band along the coastline where dune areas have become rela- tively stable (cf. Bourdeau & Oosting), and the oak-chestnut forest of the eastern United States formerly found in our range throughout much of the Appalachian Mountains. Vegetative reproduction is of considerable ecological importance in semiarid regions, which are often unsuitable for seedling establishment (Muller, 1951). Clonal thickets are formed by rhizomatous branches in several species of Quercus (Q. Hinckleyi Muller, 0. Havardii Rydb., Q. pyrenaica Willd., Q. Ilex L., and Q. virginiana). The dwarf Q. pumila also spreads in this way. Muller considers the resulting increase in longev- = to be an important factor in the ability to compete with the more xeric orms. Far too many species of Quercus have been described from fossil leaf fragments. Apparently the first identifiable oaks are found in Upper Cretaceous deposits. The leaves of the earliest oaks were entire, toothed, or scarcely lobed. Since deeply lobed leaves comparable to those of many species of subg. ERYTHROBALANUS appear much later in the fossil record, this condition is probably derived in Quercus. The relationship between Quercus and the fossil genus Dryophyllum is clouded with misidentifica- tions and the lack of a comprehensive paleobotanical study of the two genera. Berry (1923) thought that Quercus probably diverged from an- cestral Dryophyllum stock at the end of the Lower Cretaceous or the beginning of the Upper Cretaceous. 1971] ELIAS, GENERA OF FAGACEAE 189 ward to the trunk and roots. Wilting is caused by the great production of tyloses that block the xylem vessels, but toxins are also produced and may be the ultimate cause of death. Large masses of hyphae can be found on the outer surface of the sapwood and on the inner surface of the bark. Asexual reproduction is by conidia and sexual by ascospores from perithe- cia. The disease is spread by root grafts of trees growing close together or by the asexual or sexual spores being carried by insects. Most, or perhaps all species of Quercus are susceptible, those of subg. ERYTHROBAL- ANUS apparently more so than species of subg. QuERcus. Insect damage often is a serious threat and has been shown by Nichols (1968) to be the main cause of mortality of Q. Prinus and Q. alba in Pennsylvania. Heart rot, frequent in trees following fire damage, often makes the wood useless for lumber. Quercus is extremely important in the hardwood lumber industry, mature oaks being among the larger hardwood trees and yielding a hard, tough, yet durable, resilient and elastic wood that is used in heavy construction, shipbuilding, and in better furniture. Of the two subgenera, QuERcus typically has harder, stronger and more durable wood than the species of ERYTHROBALANUS, in which the wood is softer, more porous, and more open-grained. The more important timber species of subg. QuERcUS in- clude Q. alba, Q. macrocarpa, Q. stellata, Q. montana, Q. Prinus, Q. Garry- ana, and Q. bicolor; those of subg. ERYTHROBALANUS include Q. borealis, Q. velutina, Q. coccinea, Q. palustris, Q. laevis, Q. phellos, Q. texana, and Q. imbricaria. Quercus virginiana reportedly has the hardest, strongest, and toughest wood of all oaks and is highly prized. As early as the 19th century the United States government was making concerted efforts to insure that adequate supplies of Q. virginiana would be available for naval stores. The decrease in shipbuilding demands, extensive cutting, the slow growing nature of the species, and the restricted range limit the usefulness of Q. virginiana as a lumber source. Acorns have been used as a food supplement and as forage for animals, especially hogs. Among the best species of Quercus utilized are Q. alba and Q. virginiana, the former species used by the American Indians in making flour. Species of subg. QuERCUs generally have a low tannin con- tent in the acorns and, consequently, are more readily edible than those of subg. ERYTHROBALANUS, which have to be leached to remove the high tannin content to make them suitable for human consumption. All acorns are high in carbohydrates. Quercus Suber L., native to the Mediterranean region of southern Europe and the source of commercial cork, has been introduced into the south- eastern United States, where cork growing has hardly been successful commercially, This species is also an excellent ornamental shade tree, as are a number of other oaks, including the Japanese Q. myrsinifolia Blume and Q. glauca Thunb., of subgen. CycLosaLaNopsis, both of which are evergreen trees grown to a limited extent in the warmer parts of our area. As with other members of subg. CyYCLOBALANOPSIS, the cupule with the scales connate in concentric rings is characteristic. 190 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 REFERENCES: Because of the large number of references dealing with Quercus, most popular and semipopular articles and papers dealing with pathology have been omitted. Under family references see BERRIDGE; BERRY, 129-140; FormAN, 1964, 1966; FowELls, 557-640; Hyetmovist; Hurcuinson; Kurz & Goprrey, 62-104; LANGDON: LITTLE, 311-361: MartTIN; REHDER; SARGENT: SEARS; VINES, 147- 198; West & ARNOLD, 35-53, and WopeHou USE ALEXANDER, T. R. Plant succession on Key Largo, Florida, involving Pinus caribaea and Quercus virginiana, Quart. Jour. Florida Acad. Sci. 16: 133- Attarp, H. A. The hybrid oak, xX Quercus Rudkini, at reroiies Virginia. Rhodora 44: 262-266. 1942. [Q. phellos * Q. marilandic ANDRESEN, J. W. Viviparity in white oak. Jour. Forestry 53; 252: 1955, He alba. | BartLett, H. H. Regression of X Quercus Deamii toward Quercus macrocarpa and Ouercus Muhlenbergii. Rhodora 53: 249-264. 1951. Benson, L., E. A. Puitures, & P. A. Witper, et al. mehr sorting of characters in a hybrid swarm. I. Direction of slop ur. ; 1017-1026. 1967. [Quercus Douglasii and Q. ‘ca bbietla subsp. californica. Effects of solar radiation. Includes map. | BERKELEY, E. E. Marcescent leaves of certain species of Quercus. Bot. Gaz. 92: 85- 93. 1931 BILLINGS, W. D. A bud and a si to the southeastern arborescent oaks. Jour. sag 34: 475, 476. BisHop, C. N., & W. H. Duncan, oe new oak from Georgia: its associates and habitat. Jour. Forestry 39: 730, 731. 1941. [Q. oglethorpensis Duncan.] Bourpeau, P. F. Oa seedling ecology determining lana of species in Piedmont oak-hickory forests. Ecol. Monogr. 24: 297-320. & H. J. Oostinc. The maritime live oak forest in North eee Ecol- J ogy 40: 148-152. 1959. [Q. virginiana Boyce, S. G. Quercus imbricaria in the lower Piedmont of North Carolina. Castanea 14: 52. 1949. Braun, E. L. Deciduous forests of eastern North America. xiv + 596 pp. 1950. [Especially chapters 6-8.] Brett, D. W. Fossil oak wood from the British Eocene. Palaeontology 3: 86- 92. pls. 21, 22. 1960. Burk, C. J. An evaluation of three hybrid-containing oak populations we Pa North Carolina Outer Banks. Jour. Elisha Mitchell Sci. Soc. 78: 1962. [Q. phellos, Q. nigra, and Q. laurifolia.| - The hybrid nature of Quercus laurifolia. Ibid. 79; 159-163. 1963. [Q. phellos X Q. nigra a. | : Camus, A. Les chénes; monographie du genre Quercus. Texte, 2 vols. 1934-39; Atlas, 3 vols. 1934-48. CARPENTER, I. W., & A. T. Guarp. Anatomy and morphology of the seedling roots of four species of the genus Quercus. Jour. Forestry 52: 269- madi 1954. [Q. alba, Q. bicolor, Q. rubra, Q. palustris. Cuisman, H. H. On the range of Shanard red oak, Jour. Forest ry 40: 1942. [Q. Shumardii; extension northward to include Suerte ally in Franklin County, Pa J 1971] ELIAS, GENERA OF FAGACEAE 191 Cooke, G. B. Cork and the cork tree. xii + 120 pp. New York. 1961, [Q. Suber ; extensive bibliography. ] CooPERRIDER, M. Introgressive hybridization between Quercus marilandica and Q. velutina in Iowa. Am. Jour. Bot. 44: 804-810. 1957. Cutter, D. F. Anatomy of vegetative organs of Trigonobalanus Forman (Faga- ceae). Kew Bull. 17: 401-409, 1964. DeELAvAN, C. C. The relation of the storage of the seeds of some of the oaks and hickories to their germination. Rep. Mich. Acad. Sci. 17: 161-163. 1916. [Q. alba, Q. macrocarpa, Q. bicolor, Q. rubra, & Q. velutina.] DJAVANCHIR-Kuotr, K. Les chénes de I’Iran. v + 221 pp. 152 pl. Thése. Univ. Montpellier. 1967. Downs, A. A. Low forking in white oak sprouts may be hereditary. Jour. Forestry 47: 736. 1949. [Q. alba. | DvurFFIELp, J. W. Chromosome counts in Quercus. Am. Jour. Bot, 27: 787, 788. 1940. [19 spp. all 2m = 24.] Duncan, W. H. A new species of oak from Georgia. Am. Midl. Nat. 24: ‘EAP 756. 1940. [Q. oglethorpensis, abundant in Wilkes, Elbert, & Oglethorpe counties, Georgia. Quercus oglethorpensis — range extension and phylogenetic relation- ships. Lloydia 13: 243-248, 1950. [Also found in McCormick, Edgefield, Saluda, & Greenwood counties, S. Carolina; phylogenetic consideration of 8 spp. ] - Phylogenetic relationships of eight white oaks. (Abstr.) Jour. Tenn. Acad. Sci. 25: 223. 1950. [Q. alba, Q. stellata, Q. Margaretta, Q. austrina, Q. breviloba, Q. Chapmanii, Q. Durandii, Q. oglethorpensis.| ———. Preliminary reports on the flora of Georgia. 2. Distribution of 87 trees. Am. Midl. Nat. 43: 742-761. 1950. [Quercus alba, Q. cinerea, Q. falcata, Q. laevis, QO. marilandica, Q. stellata, Q. stellata var. Margaretta; includes maps. DvaL, & CA key to species of oaks of eastern North America based on foliage and twig characters. Rhodora 38: 53-63. 1936. Eames, A. J. On the origin of the broad rays in Quercus. Bot. Gaz. 49: 161- 167. pl. 8, 9. 1910. [Q. rubra, Q. velutina, Q. coccinea, Q. ilicifolia, Q. alba, Q. bicolor, Q. prinoides, Q. lamarense (fossil), & Quercinium Knowl- tonii (fossil) .] of America. Quinn & Boden Co., Rahway, New Jersey. 1941. Fercus, C. L. Oak wilt—a serious menace to our oaks. Morris Arb. Bull. 12: 3-6. 1961. [Caused by Ceratocystis fagacearum. | : Forpe, M. B., & D. G. Faris. Effect of introgression on the serpentine endemism of Quercus durata. Evolution 16: 338-347. 1962. [Includes map. |] FriesNer, R. C. Chromosome numbers in ten species of Quercus, with some remarks on the contributions of cytology to taxonomy. Butler Univ. Bot. Stud. 1: 77-103. 1930. ; Harper, R. M. Quercus macrocarpa in Alabama. Jour. Elisha Mitchell Sci. Soc. 1942. Jackson, L. W. R. & C. H. Driver. Morphology of mycorrhizae on deciduous forest tree species. Castanea 34: 230-235. 1969. [Includes Q. alba, Q. rubra, & Q. montana. |] 192 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 JeweELL, F. F. Insect transmission of oak wilt (Endoconidiophora fagacearum), Phytopathology 46: 244-257. 1956. [Several species of Nitidulidae. ] Jounsson, H. Chromosome numbers of twin plants of Quercus Robur and Fagus silvatica. Hereditas 32: 469-472. 1946. Jones, E. W. Biological flora of the British Isles. Quercus L. Jour. Ecol. 47: 169-222. 1959. [Q. Robur, Q. petraea.] Keitu, W. M. Analysis of vegetative propagation in Quercus prinoides. Rho- dora 59: 306-308. 1957. Krajicek, J. E. Epicormic branching in even-aged, undisturbed white oak stands. Jour. Forestry 57: 372, 373. 1959. [Q. alba.] Kuntz, J. E. Oak wilt: its development, spread, and control. Proc. 9th Int. Bot. Congr. 2: 209. 1959. [Jm vol. 2: Abstracts & Résumés. ] Larsste, A. M., & C. D. Monx. Some live oak forests of northeastern Florida. Quart. Jour. Florida Acad. Sci. 24: 39-55. 1961. [Ecology of Quercus virginiana. | Lanopon, L. M. The ray system of Quercus alba. Bot. Gaz. 65: 313-323. 1918. —. Anatomy of seedling buds of Quercus. Ibid. 84: 87-199. pl. 7-9. 1927. [Q. alba & Q. rubra.] Ontogenetic and anatomical studies of the flower and fruit of the Fagaceae and Juglandaceae. Jbid. 101: 301-327. 1939. [Includes Quercus rubra. | McNamara, E. F. Sprouting capacity of scrub oak. Jour. Forestry 58: 563, 564. 1960. [Chiefly control of Q. ilicifolia. “ Mazg, J. Past hybridization between Quercus macrocarpa and Quercus Gambeli. Brittonia 20: 321-333. 1968. MELvILLE, R. A metrical study of leaf-shape in hybrids. I. Some hybrid oaks and their bearing on Turing’s theory of morphogenesis. Kew Bull. 14: 161-177. 1960. [F, & F, hybrids of Q. Cerris « Q. Suber. | : Monk, C. D., G. I. Curip, & S. A. NicHotson. Species diversity of a stratified oak-hickory community. Ecology 50: 468-470. 1969. Moore, G. T. White-oak acorns as food. Missouri Bot. Gard. Bull. 12: 32-33. pl. 3.1924. [Q. alba.] Mutter, C. H. The problem of genera and subgenera in the oaks. Chron. Bot. 7: 12-14. 1942a. . The Central American species of Quercus. U. S. Dep. Agr. Misc. Publ. 447. 216 pp. 1942b. ‘ new species of Quercus in Texas. Am. Midl. Nat. 28: 743-745. 1942c. [Q. Houstoniana. | ———. The oaks of Texas. Contr. Tex. Res. Found. Bot. 1: 21-323. 1951a. The significance of vegetative reproduction in Quercus. Madroito 11: 129-137. 1951b. “— Ecological control of hybridization in Quercus: a factor in the mechan- ism of evolution. Evolution 6: 147-161. 1952. [See also Proc. 7th Int. Bot. Congr. 1950: 284, 285. 1953.] . The origin of Quercus in Cuba. Revista Soc. Cuba. Bot. 12: 41-47- . Fagaceae. Ann. Missouri Bot. Gard. 47: 95-104. 1960. [Jm: R. 5 Woopson & R. W. Scuery, Flora of Panama. | . The live oaks of the series Virentes. Am. Midl. Nat. 65: 17-39. 19618 origin of Quercus fusiformis Small. Jour. Linn. Soc. Bot. 58: 1-12. 1961b. LQ. fusiformis arose from extensive introgression between 0. simana and Q. Brandegei in a once sympatric area. Includes map. ] 1971] ELIAS, GENERA OF FAGACEAE 193 Ness, H. Hybrids of the live _ ie overcup oak. Jour. Hered. 9: 263-268. 1918. [Q. virginiana X Q. ly: Nicuots, J. O. Oak mortality in asia: a ten-year study. Jour. Forestry 66: 681-694. 1 PALMER, E. J. The red oak complex in the United States. Am. Midl. Nat. 27: 732-740. 1942. [5 spp. and 6 vars. recognized. | . Quercus Durandii and its allies. 7bid. 33: 514-519. 1945. ———. Hybrid oaks of North America. Jour. Arnold Arb. 29: 1-48. 1948. PamMeEL, L. H., & C. M. Kino. The germination and juvenile forms of some oaks, Proc. Iowa Acad. Sci. 24: 367-398. 1917. [Nine spp., chiefly of the eastern U. S. Penrounp, W. T., & J. A. Howarp. A phytosociological study of an evergreen oak forest in the vicinity of New Orleans, Louisiana. Am. Midl. Nat. 23: 165-174. 1940. [Q. nigra, Q. virginiana. | PyatNitzky, S. S. Experimental production of pon Seeing in the genus Quercus. Dokl. Akad. Nauk SSSR 52: 343-345. New hybrid forms of oak, Quercus Komarovii and bh Timjarzevii produced experimentally. Ibid. 625, 626.* é Experim ent in crossing cork oak (Quercus Suber L.) with deciduous species. (In Russian.) Jbid. 87: 291, 292. 1952.* Rasxatov, P. B. Anatomical structure of an acorn. (In Russian.) Lesn. Khoz. 4(9): 75-79. 1951.* Roserts, E. G. A disjunct community of chestnut oak [Quercus Prinus| in Mississippi. Rhodora 63: 324-326. 1961. Row ee, W. A., & S. P. NicHoxs. The taxonomic value of the staminate flowers of some of the oaks. Bot. Gaz. 29: 353-356. 1900. [Descriptions and key spp. Sampson, A. W., & R . SAMISCH. padres th and seasonal changes in —. Kelloggii. | SANTAMOUR, F. S. The osm number of Quercus dentata. Morris Arb. Bull. 13: 28-30. 1962. [mw = 12, 2n = 24 & E. J. ScHREINER. Juvenile variation in five white oak species. Morris Arb. Bull. 12: 37-46. 1961. [Q. alba, Q. Prinus, Q. stellata, Q. macrocarpa, & Q. lyrata.] SARGENT, C. S. The name of the red oak. Rhodora 18: 45-48. 1916. [Q. borealis ichx. i SAx, H. a Chromosome numbers in Quercus. Jour. Arnold Arb. 11: 220-223. 193 con CG. H. Chloralbinism in the burr oak (Quercus macrocarpa Michx.) and “0s putative modification through grafting. (Abstr.) Genetics 38: 690, 691. pear O. ‘Uber die Typologie des Eichenblattes und ihre Anwendung in der Paliobotanik. Repert. Sp. Nov. Beih. 86: 60-70. 1936. . Entwurf zu einen natiirlichen en der Cupuliferen und der Gattung SHELFORD, V. E. The ecology of North America. xii + 610 pp. University of Illinois Press. Urbana. 1963. [Quercus associations; see especially chapters 2-4, 9, 11, 18.] 194 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Sttimman, F. E., & R. S. Letsner. An analysis of a colony of hybrid oaks [Quercus]. Am. Jour. Bot. 45: 730-736. 1958. [Q. alba X Q. montana (Q. Prinus).] SmirH, C. P. Plurality of seeds in acorns of Quercus Prinus. Rhodora 16: 41- 43. 1914 SPOEL-WALvins, M. R. vAN DER. Les charactéristiques de l’exine chez quelques espéces de Quercus. Acta Bot. Neerl. 12: 525-532. 1963. [Q. petraea, Q. Robur, Q. pubescens, Q. coccifera, & Q. Ilex.] STEARNS, J. L. Distinguishing red oak from white oak by chemical color reac- tion. South. Lumberman 184(2306): 50. 1952.* STEBBINS, G. L., Jr. Variation and evolution in plants. xix + 643 pp. New York. 1950. [Quercus hybridization, 61-66.] » E. B. Marzxe, & C. Eptinc. Hybridization in a population of Quercus marilandica and Q. ilicifolia, Evolution 1: 79-88. 1947. : STeRRETT, W. D. A new oak from the Gulf States. Jour. Elisha Mitchell Sci. Soc. 37: 178, 179. 1922. [Q. Ashei. Trttson, A. H., & C. H. MULLER. Anatomical and taxonomic approaches to subgeneric segregation in American Quercus. Am. Jour. Bot. 29: 523-529. 1942. TRELEASE, W. The American oaks. Mem. Natl. Acad. Sci. 20. v + 255 pp., 420 pl. 1924, i TRIMBLE, G. R., JR. Multiple stems and single stems of red oak give same site index. Jour. Forestry 66: 198. 1968. [Q. rubra. ] Tucker, J. M. Two new oak hybrids from California. Madrofio 12: 119-127. 1953. [Q. X subconvexa, Q. durata X Q. Garryana; Q. X Howellii, Q. dumosa X Q, Garryana.] - Studies in the Quercus undulata complex. I. A preliminary statement. Am. Jour. Bot. 48: 202-208. 1961. [Includes maps. ] —— &C.H. Muller. A reevaluation of the derivation of Quercus Margaretta from Quercus Gambelii. Evolution 12: 1-17. 1958. : » H.S., A. L. Repucx, & A. R. Grove. Floral morphology of white oak Hintonianae: the genus Quercus. Bull. Misc. Inf. Kew 2: 84-95. 1939. [9 taxa from Mexico described for first time. portance of secondary xylem in delimiting Eryth us. Bull. Torrey Bot. Club 66: 353-365. 1939; II. Types parative anatomical study of the wood of Lew cobalanus and Erythrobalanus, Ibid. 115-129, : dimorphism in Quercus nigra L. Am. Midl. Nat. 33: 794. 1945, [Recognizes two distinct shapes. Py Watcur, J. W., R. T. Brycua, & K. W. Dorman, Genetic variation within geographic €cotypes of forest trees and its role in tree improvement. Jour. Forestry 56: 803-808. 1958. [Includes Q. alba.| 1971] ELIAS, GENERA OF FAGACEAE 195 YaRNELL, S. H. Inheritance in an oak species hybrid. Jour. Arnold Arb. 14: 68-75. 1933. [Q. virginiana X Q. lyrata.] ZIRKLE, C. A possible early eighteenth century record of Lea agcsatig in oaks. Jour. Hered. 41: 315-317. 1950. [Q. falcata * Q. rub e first recognized plant hybrid? Jbid. 49: 137, 138. 1958. [Q. J leas x Q. rubra ARNOLD ARBORETUM HARVARD UNIVERSIT CAMBRIDGE, MASSACHUSETTS 196 JOURNAL OF THE ARNOLD ARBORETUM [voL, 52 THE STAPHYLEACEAE IN THE SOUTHEASTERN UNITED STATES? STEPHEN SPONGBERG STAPHYLEACEAE Lindley, Syn. Brit. Fl. 75. 1829, nom. cons. (BLADDERNUT FAMILY) A small family of debated phylogenetic position comprised of five” woody genera with a total of ca. 50-60 species characterized by opposite [or alternate] compound [usually odd-pinnate] or trifoliolate [rarely simple] leaves and caducous stipules. Flowers perfect [or polygamous, rarely imperfect and the plants dioecious] in drooping panicles [or racemes], the perianth of 5 imbricated sepals and 5 petals with both series often connate at the base [sometimes connate into a tube]. Sta- mens 5, alternating with the corolla lobes and inserted either on or below the disk; pollen [di- or] tricolporate [or polyrugate]. Gynoecium [2- or] 3(4)-styled with a superior [2— or] 3(4)-locular ovary partially em- bedded at its base in a hypogynous disc [if present]; carpels with nu- merous apotropous crassinucellar ovules usually in two rows, placenta- tion axile. Fruits capsules [drupes or berries]; seeds with endosperm. Represented in our area by the type genus, Staphylea L. : Divided into two subfamilies by Pax (1897), the Staphyleoideae 1s characterized by opposite leaves, carpels not united their entire length, capsular fruits, and wood with bordered pits. It includes the eastern Asian Euscaphis, the North Temperate Staphylea, and the tropical and subtropical Turpinia. The subfamily Tapiscioideae, consisting of Huerted, endemic to Peru and the West Indies, and Tapiscia, of eastern Asia, 1s distinguished by an alternate leaf arrangement and sometimes simple “Prepared for a generic flora of the southeastern United States, a joint project of the Arnold Arboretum and the Gray Herbarium of Harvard University made Tennessee, Alabama, descriptions in bracke an asterisk have The illustratio a ‘ t : r. R. B. Channell of Vanderbilt University who se? ey materials for the illustration t grateful to Woo his ready peri 4% instructive help in the preparation of the manuscript and to Mrs. Nan y for checking bibliographic citatio e ixth genus, Triscaphis from Indo-China has been referred to the Staphylea- . " J . pings by Gagnepain. Van Steenis (1960) states that this genus is either anacardiaceous or sapindaceous, not staphyleaceous 1971] SPONGBERG, STAPHYLEACEAE 197 leaves, connate carpels that develop into drupes or berries, and wood with simple pits. It has been suggested that the subfamilial traits and other differences, which include a distinct pollen morphology (a polyru- gate type), would justify treating the Tapiscioideae as a separate fam- il Placement of the family has been problematic since the Staphyleaceae appear to have affinities with several groups. Historically, the conflicting views have centered around placement of the family in either the Celas- trales (De Candolle, Wettstein, Rendle, and Bessey) or the Sapindales (Reichenbach, Bentham & Hooker, Engler, and Hutchinson). More re- cently, Takhtajan (1959) has considered the Staphyleaceae to be a prim- itive member of the Sapindales. Cronquist (1968), although placing the family in the Sapindales, does not consider it ancestral to the order despite its many primitive features. Foster (1933), because of the similarities of the cytological aspects of Acer and Staphylea, concluded that the Staphyleaceae could have had a common origin with the Acer- aceae. Heimsch (1942) concluded on the basis of his anatomical studies that the Staphyleaceae did not belong in the Sapindales but might best be placed in the Celastrales. As indicated in the previous treatments of the genera of the Celastrales and Sapindales in the southeastern United States (Brizicky, 1963, 1964) the Staphyleaceae are thought to belong to the Cunoniales (or Saxi- fragales sensu lato) as advocated by Hallier (1908, 1912) and supported by van Steenis (1959, 1960). This alignment was first suggested by Lindley (1835) who based his genus Ochranthe on a species which was later transferred to Turpinia. Likewise, Kaernbachia Schltr. was de- scribed as a new genus in the Cunoniaceae but has been shown by van Steenis (1959) to be based on two additional species of Turpinia. The evidence for the placement of the Staphyleaceae with the Cunoniaceae hinges not only on the strong morphological resemblances (which are well outlined by Hallier, 1908, and van Steenis, 1960) but also consists of anatomical similarities (van Steenis, 1959). The admittedly scant embryological data summarized by Davis (1966) would also seem to be in accord with this placement. On the basis of distributional evidence, van Steenis (1960) has pointed out that the genera of the family “repre- sent a marked northern counterpart to the Cunoniaceae which is largely a southern hemisphere family.” Speaking of such antipodal pairs, he continues, “The meeting point of the areas of each pair and their over- lapping margins are almost always found in the vicinity of the tropical zone. It would seem that the birthplace of these pairs must have been the tropics from which their ancestors have been branched off with a northward and southward directed distribution, respectively giving way to a subsequent development (diversity) in antipodial [sic] centres.” Although the placement of the Staphyleaceae in the Cunoniales seems to be the most reasonable alternative, this position should be consid- ered tentative until further studies employing modern techniques have been made. Such a study might well include investigations of Sambucus 198 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 and Bischofia (cf. Airy Shaw, 1967; Webster, 1967, p. 312), which also seem to have affinities with the Staphyleaceae. REFERENCES: Ary SHAw, H. K. Notes on the genus Bischofia Bl. (Bischofiaceae). Kew Bull. 21: 327, 328. 1967. Brizicky, G. K. The genera of the Sapindales in the southeastern United States. Jour, Arnold Arb. 44: 462-501. 1963. ——. The genera of the Celastrales in the southeastern United States. Ibid. 45: 206-234. 1964. CHaLk, L., & M. M. Cuatraway. Perforated ray cells. Proc. Roy. Soc. Lon- don B. 113: 82-92. 1964. Cronguist, A. The evolution and classification of flowering plants. x + 396 pp. Boston. 1968. Davis, G. L. Systematic embryology of the Angiosperms. x + 528 pp. New York. 1966. Erprman, G. Pollen morphology and plant taxonomy. Angiosperms. mi + 539 pp. frontisp. Stockholm; Waltham, Mass. 1952. (Corrected reprint with addendum. 1966.) Foster, R. C. Chromosome number in Acer and Staphylea. Jour. Arnold Arb. 14: 386-393. pl. 81. 1933. Guertin, P. Développement de la graine et en particulier du tégument séminal de quelques sapindacées. Jour. Bot. Morot 15: 336-362. 1901. Hauer, H. Uber Juliania, eine Terebinthaceen-Gattung mit Cupula, und die wahren Stammeltern der Katzchenbliitler. 210 pp. Dresden. 1908. (Also published in Beih. Bot. Centralbl. 23(2): 81-265. 1908.) . Lorigine et le systéme phylétique des Angiosperms exposés @ Vaide de leur arbre généalogique. Arch. Neerl. Sci. Exact. Nat. B. IIT. 1: 146-234. pls. 1-6. 1912. Hemscu, C., Jr. Comparative anatomy of the secondary xylem of the “Grul- nales” and “Terebinthales,” of Wettstein with reference to taxonomic grouping. Lilloa 8: 83-198. pls. 1-17. 1942. [Staphyleaceae, wood anatomy, 163, 164; relationships, 189. For summary & conclusions of study se¢ Trop. Woods 71: 39. 1942.] Krause, J. Staphyleaceae, Nat. Pflanzenfam. ed. 2. 20b: 255-321. 1942. [Sy tematic treatment of the family with discussion of the phylogenetic place- ment; includes an extensive bibliography. ] EN, B. L. VAN DER. Staphyleaceae. In: C. G. G. J. VAN STEENIS, Fl. Males. I. 6: 49-59. 1960. [Systematic treatment of Turpinia; includes 1% troductory notes on the position of the family by vAN STEENIS.] anche J. Ocranthe arguta, fine-toothed palebloom. Bot. Reg. 21: fl. 1819. itor . Y & L. Cuatx. Staphyleaceae. Anat. Dicot. 1: 443-447. ee YANA, L. L. Embryology of th h . Sci. Bangalore 4% 908/404, 1608 gy of the Staphyleaceae. Curr. Sci gi — F. Staphyleaceae. Nat. Pflanzenfam, III. 5: 258-262. 1896. TEENIS, C. G. G. J. vAN. See B. L. vAN DER LINDEN, 1960. ‘del Pio ieaoe of the Papuan genus Kaernbachia Schltr. (Cunoniaceae . to furpinia (Staphyleaceae). Nova Guinea II. 10: 211, 212. 1959. AKHTAJAN, A. Die Evolution der Angiospermen. viii + 344 pp. Jena. 1959. 1971] SPONGBERG, STAPHYLEACEAE 199 Wesster, G. L. The genera of Euphorbiaceae in the southeastern United States. Jour. Arnold Arb. 48: 303-430. 1967. YAMABAYASHI, N. Types of vessel perforations in Korean woods. Trop. Woods 20-22. 1936. [Perforations multiple in wood of Staphylea and Eus- caphis. | 1. Staphylea Linnaeus, Sp. Pl. 1: 270. 1753; Gen. Pl. ed. 5. 130. 1754, “Staphylaea.”’ Deciduous shrubs, rarely attaining treelike stature with terete branches, the pith terete; bark becoming gray to black and somewhat mottled. Leaves + decussate, trifoliolate [or pinnately compound], the rachis finely pubescent (in ours), becoming glabrate; lateral leaflets subsessile, often becoming distinctly petiolulate; blades ovate to elliptic, about twice as long as wide, with a prominent midvein; the base oblique or obtuse, margins serrulate, the apex cuspidate; finely to densely pubescent (in im- mature material), the density greatest along the midrib and at the base; stipules linear, extra-petiolar, membranaceous, with ciliate margins, leav- ing distinct scars that become obscure at older nodes; rhachis with two small, stipule-like glands near the insertion of the petiolules and (in im- mature material) two obscure glandlike stipels at the base of each leaf- et. Anthesis in spring [flowers appearing before the foliage and on sec- ond year wood, or| the panicles partially obscured by the immature foliage of that year’s growth; each pedicel of a panicle with a pair of stipule-like bracts at its point of attachment to the peduncle, and with a swollen area at about the middle point of its length which appears to be a point of disarticulation. Calyx partially united at the base, only slightly shorter than the [pink or] cream-colored petals. Stamens in- serted below the disc; filaments pubescent [or glabrous] below, exserted from the perianth at anthesis; anthers 2-loculate, longitudinally dehiscent and 2-lobed below the point of filament attachment; pollen binucleate when shed. Styles 3(4), free for most of their length but connate at the expanded, blunt stigmatic apex; ovary pubescent [or glabrous], par- tially embedded in the disc, 3(4)-loculate. Fruit an inflated, mem- branaceous, indehiscent capsule, the carpels usually separating from one another at the apex during maturation to yield a 3-beaked, 3-lobed, el- lipsoid to obovoid brownish fruit. Seeds brown, subspherical, somewhat flattened, with a hard, impervious seed coat developed from the outer integument and a thicker, inner porous layer developed from the inner integument; endosperm fleshy; embryo straight with flattened cotyledons. Embryo sac of the Polygonum type. Base chromosome number 13. Lecto- TYPE SPECIES: S. pinnata L.; see N. L. Britton, N. Am. Trees 637. 1908. (Name an abbreviated form of Staphylodendron [Tournefort], from Greek, staphyle, a raceme or cluster, and dendron, tree.) — BLADDERNUT. A genus of nine or ten species with a disrupted arcto-Tertiary distribu- tion in the North Temperate Zone of both the Old and New worlds; one species native in our area, one or possibly two additional species native 200 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 x %; f, seed, X 3: g, embryo with disc-shaped, flattened cotyledons, X 3. to the New World, several infrequently cultivated. Staphylea trifolia i a small shrub becoming arborescent in the southern part of its bas be to 6 m. high and 15 cm. in diameter), is usually found growing ane ni bs or limestone soils in the understory of deciduous forests and in thic . along stream banks, in moist ravines or on the shores of lakes and ee The species ranges from southern Quebec and Ontario to Minnes a southward through eastern Nebraska and Kansas to Oklahoma, pes kansas, Tennessee, extreme northeastern Mississippi, and northern . bama (along the Tennessee River) and Georgia. According to 4 peti cited by Krause (1942), it was once known to grow in the _— er Alexandria, Louisiana. In our area it is primarily a species of the A : mont, extending into the mountains or onto the Coastal Plain in only few scattered localities. : Staphylea trifolia differs from the Californian S. Bolanderi A. a endemic to the lower altitudes of the Sierra Nevada, primarily in = pubescent ovary and anther filaments. Staphylea Pringlei S. are: described from Nuevo Leén, Mexico, is doubtfully distinct from S. i folia. Staphylea Emodi Wallich, of the northeastern Himalayas, ~~ s to be closely allied to S. trifolia. Additional trifoliolate Asian ieee : include S, holocarpa Hemsley, of central China; S. Bumalda DC., in 1g enous to central and western China, the mountains of Japan, and —-_ and S. Forrestii Balf. f. and S. shweliensis W. W. Sm., both of ie: China. Other species may be yet undescribed from this region. Staphy 1971] SPONGBERG, STAPHYLEACEAE 201 colchica Stev. is native to the southern Caucasus, while S. pinnata L. is distributed throughout Central Europe east into Asia Minor. Both have pinnately compound leaves. Their reputed hybrid, S. elegans Zobel, known only in cultivation, is the only hybrid reported for the family. Brown (1944) recorded S. splendens (Berry) Brown from the Eocene Wilcox Flora of Tennessee but noted that S. acuminata Lesq. from the Florissant beds of Colorado more closely resembles S. trifolia. The presence of Staphylea in the Wilcox Flora, as well as other fossil floras, may give support to van Steenis’s hypothesis of a tropical origin for the enus. The wood of Staphylea is a relatively primitive diffuse-porous type with scanty paratracheal parenchyma and vessel elements with many- barred scalariform perforations and spiral thickenings in the walls. The wood rays are heterogeneous, tracheid or fiber-tracheid fibrous elements are present, and the phellem is derived directly from the epidermis. The stomata, of the cruciferous type, are confined to the lower surfaces of the leaves. The few chromosome numbers recorded for the genus are the only counts available for the family. Staphylea Bumalda and S. pinnata are diploids with 2n = 26, while S. colchica, a tetraploid, has 2n = 52, and S. trifolia is a hexaploid with 2n = 72. Foster (1933) observed sec- ondary pairing in meiotic figures from pollen-mother-cells of S. trifolia, and Sax (1938) found that stomatal density decreases with increased ploidy level in these cytologically known species. Little is known about either the breeding system or the pollination ecology of species of Staphylea, but entomophilous pollination seems to be the rule. The flowers of S. trifolia are proterogynous and adapted to long-tongued insects. The hypogynous disc surrounding the ovary has been reported to be nectar secreting. The number of capsules developed per inflorescence seems to be small in comparison with the number of flowers, and ovule abortion seems to be great and seed set small. Each locule typically contains between four and twelve ovules in S. ¢rifolia, but Harris (1911b) found that of 12,072 fruit locules examined 47 per cent lacked seeds, while of those in which seed matured, 93 per cent contained only one or two seeds. The low percentage of seed set has been used as an argument for the presence of self-incompatibility and the necessity for cross-pollination, but the biological basis of the observed facts remains unknown. Shrubs of S. trifolia often occur in dense colo- nies that have been generally assumed to result from root suckering. Its inflated, bladdery fruits are quite variable both in size and degree of apical lobing, and Harris (1912) found a positive correlation between fruit length and the number of seeds per locule. The capsules remain on the shrubs after the leaves have fallen in late autumn, making the colonies conspicuous during the winter months. Both the capsules and the individual seeds of Staphylea are buoyant, perhaps in adaptation to water dispersal. Ridley (1930) also suggested that the light weight of 202 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 the fruits may allow for wind dispersal. Seeds of S. trifolia have been germinated after scarification and freezing (Dore, 1962). None of the species of Staphylea has any reported economic use other than as a garden ornamental. It has often been suggested that these shrubs should be more widely planted because of their pleasing flowers, interesting fruits, and shade tolerance. REFERENCES: Under family references see ERDTMAN, Foster, Hetmscu, Krause, and Pax. BarTKOWIAK, S., & M. GostyNsKA-JAKUSZEWSKA. Variability of Staphylea pin- nata L. seeds. (In Polish; English summary.) Arb. Kérnickie 10: 27-47. 1965, BEAL, W. J. Agency of insects in fertilizing plants. Am. Nat. 1: 254-260. 1867. [S. trifolia, 258; concludes cross-pollination is probable. | BEAN, W. J. Trees and shrubs hardy in the British Isles. ed. 7. vol. 3. viii + 664 pp. pls. 1-41. London. 1951. [Descriptions and notes on distribution of six species of Staphylea, 368-371. Browicz, K., & M. GostynsKA. Materials to the history of the geographical distribution of the bladdernut (Staphylea pinnata L.) in Poland. (In Polish; English summary.) Arb. Kornickie 3: 237-246. 1957/58. Brown, R. W. Temperate species in the Eocene flora of the southeastern United States. Jour. Wash. Acad. Sci. 34: 349-351. 1944. [Identification of Staphylea in the Wilcox Flora.] Doney, C. F. Species of bladder-nuts. Gard. Chron. Am. 43: 305, 306. 1939. [Discusses uses of the various species in ornamental plantings. | Dore, W. G. The bladdernut shrub at Ottawa. Canad. Field-Nat. 76: 100-103. 1962. [Includes description of S. trifolia {. pyriformis, discussion of pre- vious collections, map of the geographical distribution in Canada, and hy- pothetical migration routes. Fox, W. S. . H. Soper. The distribution of some trees and shrubs of the Carolinian Zone of southern Ontario. Part I. Trans. Roy. Canad. Inst. 29: 65-84. 1952. [Includes distribution map of S. trifolia and list of repre sentative specimens. Gostynska, M. The bladdernut, Staphylea pinnata, in Poland. (In Polish; English summary.) Chronmy Przyr. Ojczysta II. 16: 12-22. 1960. . The distribution and ecology of the bladdernut — (Staphylea pinnata oe Poland. (In Polish; English summary.) Arb. Kornickie 6: 5-71. Harris, J. A. The correlation between a variable and the deviation of a depen- dent variable from its probable value. Biometrika 6: 438-443. 1909. [Table I, 440; S. trifolia inflorescences with fewer flowers tend to develop rela- tively more fruits than do larger inflorescences. | . On the selective elimination occurring during the development of the fruits of Staphylea. Ibid. 7: 452-504. 1910. [Statistical study of observed low seed set in S. trifolia.] : Further observations on the selective elimination of ovaries in Staphy- lea. Zeitschr. Indukt. Abst. Vererbungslehre 5: 173-188. 1911a.* . Seed weight in Staphylea and Cladrastis. Torreya 11: 165-169. 1911b. [Lack of correlation between number of mature seeds and weight. ] 1971] SPONGBERG, STAPHYLEACEAE 203 The influence of the seed upon the size of the fruit in Staphylea. I. Bot. Gaz. 53: 204-218. 1912 Hoven, R. Some features in the anatomy of the Sapindales. Bot. Gaz. 53: 50-58. pls. 2, 3. 1912. [Staphylea, wood anatomy, 53. MAcsrip_, J. F. A new species of bladdernut. Rhodora 20: 127-129, 1918. [S. Brighamii from Ohio is described as new on the basis of flower, fruit, and foliage characters probably within range of variation ue Nf trifolia MEEHAN, T. Fertilization of flowers b agency. Proc. Acad. Nat. Sci. Phila. 1876: 108-112. 1876. [Includes interesting ieee between Meehan & Asa Gray concerning S. trifolia.] Mitosis [sic] in the pollen mother-cells of Acer Negundo L., and Staphylea fo L. Ann. Bot. 28: 115-133. pls. 9, 10. 1914. [S. tri- folia, n = Riwoie, L. C. ee of the embryo sac and embryo of Staphylea tri- foliata [sic]. Ohio Nat. 6: 320-325. 1905. [Two plates depict various stages of development. ]} Riptey, H. N. The dispersal of plants eel tps the world. xx + 744 pp. 22 pls. Ashford, England. 1930. [Staphylea, 73, 75. ROBERTSON, C, Flowers and insects. III. Bot. tab 14: 297-304. 1889. [Staphy- lea, 302, $03. SARGENT, C. S. New or little known plants. Staphylea Bolanderi. Gard. For- est 2: 544. fig. 142. 1889. [Brief description of the species and notes on the type locality.] Sax, H. J. The relation between stomata counts and chromosome number. Jour. Arnold Arb. 19: 437-439. 1938. Sax, K., & H. J. Sax. Stomata size and distribution in diploid and polyploid plants. Jour. eee Arb. 18: 164-172. pl. 205. 1937. [The polyploid levels given for some of the species of Staphylea are apparently incor- rect. The later cabkcation by Sax (1938) is in agreement with published counts WaAAnpERs, G. L., & J. J. Skvarta. Fine structure of Staphylea trifolia L. (Staphyleaceae) pollen walls. Pollen Spores 10: 475-478. 1968. Wince, O. The chromosomes, their numbers and general importance. Compt.- Rend. Trav. Carlsberg Lab. 13: 131-275. 1917. [In S. pinnata, Winge found ~ = 12, but noted m = 13 in one cell.] ARNOLD ARBORETUM HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS siete aus ounce ANAGEMENT seo cIRCULATIO Poge 1 Publisher: File two copies of this oe with yoRe pe Form Approved, | Budget Bureau No. 46-R02? pl cat of Fie 23, 1962: Section 4369 39, United oT Code) ie on page 2 1. DATE OF FILING TITLE OF PUBLICATION September 29, 197) JOURNAL OF ARNOLD ARBORCTUK 3. FREQUENCY OF ISSUE 4. LOCATION OF KNOWN OFFICE OF PUBLICATION (Street, city, county, state, ZIP codi PUBLISHER (Nume and address) EDITOR (Name and address) (January, April, July, October) | 22 Divinity Avenue, Cambridge, Massachusetts 02138 S. LOCATION OF THE HEADQUARTERS OR GENERAL BUSINESS OF OFFICES OF OF THE PUBLISHERS (Not 22 Divinity Avenue, Cambridge, Boece hecatte 02138 S OF PUBLISHER, EDITOR, AND MANAGING EDITOR MA 62438: The Arnold Arboretum of Harvard University, , 22 | Divinity gees Cambridge, i Bernice G. Schubert, 22 Divinity Avenue, Cambridge, Massachusetts 02134 MANAGING EDITOR (Name and_address) Dulcie A, Powell, 22 Divinity Avenue, Cambridge, Massachusetts 02138 = ie coins cseteenl- miata apes ane ae OWNER (If owned by a corporation, its name and address must be d and also vapaaaerakt thereunder the names and pe 3 pchbiebderd owning or holding 1 be oe or more , total amount rai stock. If n POD Form May 19668 (Signature of editor, publisher, busis mess manager. or owner) } Cater Cathet EE casi Journal of the Arnold Arboretum Published quarterly in January, April, July. and October by the Arnold retum, Harvard University. Subscription price $16.00 per year, beginning January, 1972. Subscriptions and remittances should be sent to Miss Dulcie A. Powell, Arnold < Suan 22 Divinity Avenue, Cambridge, Massachusetts 02138, U.S.A Volumes I-XLV, reprinted, are available from the Kraus ee Cor- poration, 16 East 46th Street, New York, New York 10017, U.S.A EDITORIAL STAFF B. G. Schubert, Editor C. E. Wood, Jr. T. G, Hartley CIRCULATION D. A. Powell Printed at the Harvard University Printing Office, Boston, Massachusetts COVER: The Golden Larch, Pseudolarix Gordon The cover design is the work of Norman Comeau of the Art Depart ment, — ‘Todd Company, Printers, Boston Second-class postage paid at Boston, Massachusetts JOURNAL OF THE ARNOLD ARBORETUM VoL. 52 APRIL 1971 NUMBER 2 WHAT IS THE PRIMITIVE FLORAL STRUCTURE OF ARALIACEAE? RicHARD H, Eype AND CHARLES C. TSENG “THE GENERA OF Araliaceae that are most remote from Umbelliferae are Plerandra, Tetraplasandra, and their near relatives; in general, these are the genera that are distinguished by a greater number of stamens or carpels. In an arrangement where the Umbelliferae follow the Araliaceae, one would have to place these genera at the beginning of the family. Since we can observe within the Araliaceae a very gradual transition from spe- cies with many stamens and carpels to those characterized by 5 stamens and 2 carpels, it is not improbable that we have here a reduction series and that the forms placed at the beginning of the generic sequence also represent the oldest types of the family.”” (Harms, 1894, p. 21, 22.) ““A few of the Araliaceae have ten or more petals, stamens, and carpels in a regular, symmetrical arrangement. These polymerous types have some- times been regarded as primitive within the family, but they more prob- ably have undergone a secondary increase in the number of parts of each kind. Pentamerous flowers are here regarded as primitive in the family and order.” (Cronquist, 1968, p. 278 The first quotation, translated rather freely from Harms’s monograph of the Araliaceae, cautiously expresses a view that has long prevailed con- cerning floral evolution in the family. H.-L. Li, in his revision of the Chinese Araliaceae, put it more emphatically. Of Tupidanthus, he wrote: “Its primitiveness . . . as indicated by the numerous stamens, ovary- cells, and styles, is unquestionable.” Baumann-Bodenheim (1955), a lead- ing student of fruit structure in Araliaceae and Umbelliferae, recognized a five-stage reduction series from the many-carpellate ovary to the pseudo- monomerous ovary. As far as we know, Cronquist is the first author to question the “unquestionable” primitiveness of the polymerous flower in this alliance. 206 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 characters. The most detailed observations to date (Rodriguez, 1957) show advanced xylem characters in temperate herbs (Aralia californica, A, hispida) and in the vine Hedera helix, where one would expect to find specialization, but primitive xylem characters such as scalariform perfora- tion plates and heterogeneous rays are distributed among genera that differ greatly with respect to numbers of floral parts. Of course, future work on the wood anatomy of Araliaceae may reveal correlations with floral structure that are not now apparent. It should be kept in mind, however, that at least some of the Araliaceae have evolved through re- peated radiation on oceanic islands, radiation apparently accompanied by complex changes in growth habit (e.g., trees to shrubs and back to trees again; see Carlquist, 1965, p. 191). If such changes were sufficiently widespread, the usual trends of wood evolution might be difficult or im- possible to find in this family. Since associated xylem characters are not presently of value for as- sessing the evolutionary status of polymerous flowers, we looked for an association between vascular characters— more specifically, vascular characters of the gynoecium — and the number of floral parts. One basic tenet of evolutionary plant morphology is that flowers with sepals, petals, Stamens, and carpels united in various ways have evolved from flowers with all appendages free. Similarly, a flower in which all of the princi- pal vascular bundles are separate is more primitive than one in which the principal bundles are variously united. In a group with inferior ovaries, the degree of union between dorsal carpel bundles and peripheral bundles (supplying epigynous appendages) may differ in different taxa, thus pro- viding an indication of evolutionary advancement (Eames & MacDaniels, 1947, Figure 167; Eyde, 1967, Figures 5-8). Baumann (1946) found that the Araliaceae differ among themselves in this regard,! but his ob- servations were not detailed enough for our purpose; so we surveyed the family to establish the taxonomic distribution of the differences. oe and for like reasons, we examined our material for dif- Tences in the position and the degree of union of ventral bundles. MATERIAL, METHOD TABLE 1 lists the collections fro gether with the manner anatomical study without detriment to the collection. i removed two to four flowers for serial sectioning and seribdng the lcs known as Baumann-Bodenheim) had his own manner of de- orm of the gynoecial vascular system. Where we say the dorsal carpel bundles are united wi : (Baumann, 1946, p. 63 Peel a aime he would say the carpel bundles are absent In most cases, we 1971] EYDE & TSENG, ARALIACEAE 207 a similar number for clearing; however, we occasionally made our ob- servations from a single flower, found nothing of much interest, and declined to remove others from the sheet. We sectioned the flowers transversely, supplementing with longitudinal sections only in the case of Tetraplasandra (on which we have written a separate paper; Eyde & Tseng, 1969). We cleared whole flowers by treating them with NaOH and chloral hydrate, then passed them through an ethanol series into toluene for examination. If this treatment did not reveal the vascular system in sufficient detail, we next transferred the cleared flowers from toluene to melted paraffin, cast the paraffin into blocks, and cut the blocks into pieces with a razor blade, using a dissecting microscope to orient the smaller specimens. When we subsequently dis- solved away the paraffin matrix with toluene, the cut pieces often showed vascular detail not visible in whole cleared flowers. OBSERVATIONS Tastes 2 and 3 summarize the important features of gynoecial vas- culature for each of the examined species; the tables also list the number of petals, stamens, and carpels for each. Sepals are not listed because the calyx is often poorly developed or lacking in Araliaceae. In TasLe 2, species are arranged in four groups, according to whether their flowers are best described as polymerous, 5-merous in all whorls, 5-merous with 2- merous gynoecium, or 5-merous with 1-locular gynoecium. Assignment is rather arbitrary in some cases, but most species fall readily into one or an- other of these groups. Literature citations accompany a few of the entries; in such cases, the vascular characters are not our own observations but are taken from an illustration or from descriptive comments in the cited work. Tetraplasandra species vary so greatly with respect to meristic characters that we decided to list them separately (TABLE 3). In many species of Araliaceae — especially those with polymerous flow- ers — the number of floral appendages in each whorl varies from flower to flower. Obviously, we could not expect to establish the limits of varia- tion for such species by examining only a few flowers of each (moreover, some of the flowers had petals or stamens missing) ; therefore, the counts given in TaBre 2 rely heavily on published descriptions. Several of our entries for petals and stamens of Oreopanax species are based on Smith’s (1941) generic description in North American Flora, in which flowers are said to be “5 (rarely 4- or 6)-merous.” In general, Smith’s treatment does not give the numbers of floral parts for individual species. We use quotation marks wherever we have made no count of our own. ‘For the most part, however, we did make one or more counts which fell within the limits given in the literature; in such cases we list the literature limits without quotation marks, When our count deviated from the counts of other authors, we usually expanded the literature limits to accommodate our observation. For instance, Clarke’s treatment of Schefflera (Hepta- pleurum) khasiana in Hooker’s Flora of British India (see generic de- 208 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Ficures Transverse sections through inferior ovaries of Araliaceae. 1+4, FIGURE 1. Oreopanax nitidus, Cuatrecasas 27539: dorsal carpel bundles (arrows) 1971] EYDE & TSENG, ARALIACEAE 209 scription, vol. 2, p. 727) would indicate that the flowers are completely 5-6-merous, but we count 9 ovary Jocules, so our entry in the GYNoECIUM column of TABLE 2 is 5—9 to accommodate Clarke’s count as well as our own. In two cases, however, we list only our own count, because we are not confident that the specimens are correctly identified. One of these is Boerlagiodendron novo-guineense: our preparation has 12 carpels, whereas Harms (1894, p. 31) lists this species among those with “Frucht- knotenfiacher 7-9.” The other is Gastonia boridiana, reportedly 8-10- carpellate (Harms, 1938a), but 15-carpellate in our preparation. It is not always easy to characterize the degree of union between dor- sals and peripherals. To be sure, there are taxa (Peekeliopanax, for in- stance) in which all dorsals are distinctly separate, even in sections taken near the base of the ovary; and there are others in which the dorsals are intimately united with peripherals up to the summit of the ovary. But there are also taxa in which the dorsals are united with peripherals in the lower part of the ovary and free from the peripherals in the upper part. To complicate matters further, the several dorsals within a single gynoecium may be free in varying degree.2 At first we planned to use only two sym- bols, + for united and — for separate, in the Dorsars column of our tables, but we found we needed a third symbol (int) for the many taxa that are intermediate. Of course, the addition of a third category did not eliminate uncertainties attributable to poorly differentiated vascular bundles, to less than satisfactory preparations, or to omissions in the works of other authors, These uncertainties are indicated by question marks. Characters of the ventral bundles are treated in two vertical columns, one expressing the degree of union of each pair of bundles, the other their position. In most species, the pairs are readily characterized as separate or united (if the separation is at the placental level only, a frequent con- dition, we consider the ventrals united), but an intermediate designation was necessary in certain cases. The ventral supply of a few species is anomalous. In Meryta sinclairii, for instance, it consists of a complex of slender anastomosing strands with no apparent symmetry. In Tetra- plasandra racemosa the ventral system is seen in cross section aS a series of vascular patches arranged in a circle, with most patches containing several discrete xylem strands. The ventral system of Plerandra vitiensis has a somewhat similar appearance, but only in the upper half of the *This is true of our specimens of Hedera helix, for instance. Apparently, other authors have observed little or no union of dorsals and peripherals in this species (Eames & MacDaniels, 1947, p. 352; Philipson, 1967). united with peripheral bundles, duplex ventral bundles heterocarpellous re aligned with the septa, see also Ficure 10); X 26. Ficure 2. ionising ig oerstedianus; somewhat oblique section; ventral bundles fewer than carpels, ap- is; do bundles (inner arrows) and peripheral bundles (outer arrows) separate, ventral bundles anomalous ik text): < 20. Ficure 4. Hedera helix; double ovary (see also Ficure 8); ventrals on left are fewer than carpels; X 15. JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 210 sx ‘gy eryeuInS (V) Obs Sa040T 1s oUYDY ‘bi, wnpyyouuid wnipypap [> ‘sx ‘yd ‘loqiy ‘eN ‘S'Q “7N9 "| vsouids Dyvap [> ‘sx ‘yd Joqiy ‘eN ‘S'9 “7N9 ‘wiaag (‘bIP,) 07Dja DYDAp sx ‘IyY yadnyy (HO) £622 fauay sx ‘By ueunyy (Vv) 96Z JazzD FW-japun H “LI ("'T) snqyoyofi4y xpundoyyuvap sx “1Jy uedef (V) 9S9¢ 1oAn]Y sx ‘gy uvdef (Vv) ,.€0008b,, 2yNZNS “bI, CHT) snsouids xoundg oyjuvapy sx ‘1yY Isuays (Vv) “u's Ipppaiy sx ‘YY yedny (V) 896[ uosp 4 S[9IC] X9 SULIV]Y Sisuauanyojas xpundoyuvap sx “By yodox (Vv) £0E ysnms "W99g (uixepy Q Idny) snsopyissas xpuvdoyzuvdIp sx ‘yy eunyouryy (V) ‘u's 2oz440045 SULIvFT (urxeyy “Idny) snsoa1uas xpundoyjuvap sx ‘yy yodny (V) L96[ “ost suey (‘AIQ) snztyssoona xouvdoyjuvap sx ‘yy va10y (V) 9162 OW] IBYeN Snuvasoy xDundoyuvIp sx ‘py uedef (V) 8669 uosM reyeN (‘9907 FR ‘qats) suvrouut xpuvgoipoagd = ‘Aes ® "youvsy (:99NZ BW ‘qais) supaouut xpuvdoyjyuvap sx “ayy yedny (v) LL6L uospM sx ‘py yednyy (H9) 2E8p¢ Kauay suey (‘AI[Q) Msuay xpuvdoyjuvap sx “134 MOYIIIM (¥) 882 094) & ‘on1yD ‘pavmars P ‘BY yedny (SA) POLE uOsTLM sx ‘py Surrey) (Vv) [8¢0z2 Surya “WS “MM SMpAysiptavas xvuvdoyjuvIp sx “ayy UBMYIIZS (9) Leg IutyD sx “py UBMYIIZS (Vv) [LOI uostt P ‘sx “py Isuays (SA) Shp mwopang suey upjvss xvuvdoyjuvop NOLLVaVdaag ALITVIO'"T . 44aWON ® ANVN S,AOLIATIOD Sa1oadS suonesedaig ‘suonsa}0D “| A1av I 211 EYDE & TSENG, ARALIACEAE [2 ‘sx ‘py sx ay P ‘By 19 ‘sx “py sx ‘P ‘BY pouInr) MON vIUOpaTea MAN BIuOpa[ea MAN ISsUPM seinpuoyy ug ‘day uvoumm0g ‘ST PIBMPUT AA eluopatea) MeN BoMyY WINS HUMeAy ITRMEFT ueuuny” "SJ WOUIOTOS eauIny) MAN eyerjsny saurddiyg (V) O€ITET Ysaajssay G ssvag (Vv) £6012 pav M-uopsury (sn) c99¢ zauawis (V) Z66ZI 40}904g (H9) Zeeg Kajo09 (V) Ob Suamary (V) [Z99¢ ajqnjsuoy G uosuyor (H9) Log] supyyty SULILE] DUDIPLLOg DINOISD+) PALL “AO Vsoyt9 vaquovy ‘yourlg W susaq (qunyy]) vauodovl visyoy ‘T[leq DyDysnsuv xoundowaay ‘TIN QBIA (YONA) Dwutssyunsaja vaay,03K21q ‘Z2RIN~ purpy snyzuvkyovjs xvund ojdiqr “Wourid F ‘ausaq ([qny) 1020,0140m xpundowxk pig ‘Yury, (MS) snyonuazzn xpundowmdpiq ‘Wg “UUOg 4a94anb xDundoipuaq ‘WS ‘Dy (Ws ‘uU0d) snpodoznu0s xpundoupuag ‘LI (SIA) t4aypaays xvundoapuaq ‘TILA Dxopoand vasqanjaq SULIBVE] UY IDQ4ajND] VaAqaDjaqy *YAIZ BW PYOY Vpojnawupg vuossny JIJPH (pnery) wnudks144 uospuaposayy) ‘{ely asuainny uospuapo.snay) “WaaS Dpidsiy sisdowssvaig MYM LO Mnapuvija] Uuospuaporsnjsaog suey (“HoqS) ASUAIUINS -OLOU UOAPUAPOLSDIAIOg ‘Id vso02204f Dy I1.40.4SP ‘TA mniyofissaaip unyxydgosyjAp ‘Aa Unuviusayo wnycydosyqap JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 212 PD ‘sx ‘yd [PP ‘sx ‘yd [Pp ‘sx ‘yd a ‘BY 1 ‘sx ‘py fo “sx ‘ary eIquUIO]OD BIQUIO[OD eIquO]O.) Bysely “puryeaz MON eBay vouiny) MON seoonjoyy yWNos (3nd) epnuioag (SA) 6€¢2z2 Sspspraaqony (SQ) p2rLzZ SPsvIaa4jongD (SQ) [[pZzZ spsv2aajon5 (sa) S€6 uossapUupy (sa) L0gZ appoyrs @ punjsoox (Vv) OzzZz qq "ALIS “10.J ‘Puy “YIAN (AN) £297 Ka1q40 MY ‘u0jj14g ‘uaoug ‘VEND snpiyiu xpuvgoasC) ‘qourlg ® “eudeq CM'A'H) snpungtuoy xvungoas—9 ‘by Cus “a “[) suprssoy xvungojdo ‘BIA preqnq snzosonjoaut sndav20poa py “WI99G MIDI IUIS DILL PY I'T (Surly) t4ajzsy xvund onptssa py “WY Las xoundg opps a py "WII (“[[RM) Suzojnpun xvuvdo.s20 jy [Sula] pssp49 piguryopuy "UIdaS ("qxOY) SuDssv4f xDUDd 042127 VJ (uund) svsuauzyo xpundosaja yy "I *9Y Déapay] "UINYIS “YY uUsusvy xDuvd o1smAD FT ‘bi, Duondndg viuojspy “Wey DIsuOgsS14NI DINOISD+) q NOILVavdaid ALITVIOT » WAIWAN ® ANVN S.AOLIATIOD Sa10adS (panusquo?) suonvsedarg ‘suonsa][oQ “| ATAV I, EYDE & TSENG, ARALIACEAE 213 1971] Pp ‘sx ‘BY [9 ‘sx ‘ayd P ‘gd [9 ‘sx ‘yd sx ‘yd PIqUIOTO?) pur[eaz MON saurddiryd ‘eg ‘Ajunod ‘uourqay PIULBALA ruleueg eIquio]o) eureueg (SA) POZE Jamog ‘loqay BurqAqys ‘yn (sa) ZIT wossepny (V) [p98 suamayy ‘u's apag (SA) 9ITT 4aypoy (SQ) [86] nsuakp ‘apky ‘usays (sa) ¢¢8Zz sOsU70440ND 0D ‘Igs yipoyy if 717X ‘prey *}Og I0Z0g “3nd suey (Cy a H) sisuanpumb vaaygayos suey (“youedy) upiapp xpupgopnasd Poy “dD (Cuun) ‘y xe ‘[OS) Ssnyofissps9 xpungopnasg uosdiiyg (inj ) snasoqgaD xpuDgopnas g SULIeET press - Dsnliyjnu s19skJog “HIOA DYOf{ipuDss svi9skjog ‘Teg (Weag) sisuarya vapuvsayg uosdifiyd sisuauowmojos papunsatg “WS *D “V 071 0Sul Dépudsayd “W9IS (‘T]eA\) SNyDpsooqns xoundojuag ByeARHY Djospisdounjsp2 xoungojuag suey syignjaags xpuvgouayaag "I Snuofiag xDung ‘I snyofanbuinb xoupg "yourlg Q ‘audaq (° ¥ G'H) sisuadnjox xpuvgoaic:) ‘O = “SUA “T snyojisnjgo xpuDg 024g) JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 214 (sa) Oss6r KD @ ‘uv ‘sauaseaq (T9TT60I SQ) ‘u's oun (H9) 8808 y20y (sa) Orgel yoy (sa) grore 44a ‘adtyspinyy ‘sauasaq (68+608 SA) ‘u's punsgayny (Vv) I8IzZ ayqaty sauasaq (986069 SQ) ‘u's uDwagoy (SA) OZOIT saunas py ) sauasaq (V) L288 4apyiM @ Y20N (SQ) S680T uostt | (V) ObL 094) @ ‘ontyD ‘panmarg (H9) UMOUYUN 107997109 (SA) £066 4ayID 4 (V) PEL6 UosTt (sa) (9) poor uayty (V) L6z2d ynms (SA) 69¢2 adnay & pypqoay J st ‘BY TeMU FT st ‘sx ‘gy TeMeET sx ‘Gy TeMePT 1 BY Heavy sx “BY TeMeH 1? ‘st ‘sx ‘gy TEMPPT [2 ‘st ‘sx ‘gy TeMEPT 1 “ayy TeMeH 19 “By TeMeyy 7 ‘sx ‘gy TWeMeyT 19 ‘py ‘sx ‘ay TEMGRT P ‘By TIeMPY] 19 ‘S| ‘sx “py TeMe HY sx ‘py TeAey P ‘By UBMIET, sx “IjYy MOYITIOM sx ‘gy 1? ‘sx ‘gy puvlea7 i N sx ‘py UPMIE TL, P ‘sx ‘py eueurg [PP ‘sx ‘ayd rurueg sx “Ayy vjanzaue A sx ‘py eoreure sx ‘yy BOTY BISOD P “ayd uojéax q NOLLVavdaug ALITVIOT suey (AQa[[tH) Dépuviawu vapuvsnidn.ze YOY Sesuaiwuny vapuvsyjgdaqa J Joys (uueyl "H) sisuatavy vipunsyjgn.1ya J SULIEFY (‘AQAT[IEZ) avjovy vapunsyjgnaja J ACID) “Y Sisuaimapy Dipunsyydn.442 J HAYS CAG) V¢svr0uwK3 vapuvsvyjdovaja J (oy “Y (Yoo) snsafrskdnd xpuvdo.aya y Avi) "VY (journboef » uoIquIoF{) stzvj0d vdavz0qjuUS5 "'GASUY) mnsjarxa UOApuapopntIs ‘pour ‘ds psayfayas ‘BIA CAG ‘A'N) vepyaqun vsayfayrs suey (IYSIM) Dsomaons Daayfayas » WGNON BP AINVN SAOLATIO|D SaIoaIdS (panusjuo2) suonesedarg ‘suonraqjoQ *| alav.y EYDE & TSENG, ARALIACEAE 215 1971] ‘uotjeiedeid paiva[s =p ‘suonoas [eUIPNzZuo] = s][ ‘suotjes sso1d = sx “yinay papyoid = syd ‘1amoy papyord = yd 4iniy WNteqiey = ayy ‘1aMOy wniueqiey = YY, *10}9a]]09 24} JO UOTSsassod ay} UT [[IJS JAYDNOA ¥B sazBdIpUT YOTYyA ‘OD ydaoxe ‘wnsouDgsaH XapUy MOT[OJ SAVYINOA AO suOTeUsIsop WINTIeqiayy , sx ‘ayy ‘sx ‘yy po ‘asd ‘yo ‘gd [> ‘sx ‘ud [1D ‘sy ‘sx ‘yy Ip ‘sx BY (989M) PISARTR]L ueuunyz ueuunyz Heme Teme yy eM Bouinr) MON eae yy eae (V) 89SZE ON Platy asognsurs (sa) ) 088 4204 (sa) 2g¢T oy ‘ysinbpawy “usayy (sa) £I012Z sday > ‘asiyspanypy ‘dauasaq (V) LPOET DuusnioH 4 duyauny (8L£608 SQ) “u's punsgayiiy (Sa) 6SIbz2 DYSAG | sauasaq "SWIOYL, QJ “YOOR snqn47g¢Kyoo snyquvpign T, ‘ds pisanasy oY apppajpipa papunsnjdvaga L Yaays (saqio) wnsowasns uoLpuapiosun Py = $aqioJq DsomarDs DApUDSDIgDAJA J "qs}JOYS sisuaayndnd vapunsyjgnaja J ‘bi, Suapionnd vapunsnjdn.ja J YlIYS VYyIUDAIIU DApUDSD]gDAJA J, 216 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 ovary; at lower levels there are paired xylem patches in the septal radii, as in Plerandra solomonensis. Schefflera quinduensis shows an anomaly that we found nowhere else in the family: the ventral system is a double series (FicuRE 3), arranged so that each septal radius has two well defined vascular strands, one internal to the other. Another species of Schefflera, S. khasiana, also has supernumerary bundles in the center of its gynoe- cium, but they are not so symmetrically arranged with respect to the septa Ventral bundles can be united, or merely associated in pairs, in either of two ways: if the united bundles belong to adjacent carpels, the union is heterocarpellous; the union is homocarpellous if the united ventrals belong to the same carpel.* Since homocarpellous ventrals lie on the same radii as the dorsal bundles and heterocarpellous ventrals are on the septal radii (Ficure 1), the two types of union are readily distinguished. Oc- casionally, however, one finds flowers in which the duplex ventral bundles are fewer than the carpels; the bundles may then be arranged asym- metrically with respect to the septa and to the dorsal bundles. This is e case in our material of Hladik 551 (Oreopanax aff. oerstedianus, Ficure 2) and we are therefore uncertain as to the original position of the ventrals, although they appear more homocarpellous than heterocar- pellous. It is also true of Reynoldsia sandwicensis, and of some specimens of Hedera helix (Ficure 4), but in these species the asymmetry does not obscure the true position of the ventrals. Our uncertainty concerning the position of the ventrals in Harmsiopanax harmsii and in Cheirodendron kauaiense is due to the difficulty of following weakly differentiated vas- cular bundles in sectioned herbarium flowers. Baumann (1946) supposed homocarpellous ventrals to be the usual condition in araliaceous flowers (see his Figure 3, showing a “typical” Synoecium in cross section), even attributing this type of venation — quite wrongly — to Schefflera and to Polyscias. Actually, the heterocar- pellous condition is far more common. We have found homocarpellous ventrals only in Trevesia, Aralia, Panax, Stilbocarpa, Harmsiopanax, Mackinlaya, and certain species of Oreopanax and Acanthopanax. Miscellaneous observations on the-lateral bundles of the carpels, on crystal complement, and on secretor : t structures are covered in the dis- cussion section. 4 DISCUSSION, CONCLUSIONS P ieee | "6 ied x Po primitive. —If our criteria for recognizing advance- rellable, the question posed as our title is answered unequivocal- ts as “homocarpous” and “heterocarpous,” but and 1971] EYDE & TSENG, ARALIACEAE 217 ly by the data in the tables. Ventral bundles retain their individuality only in certain species with polymerous flowers and nowhere else, even if we should count the intermediate (int) cases as having paired ventrals. Group A also has the largest proportion of taxa with discrete dorsal bun- dles: nine of the 11 Group A genera that we examined have one or more species with free dorsals. One of the two exceptions, Boerlagiodendron, is recognizably specialized by virtue of its sympetalous corollas. It is possible that the Group A genera missing from our survey (TABLE 4) contain a few additional species with advanced vascular characters. The insertion into Group A of additional taxa with derived vasculature would not alter our principal conclusion, however, that this group retains within it the primitive floral characters of the family. Group B has relatively few species with discrete dorsal bundles, and these species belong to only four of the 15 genera listed. Schefflera, the only one of the four with species in both Group A and Group B, appears to be least advanced. Specialized characters found in the other three genera include capitate inflorescences (in Oreopanax), simple leaves (in Dendropanax), and prickly stems (in Acanthopanax). No one doubts that 2-merous gynoecia are derived; therefore the pre- ponderance of fused bundles in Group C serves to support our use of vascular characters as a guide to evolutionary advancement. This part of the table is not, however, without surprises. One unexpected finding is that Tetrapanax, with many specialized features, has markedly discrete dorsal bundles. It is less surprising to find discrete dorsals in the ovary wall of Didymopanax, for Frodin (ms) * merges Didymopanax with Schefflera, where discrete dorsals are not out of place. The relatively primitive vascular system of Delarbrea, together with its peculiar secre- tory structures and its geographic isolation on New Caledonia, place this genus (and its allies Porospermum and Myodocarpus) on an evolutionary line that diverged from other Araliaceae perhaps as long ago as the Cre- taceous Period. : : Comparison of species within certain genera provides additional evi- dence that our criteria are valid and that separate vascular bundles tend to occur in polymerous flowers. In Plerandra, the two typical species Pi vitiensis and P. solomonensis have very primitive vasculature, as would be expected if polymerous flowers are primitive. The exception 1s P. insolita, which we assume to be a specialized member of the genus be- cause of its highly modified inflorescence (see Smith & Stone, 1968, p. 477). In Tetraplasandra, the polymerous species T. paucidens and T. hawaiensis have the least specialized gynoecial vasculature; T. racemosa, the only other polymerous species examined, has an anomalous ventral supply, in keeping with its somewhat isolated position ° within the genus. “D. G. Frodin kindly allowed us to xerograph his impressive handwritten “manu- script “The Complex of Cephaloschefflera” and to cite it here. The work is an aa oo. —— as part of Frodin’s preparation for the doctorate in the Botany a i i . " its a a Tetraplasandra racemosa bears its flowers in racemes, rather than in umbels. Sherff (1952) considered it a separate monotypic genus. 218 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Meristic differences are similarly associated with vascular differences in Reynoldsia, The Hawaiian species R. sandwicensis, ecologically special- ized and belonging to a derived flora, has fewer carpels than R. pleiosperma and other extra~Hawaiian species; it also has the more specialized vas- cular system of the two species we examined. Another genus with vas- cular characters that vary according to the species is Scheflera. Although Scheflera exhibits no simple relationship between meristic characters and opinion was overlooked or ignored by Li (1942), who selected T. calyp- tratus as the most primitive living araliad. of the several systematic treatments incorporating that concept. Indeed, we are convinced that no author has yet subdivided the family in a way that - una Plerandra, i cui fiori sono normal i i i i : " mente e stabilmente divenuti dopii eres ms — di 2 fiori . - -” In support of this notion, Rippa figured a flow ; tga us with one pedicel, two congenitally united ovaries, two androecial circles (side by side), and two Sets of sti = oie ag Correct, however. The td “fusion,” and Rippa’s r terior: oo gured flower, seem to imply a gradual phyletic union, be g in ‘whch en mation of fi d proceeding through intermediate stages pe a en the ovaries are uni The fasciation of flowers in changes are known to Play a major role in the f t f cultivated tomato fruits (Houghtalin asciation of cultivate double flower is pag ss es 1943; Zielinski, 1945). If this is true, Rippa’s Plerand; jes ees Deen questioned: Frodin (ms) considers Tupidanth sely group pet cea sts Subintegra Craib (which is actually a Schefflera, Paratropia : 8 to Frodin) than to any species of Plerandra. 1971] EYDE & TSENG, ARALIACEAE 219 Ficures 5-7. A fruit of Tupidanthus calyptratus, Henry 12298, X 1.5. Fic- URE 5. Viewed from side. FicuRE 6. Viewed from above. Ficure 7. Cut transversely through upper part of locules, most of which are sterile. No at- tempt was made to represent all of the more than 200 stigmatic lobes and Pyrenes shows evolutionary alliances. Harms (1894/ 1897), whose monograph is still the most complete — though it should not be used without also con- sulting his supplementary contributions covering the next 30 years — divided the Araliaceae into three principal groups. Earlier, Bentham and Hooker (1867) had recognized five main groups (as series), and Viguier (1906) later recognized 10 groups. Each of these treatments is artificial 220 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Ficure 8. Hedera helix; fruit Shealnise from double ovary; X 7.5. A sec- ond example from the same plant is shown in transverse section as FIcuRE 4. to some extent, but the modification proposed by Hutchinson (1967a) is even more s Hutchinson’s arrangement of the family follows that of Bentham and Hooker, except that he treats their series as tribes and adds two new tribes for species with solitary flowers in the inflorescences. The new tribes are placed first, in keeping with Hutchinson’s view that non-um- belliferous inflorescences are in all cases more primitive than umbellif- erous ones. Rigid adherence to this view obliges him to break up genera that are homogeneous except for inflorescence structure. Smith and Stone (1968) cite some of the weaknesses of Hutchinson’s arrangement, point- ing out the contradictions that arise in the genera Polyscias and Plerandra, among others, if characters of the inflorescence are overstressed, and Bernardi (1969) gives similar arguments against Hutchinson’s treatment as it pertains to Madagascan araliads. Perhaps no genus confounds Hutchinsen’s emphasis on the inflorescence more effectively than Meryta, for the plants are dioecious, with staminate and carpellate inflorescences sometimes es markedly Hutchinson’s key (1967a, p aaa ap- = EYDE & TSENG, ARALIACEAE 221 1971] ‘(9x9} aa) SIOYINe JayIv JO avarpuvIa[q ay) SeprAIp 1 yey) St WeASEIpP ay) Jo oINWeey [edioULId Y “UOTTPUOD paAuiap B aq 0} savadde spejed yo uorjyeotquit ‘avaorrery oy) UITIIM, ‘Jedievd yoRa 10} saypung jeUaA jo ated v YIM ‘snold 1SVII0) SIABIT dNOYD IWWALSIINV xpundowealy Dipubilald * ah & enyiubeiere £ wnwuedsoiog D204IOIIO™ 9" SNOsaAjOd ) ~ S4944UD posoquinyD.~g D191jJ9YIS $ Ss : 2 | sndip2opohw pDIayjOBAzZIg : fPuPeses ‘doijayxeeg J (snosawi—g ) 4914}j9425 s9ypBeiBas pup ( 84D]|89d109~-Z) BAe |s39N45 - “vw eee fee y ae pIuoSNH \Y et? sy"? y sp12sAjog GNNOdWO)D ATILVNNId S3AVI1 _——- |___. eee lt QNNOdWOD YO G3801 AldLVYW1Vd S3AVI1 222 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 assigned to different tribes. Yet these two genera are anatomically alike (Viguier, 1906, p. 136); they both occur in New Caledonia; and they share the 8-chambered condition of the anther found nowhere else in the family (Viguier, 1906, Figures 26 & 41). Further evidence of the artificiality of a group based on polyandry alone is that it must include genera with pinnately compound leaves as well as genera with palmately compound or palmately lobed leaves. A more natural classification of the Araliaceae would result, we believe, if the primary division were to separate the palmate-leaved genera from the pinnate-leaved genera, and if this separation were continued wherever possible throughout the family. Ficure 9 illustrates in a general way what we have in mind. The diagram includes only genera with primitive flowers plus a few others selected for their special interest. To include additional genera at this time would be unduly speculative; furthermore, we have no desire to duplicate the entanglements achieved by Harms (1894, p. 23) and by Viguier (1906, p. 176) when they attempted to link all genera of Araliaceae diagrammatically. Our diagram breaks up the Plerandreae of Bentham and Hooker (1867) and of Viguier, making the component genera basal or near-basal in two main evolutionary lines. Plerandra and Tupidanthus, with palmately compound leaves, are near the base of an evolutionary line leading through the species of Schefflera with polymerous flowers to Schefflera species with 5-merous flowers and then to those with a 2-merous gynoecium. Tetra- plasandra, with its close allies Reynoldsia and Peekeliopanax, is near the base of an evolutionary line leading through Gastonia to Polyscias and its derivatives, all with pinnately compound foliage. For various reasons, some of which are mentioned elsewhere in this paper, it seems likely that Stilbocarpa, the other a more ancient line leading to Eremopanax and the M yodocarpus group. While arguing for greater emphasis on the form of the foliage — pin- nate versus palmate — in subdividing the Araliaceae, we remain aware that certain exceptio has simple leaves,7 If th Lace) complicate a subdivision of the family in which the arrangement of €atlets is given primary importance. * Simple F ene It is — rei! re be ignored in this discussion unless they are palmately lobed. repeatedly in th ern leaves with entire margins or dentate margins have evolved © Araliaceae, for they occur in very dissimilar species. 1971] EYDE & TSENG, ARALIACEAE 223 Position of ventral bundles and its significance. — From what is known of evolutionary trends in angiosperms, it is reasonable to as- sume that the syncarpous, inferior gynoecium of Araliaceae evolved from an apocarpous superior gynoecium, as we have indicated diagrammatically in Figure 10. Other changes shown in the figure are: reduction in the number of carpels, reduction in the number of ovules per carpel, homo- carpellous union of ventral bundles, and heterocarpellous union of ven- trals If ventral bundles were always positioned with diagrammatic sym- metry and if the position, once established, could not change, taxonomists would have a superb tool for classification. Species with homocarpellous ventrals would perforce be assigned to a different evolutionary line from those with heterocarpellous ventrals. In reality, however, such a simple interpretation cannot be presented without qualification. One complication, already mentioned under OBSERVATION S, is that the locules occasional- ly outnumber the ventral bundles; in such cases the designations “homo- carpellous” and “heterocarpellous” may be difficult or impossible to apply. Another complication is that there are taxa in which the apparent posi- tion of the ventrals changes with the level of the cross section. In scan- ning a series of cross sections from Hedera or from Fatsia, for instance, one observes that the ventral bundles are heterocarpellous through the lower levels of the gynoecium but converge in the upper half of the gynoecium into a central plexus; just below the insertion of the ovules they become reestablished as discrete bundles in the homocarpellous position. Singh’s (1954) drawings of transverse and longitudinal sections show these changes very nicely for Hedera nepalensis. The diversity of ventral vasculature within Oreopanax and Acantho- panax is the most bothersome complication of all. Oreopanax has not been monographed. It appears to us to be an artificial genus, and this could account, at least in part, for its vascular heterogeneity. Acantho- panax, on the other hand, has had the attention of Harms (1918) and of Li (1942). We are satisfied that in this case the vascular variations oc- cur within a group of related species. Among the Group B species of Acanthopanax that we have examined, A. henryi and A. leucorrhizus seem least aberrant: the ventral bundles are in the heterocarpellous position except for a narrow asymmetric-to-homocarpellous region in the vicinity of ovular insertion. In A. senticosus, the ventrals are heterocarpellous below the locules, homocarpellous near the ovules, and asymmetrically ar- ranged at intermediate levels. The ventral supply of A. setchuenensis and A. giraldii can be either heterocarpellous or homocarpellous; that is, both patterns are found in the same species. In this respect, these two species differ from all others in our study. The evolutionary causes for this variation in the ventral vascular sup- Ply probably involve genetic modification of the direction of vascular differentiation. Data on vascular differentiation in flowers are scanty. It is known, however, that the procambium develops acropetally and con- tinuously in some carpels, including those of certain primitive angiosperms 224 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 __ FIGURE 10. Origin of homocarpellous and heterocarpellous ventrals accord- i A) Polymerous, apocarpous more ovules per carpel. (B) with united carpels, inferior ovary, 1971] EYDE & TSENG, ARALIACEAE 225 peered 1959, 1961; Tucker & Gifford, 1964). In the gynoecia of other ecies, however, procambial differentiation is basipetal or bidirectional, p siaadenten by Esau, 1954, and Tucker, 1959; see also Paterson, 1961). We conjecture that the ventral bundles of primitive Araliaceae developed acropetally and that subsequent genetic changes have introduced vary- ing degrees of basipetal differentiation — from the ovule toward the re- ceptacle — in certain taxa. Basipetally differentiating procambial strands would tend to be in the homocarpellous position because of the more or less median insertion of the single fertile ovule, and the junction of these homocarpellous strands with acropetally differentiating heterocar- pellous strands would necessarily be a region of asymmetry or anastomosis. This conjecture is in keeping with our observation that the ventral bun- dles are always heterocarpellous in Araliaceae with polymerous flowers (Trevesia excepted), whereas the mixed or transitional condition occurs only in a few Group B taxa. The existence of these few transitional forms suggests the manner in which the homocarpellous pattern could have arisen here and there among the 2-carpellate (Group C) Araliaceae. Similar reasoning can be invoked to explain the diversity found in the Umbellif- erae, in which the ventral bundles can be in either position (Jackson, 1933; Tseng, 1967), apparently without regard to major taxonomic groupings, This is not to say that the hypothetical sequence in Ficure 10 is com- pletely incorrect or that the position of the ventrals has no taxonomic value. We suspect that the change from B to C illustrates fairly accurately the origin of the homocarpellous condition in Trevesia, because the flow- ers of this genus do not seem greatly advanced, and because one may find specimens in which the original paired condition of the ventrals is partially retained (Eyde, 1967, Figure 11). These pairs are always in the homocarpellous position, with no indication of secondary complexity; therefore, we have constructed FicurE 9 so as to make Trevesia an early evolutionary offshoot of the Schefflera line. For like reasons FIGURE 9 shows the early divergence of a line leading to Aralia, Stilbocarpa, and Panax (perhaps Harmsiopanax belongs here too). Homocarpellous vas- culature seems primary in this case——we have seen no exceptional or ambivalent specimens — and agrees with other features such as the form of the style and the estivation of the corolla in linking these genera to- gether. Parenthetically, it should be added that differences in ventral vasculature may have value for distinguishing genera and species of Araliaceae even if the phylogenetic concepts expressed in Ficure 10 should eventually prove false. Additional characters.— While recording the vascular characters listed in Tastes 2 and 3, we scanned our serial sections for other features single ovule per carpel; ventral — still in pairs. At still later stages, the paired ventrals unite as homocarpellous ventrals (C) or as heterocarpellous: ventrals (D). This pa explain the origin of some, but not all, ventral vascular systems in Araliace 226 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 TABLE 2, Gynoecial vasculature and meristic characters in selected Araliaceae Symbols in the Dorsats column show whether the dorsal bundles are predominantly separate from the peripheral bun dies, —; united w with the peripherals, +-; or in nter- paired ventrals, the ultimate > of union being a solitary central strand, ce. Pair- ing or union of ventrals may be homocarpellous, ho, or heterocar pellous, he; see Ficure O. Other o> cuntiins identify anomalous vasculature, ag at asymmetry, asym; and items that are not applicable, NA. Further explained in tex VENTRALS Dorsats Union Posrrion CorottaA ANDRoEcIUM GYNOECIUM Group A, FLOWERS MORE OR LESS POLYMEROUS Tupidanthus ntl heey - +? he NA many up to ns Plerandra vitiens _ — he 5 (SCA. ATS iat Plerandra facie + + he “4-6” “50-75” apes Plerandra solomonensis — he many Peekeliopanax spectabilis — +P he 6-12 25-66 6-22 Boerlagio novo-guin + int he 11 11 12 Bocragiodendron tetrandru ai = he 4 4 Reynoldsia sists = a he 10-11 10-11 20-22 Reynoldsia sandwicensis _ + he 8-12 8-12 7-12 Indokingia crassa _ int he NA upto 100 14-18 Gastonia — int he ? 15 Gastonia papu — int he - 6-12 Gastonia eae —? + he 10-12" 19-12” 9-12 Trevesia — + oO 7-12 7-12 7~12 Sidodendron a a a he 9-10 9-12 9-12 egalopanax _ + he 7-10 7-10 7-10 Scheflera Sibiidiesing — “- he (anom) 8-9 8-9 8-10 Schefllera digitata + he “5 or “Sor 6-10 (or more” more” more?) Scheflera khasiana _ + he (anom) 5-6 5-6 5-9 Schefflera racemosa =e + he 59 ay 5-9 Group B, rLowers More oR LESS 5-MEROUS Schefflera decaphy : + . he 5 5 5 Schefflera icuas ae, — + he 5 5 5 Scheflera bangii - + he 5 5 5 Scheflera robusta ee aa 5 5 5 Schefflera s (Hladik al. 5 5 5 Scheflera umbella +? ‘3 he 5 5 3-5 Oreopanax capitatus (Baumann- Bodenheim, 1 ho 4-6 4-6 5-10 Oreopanax obtusifolius - - h 4-6 4-6 5-7 ‘Oreopanax xalapensis + he 4-6 4-6 4-6 ‘Oreopanax nitidus int + he 5 5 5 Oreopanax ig + he? 4 5 5 Oreopanax aff oerstedianus — a ho? 1971] EYDE & TSENG, ARALIACEAE 227 TABLE 2. Gynoecial vasculature and meristic characters in selected Araliaceae ead continue VENTRAL Dorsats UNION PosItIon CoroLLA ANDROECIUM GyYNOECIUM Oreopanax floribundus int + he 5 5 4-5 endropanax chevalieri — + he 5 5 5~6 Dendropanax querceti — oh he 5 4-5 oe gonatopodus int ao he =] 5 5—6 Aralia e + ho 5-6 5-6 5-6 Aralia ae nOSa + ho 5 5 5 Dizygotheca elegantissima + he 5 5 5 Pentapanax castanopsidicola _ + he 5 5 5 Pentapanax subcordatus 4. he 5 5 5 Pseudopanax crassifolius + he 5 5 5 Fatsia japoni int + he 5 5 5 Hedera helix int a he 5 5 3~5 Hedera ne palensis (Singh, 1 + he s 5 5 Acanthopanax sentic + asym 5 5 5 Acanthopanax leucorrhizus int + he 5 5 5 Acanthopanax setchuenen. + int heorho 5 5 5 Acanthopanax henryi int + he 5 5 3-5 Acanthopanax giraldii ths + heorho 5 5 3~5 Gamblea ciliata + a 4-5 4-5 3-5 Cheirodendron kauaiense + + he? 5 5 3-5 Cheirodendron trigynum + + e 5-6 5-6 2-5 Meryta sinclairii + anom 4-6 4-6 3-6 Stilbocarpa polaris of + ho 5 5 3-5 Polyscias nodos + 4 he 5 5 5 issodendron australianum + + he 5 5 3-5 Group C, GYNOECIUM 2-LOCULAR Polyscias multijuga + + he 5-6 5-6 2 Polyscias grandifolia +b. he 5 5 2 Tieghemopanax elegans (Baumann, 1946 + + he 5 5 2 Panax trifolius + ho 5 5 3 Panax quinquefolius int of ho 5 5 2 opanax horridus int a he 5-6 5-6 2-3 Didymopanax attenuatus — of he 5 : 2-3 ymopanax morototoni — + he 5 5 2 Brassaiopis hispida int + ce 5 5 ~ Macropanax undulatus int + ce 5 5 - Pseudopanax arboreus nt +. he 5 5 2 Pseudopanax davidii + ce 5 5 2 Acanthopanax gracilistylus — + he 5 5 . Acanthopanax sessiliflorus + he 5 5 - Acanthopanax korean int + ce 5 5 2 Acanthopanax 5s int sf. he 5 5 2 Acanthopanax trifoliatus int + he 5 5 2 canthopanax innovans - a ho 5 5 2 Kalopanax pic en of. he 5 5 2 Merrilliopanax listeri int + he 5 5 . 228 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Taste 2. Gynoecial vasculature and meristic characters in selected Araliaceae (continued) VENTRALS Dorsats UNION PosITION CoroLttA ANDROECIUM GYNOECIUM ussonia paniculata os he 5 5 2 Heteropanax chinensis + + he 5 5 2 Heteropanax fragrans o + he 5 5 2 Astrotricha floccosa + + he 5 5 2 poh ta celebica + + ho 5-6 5-6 2 Mackinla issued (Baumann- Bodenheim, 1955) + a 0) 5-6 5-6 2 Harmsiopanax harmsii _ ho? 5 5 2 Harmsiopanax leatus (Philipson, 1970) = + ho 5 5 2 Delarbrea oe — - he 5 5 2 Delarbrea parad ? + he 5 5 2 M yodocarpus i aiaaaiie int aa ho? 5 5 2 Porospermum michieanum + + hoorce 5 5 2 Stilbocarpa lyallii ~ + 4-5 5 2 Tetrapanax papyriferus _ a he 4 a 2 Group D, GyNOECIUM 1-LOCULAR Diplopanax stachyanthus — + NA 5 5-10 1 Arthrophyllum ahernianum — + NA 5 5 1 Arthrophyllum diversifolium — + NA 5 5 1 remopanax angustata = + NA 5 5 1 Aralidium pinnatifidum — + NA 5 5 1 TABLE 3. Gynoecial vasculature and meristic characters in Tetraplasandra [Symbols and abbreviations as in TABLE 2] VENTRALS Dorsats Union Posttion CoroLLA ANDROECIUM GYNOECIUM ig! sca = int he 8-9 20-40+ 9-11 T. hawa int int 8 20-30 6-13 r piceeabogg anom “56” “10-20 rarely5)” 11-14 T. —- int + he wing fo 5-7 T. waialealae int a he 5-7 20-28 pis poe ier ae as he 6-9 2-5 - = seine ~ + he 6 12-18 3-4 - _ ra + + he ‘6 5-8 2-5 i TENS: ~ he 5-6 10-14 3 . micrantha + + he 8 TG 2-4 T. gymnocarpa A es he 5-9 5-9 2-5 that would distinguish taxa or suggest evolutionary trends. We especially wanted to learn whether secr secretory canals, known to have taxonomic utility in the Umbelliferae, might aid in classifying the Araliaceae. Since many of our observations are based on herbarium flowers, with no con- 1971] EYDE & TSENG, ARALIACEAE 229 trol over developmental stages or quality of preservation, it would be pointless to attempt a description . secretory structures of each species, but a few generalizations can be made. We found, for instance, that the an of secretory canals can differ greatly in gynoecia of different species, even closely related species. Hetero- panax fragrans has an abundance of canals, whereas H. chinensis has few. Canals are abundant in flowers of Acanthopanax senticosus, rare in A. giraldi and A. sessiliflorus, confined to the vicinity of major vascular bundles in A. koreanus, and lacking in A. leucorrhizus and A. setchuen- ensis. Secretory canals have a prominent place in the phylogenetic specula- tions of Baumann and Tikhomirov. Tikhomirov (1961) called the secre- tory structures of Araliaceae ‘“‘rebernye kanal’tsy” (costal canals), indi- cating that they are found only in association with vascular bundles. If this were a reliable generalization it would tend to support his contention that Hydrocotyle and Centella, with closely associated canals and bundles, resemble Araliaceae more than they resemble apioid Umbelliferae.* We Oplopanax, Panax, Tetrapanax). Baumann (1946) concluded from his studies that the ancestral Araliaceae were without secretory structures (see his summarizing table), but we see no evidence in support of this view. All of the polymerous flowers that we examined contained secretory canals, and the canals are scattered; that is, they have no noticeable tendency to accompany major vascular bundles. We failed to find secre- tory structures in the flowers of a very few species belonging to Groups C and D; viz., Diplopanax stachyanthus, certain species of Acanthopanax, Aralidium pinnatifidum, and Stilbocarpa lyallii.® Aralidium and Stilbocarpa would be considered advanced by any reasonable standards, Our evidence indicates very emphatically that the gynoecia of ancestral Araliaceae were well supplied with scattered secretory canals, that the canals have been lost in a few of the derived taxa, and that they have been localized in others, including the Umbelliferae. In umbelliferous fruits the canals may be confined to ridges formed by major vascular bundles Be candies) or to the grooves alternating with the ridges; some Umbelliferae have canals in both positions. Many older works call all of these canals vittae, whereas modern authors tend to use a separate Tikhomirov also claimed, as Hakansson (1952, p. 43) did, that the locules of mission of pollen tubes (see also Hakansson, p. 3 *Grushvitskii et al. (1969) also found no secretory send ga in se lyallii, but attributed this to the poor quality of their materi e are somewha uncertain about Diplopanax, for our observations were hindered i a darkly staining substance in the floral tissues. 230 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 term for canals that run with, and external to, the peripheral vascular bundles: extravasale Sekretkanaéle (Baumann, 1946); ékstrafastsikul- yarnye kanal’tsy, rebernye kanal’tsy (Tikhomirov & Galakhova, 1965); companion canals (Tseng, 1967). Companion canals develop somewhat differently from canals that do not accompany bundles (Kovacs & Sar- kany, 1968). This suggests that the two kinds of canals could contain different substances. It should be noted, however, that Lassanyi & Lérincz (1970) report no histochemical differences between “pith canals” and “canals of the phloem” in stems of coriander. Flowers of Araliaceae differ not only with regard to the distribution of brea and Porospermum, as in the related genus Myodocarpus (Baumann, 1946, Figure 2), druses are especially abundant in the ovarian septum. A similar concentration of druses occurs in the commissural region of certain Umbelliferae (Tseng, 1967, p. 39), Our survey has not by any means exhausted the potential taxonomic utility of the gynoecial vascular system. Each carpel of an araliaceous synoecium commonly has two or more conspicuous bundles in addition to the dorsal bundle and the ventral bundles. In some taxa, these inter- fae Pe aa are arranged in a distinct circle around the endocarp ceae, such : jr Tetrapanax). Some of the other 2-carpellate Aralia- lar HA a a floccosa, show the typical umbelliferous vascu- inferior fei — ‘ re 10 conspicuous bundles in the wall of the . ane Clearly representin i dle eo a peripheral bundle. P & the union of a carpel bun ed i : cumin 7 to leave the subject of gynoecial vasculature without @ ‘eature that may prove useful in distinguishing the Um- 1971] EYDE & TSENG, ARALIACEAE 231 belliferae — at least the apioid Umbelliferae — from the Araliaceae. In both families, it is common for intermediate carpel bundles (lateral bundles of Umbelliferae) to turn inward through the upper part of the ovary and join the ventral bundles. The junction may take place at the level of ovular insertion (Panax) or very close to this level, In Araliaceae, the usual position of the junction is above the attachment of the ovule, but in the apioid Umbelliferae the junction is usually below the attach- ment of the ovule. The characteristic pattern for Araliaceae is shown in Philipson’s (1967, p. 147) diagrams of Hedera helix and Pseudopanax arboreus; the pattern for Umbelliferae is exemplified by Jackson’s (1933, p. 123) diagrams of Osmorhiza longistylis. Ancestry and affinities of the Araliaceae-Umbelliferae complex. — Taxonomic opinion is divided as to the nearest allies of the Araliaceae and Umbelliferae.1° Thorne (1968), Hutchinson (1967a, 1969), and Takhtajan (1969) adhere to the traditional concept of a close tie with Cornaceae. Others oppose this view (Eyde, 1967; Cronquist, 1968; Hegnauer, 1969; Philipson, 1970). We remain with the opposition, for our evidence indicates that the ancestral Araliaceae had, in addition to polymerous flowers, compound (probably pinnately compound) leaves and a normal ventral vascular system, features not found in Cornus or in any of the genera that can confidently be allied with Cornus. According to Cronquist (1968, p. 278), the Araliaceae ‘‘would be per- fectly at home in the Sapindales”’— and close to Burseraceae —if the Ovaries were superior instead of inferior. At his suggestion, we ran Tetra- plasandra gymnocarpa through Hutchinson’s (1967b) key to plant fam- ilies: because of its hypogynous flowers, the species came out in Bursera- ceae as Cronquist predicted it would. Of course, the key could be adjusted quite easily to place T. gymnocarpa where it belongs, and we mention this point only to emphasize the similarity of certain Aralia- ceae to the Burseraceae. Secretory canals provide one link between the two families, pinnately compound leaves another. The ovules of Bursera- ceae are oriented in the same manner as those of the Araliaceae, but there are usually two fertile ovules per locule, a more primitive condition than that found in the Araliaceae. In other respects, the Burseraceae are more advanced. Vessel members of Burseraceae are always simply per- forate (Metcalfe & Chalk, 1950), and none of the species have polymerous flowers. The seeds lack endosperm. Also, the ventral carpel bundles of the few burseraceous flowers that have been examined anatomically (Narayana, 1960a, 1960b) are united in the heterocarpellous position. From the mixture of primitive and advanced characters, it appears that neither family can be derived from the other. If the Araliaceae and the view, citing chemical similarities. The differences in floral structure separating Pittosporaceae from Araliaceae and Umbelliferae are so great, however, that we can- not envision a close common ancestry. 232 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Burseraceae are allied, as the evidence indicates, both must have arisen from a common ancestral group. It then follows that the Araliaceae are also allied with the Rutaceae, for the affinity of Burseraceae to Rutaceae is well established. The addition of Rutaceae to the discussion makes Araliaceae seem even more “at home in the Sapindales” (Rutales of some authors), for a few of the Rutaceae have moderately polymerous flowers, and scalariform perforation plates have been observed in rutaceous woods (Solereder, 1908). The ventral bundles of Rutaceae with syncarpous gynoecia may be separate (Coleonoma) or united in either the homocarpel- lous (Skimmia) or the heterocarpellous position (Ruta; see illustrations in Gut, 1966). The likelihood of a connection with Rutaceae recalls a suggestion put forward by Ehrlich and Raven (1967). Citing Dethier’s ( 1941) finding that certain Rutaceae and certain Umbelliferae contain the same essential oils attractive to swallowtail butterflies, these authors suggested that the two families might share “a closer ancestral tie than had been suspected.” Although a chemical link between Rutaceae and Umbelliferae seems to Support our arguments, it also raises a question: Why have these es- sential oils not been found in Araliaceae? (At least one of them, methyl chavicol, occurs in Burseraceae; Roberts, 1923.) Of course, the answer may simply be that no systematic search for these compounds has yet been undertaken in the Araliaceae. W. R. Philipson, whose recent work on Araliaceae has been directed least some of Philipson’s examples actually represent a secondary loss of umbels, For instance, Li’s (1942 ) illustration of the spike-like inflor- Ww . ? ‘ higenag that the racemose or partially racemose inflorescences occurring genera may never have passed through an earlier umbellate stage. 1971] EYDE & TSENG, ARALIACEAE 233 We suspect, however, that some of the ancestral Araliaceae had more loose- ly branched inflorescences than the term “racemose” would indicate. The panicles of Reynoldsia pleiosperma and the more or less paniculate in- florescences of certain Tetraplasandra species suggest the manner in which various early Araliaceae may have borne their flowers. Our conclusions concerning the ancestral characters of the Araliaceae agree well with current phytogeographic concepts in that the taxa we consider most primitive are all tropical. Moreover, many of them are found in southeast Asia and on islands of the western Pacific, that is, in the region believed to be “the cradle of the flowering plants” (Takhtajan, 1969; Smith, 1970). SUMMARY This investigation was undertaken in response to conflicting claims concerning the relative antiquity of polymery and 5-mery in Araliaceae. Assuming the degree of fusion of vascular bundles should be least in the least advanced flowers, we surveyed gynoecial vasculature throughout the family using standard clearing and sectioning methods. If the initial as- sumption is correct our findings show that 5-merous flowers are derived from polymerous flowers. Older taxonomic treatments err, however, in making Tupidanthus most primitive because of its highly polymerous flowers; in this genus, as in certain species of Plerandra, floral parts have increased. Furthermore, most treatments mistakenly place all polyandrous genera in the same subdivision of the family. We believe the polyandrous forms belong to two main evolutionary lines, one with pinnately compound leaves, the other with palmately lobed or palmately compound leaves. Position of ventral carpel bundles is of theoretical interest and may prove to be an aid to classification. In most Araliaceae the ventral bundles al- ternate with the locules. There are several taxa, however, in which the ventrals are aligned with the locules, and the genus Acanthopanax in- cludes species that exhibit both patterns, apparently indicating the man- ner in which one type of ventral vascular system can be converted to the other. We judge that the ancestral Araliaceae had pinnately com- Pound leaves, that at least some of the inflorescences were paniculate, and that the flowers were well supplied with secretory canals. We agree with others who have suggested an evolutionary link with Burseraceae and Rutaceae. ACKNOWLEDGMENTS We acknowledge with pleasure a profitable exchange of ideas and in- formation with D. G. Frodin and W. R. Philipson. Helpful suggestions were also obtained from conversations with A. Cronquist and H. E Robinson. For pickled specimens, we thank J. C. Cuatrecasas, F. R. Fos- berg, A. Hladik, F. M. Hueber, E. McClintock, W. Soegeng-Reksodi- hardjo, and W. L. Theobald. Tseng obtained much of the herbarium ma- JOURNAL OF THE ARNOLD ARBORETUM [voL, 52 234 ‘sjeyad ayeouquar Aq mypsp 0) payury *(32] “feat [ ATWO saumMawios) aze[oroy-¢ ‘saquity> Apoowy “erpeajsny fds 7 ‘SDIISKIOT PUL UOsPUapOsstY UIaAIaq B}UIPIULaIUy “vaUINT) MAN f'ds | "SpIIskjog 0} payejay “s— Ayawmog {dds z 30 | ‘smtaskjog 0} payepy ‘awoseSepepy ‘dds ¢ YVTASIOT-Z WAIDFONAD ‘9D anor “mantporrs “Opnasd pue vkojuryan py ut se ‘spejad pamela “eIuOpayey) MAN S-dds + ‘DISD OF payejay “s— uruog ‘dds z ‘SafAQs ayeuuOD ‘saavat azeforoy-¢ ‘ensutuag Avjepy ‘iodds z SQOWAN-¢§ SSAT XO AAOW SYAMOTA ‘gq aNoUr) *(dnoiz_ umnyyAyd “BpOls) vaayfayrg YIM sty) sazrun (st) uIporg ‘saqjayodag fds 7 “Spay Ul aISsas SIOMO]T “seIINJOP ‘BauIny) Many {dds z DISALAdT OF pPazejar Ajqeqoig “siaMoy adie] ‘saqoy xAyeo Suoy ‘S@A RI] punodwios Ajayewyed yy 9a} 4yyoud yfewg “vung qaddq ‘ds { S[oyM ¢ Aq vIa4I03KzIq Wo] payeiedag ‘vruopayea Man ‘dds z 10 I ‘SOARIT paqoy AyezeUL -[ed ‘saqoy xAyeo papraip Aq Daépuddafd Woly payeredag ‘uryuoy {ds [ SQOXAWATOd SSAT YO IMOW SAaMOTA ‘Y anoXy z S$ Si z S S z § S ¢-z S § t-Z s S s $ s $ s s OI-s OI-8 OI-s OI-8 OI-8 OI-8 £I-8 $ s OZ-ST ST S or Aueut Or Di}ams9y30 py DIY 904quapundrsaw[od Dpasatunog round opo17¢ wnyojiag od y Dis70{ ourwog sisdosapayy xoung oa uOjKXOWSE OPUINGPOO 09943029Q sisdospusa}g WOATOWONAS) WAIAOMINY VITOXOD S}USUTWOD PUP S19}2eIeYD INstIIUT :AQAINS UT papnypouT jou viwUIy “p ATAV EYDE & TSENG, ARALIACEAE 235 1971] ‘sp19skjOg Wo01y paauap Ajqissod pue 0} payepay “(671 “d ‘6967 ‘UOsUIyDINF_) aus Isaz0yur [eNsN -un ‘saavay punodwoo Ajayeuutg “iedsesepeyy ‘(9961 ‘tpreuseg) “dds ¢ YVTNIOT-T WAIDFONAD ‘q] dno) ‘OL6I ‘uosdiryg aas ‘ynay jo UOT}IAS SSOID 1OJ “SAAvaT a[dUIIS YIM 9a] paue [[eUg “yeIag ‘*ds | ‘mnjojadoigy ‘pXvjuryoo py 0} payejei Ajpasoddng ‘JIMOY JO UOTL[NINAe JeNsNuN ‘saaval sjeuUlg “eIUOpaley MeN ‘ds | “DIMOs “NJ WOIJ pazeFaaBag ‘saavay ayeuyed paqoj-¢ “eouyy yyNog ‘ds | ‘(@z “d ‘9681 ‘SULILET) DYDAP PUL DIaMsayjJoFY 0) Paye[aYy “eyeIysny ‘ds | sndap20ygny DIMap 40 Af uniporasopnas q DiDsDuUDW IIS DipDavjoydsD 236 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. 52 terial from Harvard University while he was a Mercer Research Fellow at the Arnold Arboretum; courtesies extended by the directors and the curators of the Arboretum and of the Gray Herbarium are greatly appre- ciated. Drawings are by Anne V. Donovan (Ficures 5, 6), Diane Rob- ertson (Figures 7, 8), and C. H. Reinecke; photographs by Smithsonian photographers Andrew Wynn (Ficure 3) and Victor Krantz. The Na- tional Science Foundation and the Smithsonian Research Foundation pro- vided financial support through grants to Tseng (GB7988) and Eyde (SG625005) respectively. LITERATURE CITED Note added in proof. While reading galley proofs for this article, we received two additional contributions by W. R. Philipson, “The Malesian species of Gas- tonia” (Blumea 18: 490-495. 1970) and “‘A redefinition of Gastonia and related (Harms) Philipson. If Tetraplasandra is to include only Hawaiian species, it should occupy a slightly more advanced position than Gastonia in our FIGURE 9; otherwise our conclusions are not affected. BAUMANN, M. G. 1946. Myodocarpus und die Phylogenie der Umbelliferen- Frucht. Ber. Schweiz. Bot. Ges. 56: 13-112; Tab. 1, Taf. 1-6. . 1955. Ableitung und Bau bicarpellat-monospermer und pseudomono- carpellater Araliaceen- und Umbelliferen-Friichte. Jbid. 65: 481-510. wi G., & J. D. Hooxer. 1867. Araliaceae. Genera Plantarum 1: 931- ee L. 1966. Species novae (13) et nomina mutata (2) in Araliae amilia insulae Madagascariae. Ber. Schweiz. Bot. Ges. 76: 352-395. . 1969. Araliacearum Madagascariae et Comores exordium. 1. Revisio t taxa nova Schefflerarum. Candollea 24: 89-122. aed . 1965. Island life. Nat. Hist. Press, Garden City, New York. ST, A. 1968. The evolution and classj i ; j Houghton Miffiin, Boston, nd classification of flowering plants. DELPINO, F. 1883. Teoria generale della fillotassi. Atti R. Univ. Genova, vol. 4, parte 2, ‘ Deruier, V. G. 1941 Chemical f ini Fes ee : actors determining the choice of food plants se Agate Am. Naturalist 75: 61-73. : ES, A. J., & L. H. MacDaniets. 1947. An j i : Ed. 2. Mc Geioi-tilll New Yoon - An introduction to plant anatomy. HRLICH, P, R. & Po Ra i : 216(6): 104.111. VEN. 1967. Butterflies and plants. Sci. Am. Bis pe 1954. Primary vascular differentiation in plants. Biol. Rev. 29: Eype, R. . é H. 1967. The peculiar gynoecial vasculature of Cornaceae and its Be opin pimp acs Phytomorphology 17: 172-182. [Published 1968.] aes -\. LSENG. 1969. Flower of Tetraplasandra gymnocarpa: hypogyny ©pigynous ancestry. Science 166: 506-508. 1971] EYDE & TSENG, ARALIACEAE 237 GrusHvitskii, I. V., V. N. Trxuomirov, E. S. AKseNnov, & G. V. SHIBAKINA, 1969. Sochnyi plod s karpoforom u vidov roda Stilbocarpa Decne. et Planch. Pp eat Byull. Moskov. Obshch. Ispytat. Prirody, Otd. Biol. 74(2): 6 Gor, B..}. Fie Beitrage zur Morphologie we Nh alana und der Blii- tenachse einiger Rutaceen. Bot. Jahrb. 85: 47. HAKANssSON, A. 1952. Seed development in eine tenera. Bot. Not. 1952: 33-4 Hat, B. A. 1954. Variability in the floral anatomy of Acer negundo. Am. Jour. Bot. 41: 529-532. Harms, H. 1894/1897. Araliaceae. pp. 1-62. Nat. Pflanzenfam. III. 8: 1-62 : 1896. Zur Kenntnis der Gattungen Aralia und Panax. Bot. Jahrb. 23: 1-2 ———. 1907. Araliaceae. Nachtrage zu Teil III. Abt. 8: Nat. Pflanzenfam. (Erganzungshefte II) 253~255. 1918. Ubersicht iiber die Arten der Gattung Acanthopanax. Mitt. Deutsch. Dendr. Ges. 27: 1-39. ———. 1938a. Neue Araliaceen aus Papuasien. Bot. Jahrb. 69: 277-283. . 1938b. Zur Kenntnis von Meryta sonchifolia Linden et André und einigen anderen Arten der Gattung. Notizbl. Bot. Gart. Berlin 14: 315- S215 HEGNAUER, R. 1969. Chemical evidence for the classification of some plant taxa, pp. 121-138. In: J. B. HarporNe & T. Swain, eds. Perspectives in phytochemistry. Academic Press, New Yor HoucutTatine, H. B. 1935. A developmental apatyaa of size and shape in tomato fruits. Bull. Torrey Club 62: 243-252; 14. HUTCHINSON, J. 1967a. The genera of flowering plants Dicotyledones, vol. IT. Cla rendon Press, Oxford. [Araliaceae, pp. 52-81, 622-624.] 1967b. Key to the families of flowering ae of the world. Claren- don Press, Oxfor 1969. Evolution and phylogeny of flowering plants. Academic Press, London/New Yor Jackson, G. 1933. A study of the carpophore of the Umbelliferae. Am. Jour. Bot. 20: 121-144 Jay, M. 1969. Chemotaxonomic researches on vascular plants. XIX. Flavo- noid distribution in the Pittosporaceae. Bot. Jour. Linn. Soc. 62: 423-429. Kovacs, A., & S. SARKANY. 1968. Formation nf volatile oil ducts in the de- veloping pistil of Heracleum mantegazzianum Somm. et Lev. Acta Agron. Acad, Sci. Hungaricae 17: 359-372. LassAny1, Z. & C. Lérincz. 1970. Test on terpenoids present in parts of Coriandrum sativum L. Il. Histology of the developing structures of Coriandrum sativum L. var. “Lut” and eta of volatile oil canals. Acta Agron. Acad. Sci. Hungaricae 19: 25-32 Li, H.-L. 1942. The Araliaceae of China Sargentia, N Lucxwn, in kc. Se, ae evolution of the wenn tomiato. Jour. Roy. Hort. Soc Pts 19-2 Metcatre, C. R., & L. cniie 1950, Anatomy of the dicotyledons. Clarendon ee Oxford. Narayana, L. L. 1960a. Studies in Burseraceae—I. Jour. Indian Bot. Soc. 39; 204-2 209, 2 vols. 238 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 ——. 1960b. Studies in Burseraceae —II. bid. 39: 402-409. Paterson, B. R. 1961. Studies of floral morphology in the Epacridaceae. Bot. Gaz. 122: 259-279. Puitipson, W. R. 1951. Contributions to our aay of Old World Araliaceae. Bull. Brit. Mus. (Nat. Hist.), Bot. 1967. Griselinia Forst. fil.— anomaly or ah . Z. Jour. Bot. 5: 134-165. - 1970. Constant and variable features of the Araliaceae. In: New re- search in plant anatomy. Supplement 1 to Bot. Jour. Linn. Soc. 63. Aca- demic Press, London Prjt, L. VAN DER. 1969. Principles of dispersal in higher plants. Springer Verlag, Berlin/Heidelberg/New Yor a * 1904. Sul genere T apidunithus. Bull. Orto Bot. Univ. Napoli 2: 145- ican O. D. 1923. Constituents of the essential oil from the gum resin of ph onein serrata, Roxb. Jour. Soc. Chem. Ind. 42 imac section): 486T-488T, Ropricuez, R. L. 1957. Systematic anatomical studies - ne and other woody Umbellales. Univ. Calif. Publ. Bot. 29: SHERFF, E. E. 1952. Munroidendron, a new genus of toes trees from the island of Kauai. Botanical Leaflets 7: 21-24. [Published by the uthor SiIncH, D. 1954, nian morphology and a of Hedera nepalensis K. ach. Agra Univ. Jour. Res. Sci, 3: SmirH, A. C. 1941. Araliaceae. N. Am. i a 3-41. - 1970. The Pacific as a key to flowering plant history. Univ. of Hawaii Harold ‘3 Lyon Arboretum Lecture No. 1. C. Stone. 1968. Studies of Pacific Island plants, XIX. The Aras 0 of the New Hebrides, Fiji, Samoa, and Tonga. Jour. Arnold r 1-501 SOLEREDER, H. 1908. Systematic anatomy of the dicotyledons. Transl. by L. . A. Boodle & F. E. Fritsch. 2 vols. Clarendon Press, Oxford. AKHTAJAN, A. 1969, Ae ig plants — origin and dispersal. Smithsonian Institution Press, Washin THORNE, R, F. 8. oo - a putatively phylogenetic classification of : the flowering plants. Aliso 6: 57 7-66. IKHOMIROV, V. N. 1961. O sistematicheskom polozhenii rodov Hydrocotyle B 46: 584-586. 9 L. i Centella L. emend. Urban & N. GaLakHova, 1965. Material ‘ y po morfologii gruppy Angelicinae. i rage anatomii ploda Angelica sylvestris L. kak lektotipa roda z nee L. Byull. Moskoy. Obshch. Ispytat. Prirody, Otd. Biol. 70(1): sapere C. Anatomical studies of flower and mei in the Hydrocotyl- : oideae (Umbelliterny Univ. Calif. Publ. Bot. 42: 1-79. asin S.C. 1959 Ontogeny of the inflorescence at the flower in Drimys nteri var. chilensis. Univ. Calif. Publ. Bot. 30: 257-336. ———~ 1961. Phyllotaxis and vascul fuscata. Am. i Bat 48: vlan ar organization of the carpels in Michelia & Gtrrorp, Jr. 1964. C a ceolata, “Fhytemotpiney 1 iabac a vascularization of Drimys la 239 EYDE & TSENG, ARALIACEAE 1971] bes R. 1906. Recherches anatomiques sur la classification des Araliacées Sci. Nat. Bot. Ser. 9. 4: 1- 1945. Fasciation in horticaltueal plants with special reference —268. neato Q. 19 to the tomato. Proc. Am. Soc. Hort. Sci. 46: 263 SMITHSONIAN INSTITUTION WINDHAM COLLEGE DEPARTMENT OF BOTANY DEPARTMENT OF BIOLOGY WASHINGTON, D.C. 20560 PuTNEY, VERMONT 05346 240 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 SHOOT GROWTH AND HETEROPHYLLY IN ACER! WILLIAM B. CRrITCHFIELD In Acer and many other woody genera, leaf form changes drastically and often abruptly during the life of the plant. Heterophylly is of two principal kinds in these woody plants: (1) changes during seedling and post-seedling stages, and (2) changes during the development of single annual shoots of adult plants. Type (1) is ubiquitous in Acer; type (2) is a regular feature of some species but uncommon or poorly expressed in others. These two types of heterophylly tend to converge; deviations in leaf shape on adult plants are often in the direction of seedling leaves. This tendency of woody plants to produce reversions, among other aspects of heterophylly, attracted a great deal of attention around the end of the 19th century, and much of our knowledge of these phenomena derives from that period. The principle that ontogeny tends to recapitulate phy- logeny (Haeckel’s biogenetic “law”) was widely accepted by biologists at the time, and variations in leaf form during development offered many apparent illustrations. Partly in reaction to such phylogenetic interpretations of variation in plant form, Goebel ( 1900) formulated the concept of heteroblastic de- velopment, which dealt with the same kinds of variations in ontogenetic and Physiological terms. In plants exhibiting heteroblastic development, the Juvenile and adult stages are markedly different, and the adult some- times produces reversions to the juvenile state. Both features are illus- trated by several common maples (Jackson, 1899). Their seedling leaves 2 ° Cine ot a begun while the author was on the staff of the Maria Moors ing the tenure ri ck Botanical Research at Harvard University, and completed dur- I am grateful to Ral ae tspita Forest Research F ellowship at the same institution. of the manuscript. pi *t, Wetmore and Brayton F. Wilson for their helpful reviews 1971] CRITCHFIELD, ACER 241 biloba and representatives of Betula, Cercidiphyllum, Liquidambar, and Parthenocissus (Clausen & Kozlowski, 1965; Critchfield 1970a, 1970b; Smith, 1967; Titman & Wetmore, 1955). This link between heterophylly and shoot specialization is conspicuous in two maples of eastern North America: Acer rubrum L. (red maple) and A. pensylvanicum L. (striped maple). This paper describes shoot development in these species, sur- veys the patterns of shoot growth and incidence of heterophylly in other members of this genus, and compares Acer to other groups with similar patterns of shoot development and leaf variation. MATERIALS AND TERMINOLOGY The growth of leaves and internodes was measured at 4-day to 4-week intervals on long and short shoots of red and striped maples. Total leaf length was measured in 1957-58; in 1959 blades and petioles were mea- sured separately. Observations of red maple were made on 5- to 10- year-old trees growing in natural stands on the Harvard Forest at Peter- sham, Massachusetts (1957, 1959) or in a plantation of Harvard Forest and other Massachusetts origins on the Case Estates of the Arnold Arboretum at Weston, Mass. (1958). Shoot growth of striped maple was measured in 1958-59 on young trees in two stands on the Harvard For- est. Sprouts were induced to develop by partial stripping of the buds and branches from small trees. Other species of Acer were observed mostly in the Arnold Arboretum. Illustrated shoots and leaves are from Arboretum trees except Acer orien- tale L. (A. creticum L.), which I collected in the mountains of western Crete. Dried specimens were examined in the Gray Herbarium and the herbarium of the Arnold Arboretum. Terminology follows that of an earlier paper (Critchfield, 1960). The early leaves expand when the buds open in the spring; the late leaves are produced subsequently. Transitional leaves, present on some annual shoots, are intermediate between early and late leaves in time of appear- ance, position on the shoot, and form. Leaves are numbered from the base of the annual shoot, and an internode has the same number as the leaf at its upper end. The phyllochron (Bond, 1945) is the time interval between corresponding developmental stages of successive leaves, exclud- ing initiation (to which plastochron refers). The stage on which the phyllochrons are based here is leaf maturation, defined as 90 percent of fi- nal leaf or blade length. Blade length was preferred because of the great variation in growth duration and final length of petioles, particularly in re The term epigenetic is applied to the ontogeny of all leaves that are not preformed in winter buds. In contrast to the distinctly two-stage growth of preformed leaves, epigenetic development is continuous from initiation (or a very early stage of ontogeny) to maturation. 242 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 ORGANIZATION OF THE SHOOT SYSTEM Acer pensylvanicum. — The long and short shoots produced by striped maple during a single season are distinct from each other in both stem length and leaf production. Short shoots greatly outnumber long shoots after the first few years of growth, and suppressed understory trees may not produce any long shoots. The short shoots almost invariably bear a single pair of the large leaves which typify this species, and terminate in a bud or inflorescence. The stem of the short shoot rarely exceeds 5 cm. in length, and con- sists mostly of the internode below the single leaf-node. In a sample of 40 short shoots on herbarium specimens from widely scattered localities, the stems had a mean length of 2.0 cm. (range 0.3 to 4.6). An exceptional short shoot from Quebec had two similar pairs of leaves and a length of 2.0 cm., but all others had only one pair. Long shoots of striped maple have more leaves (two or more pairs) and much longer stems than short shoots, and the length of the stem is roughly proportional to the number of leaf pairs. In a sample of 37 long shoots from six trees used for growth observations, 27 shoots had 2 to 4 leaf pairs and the others had 5 to 9. The stems ranged in length from 10.7 to 102.6 cm. Shoots with only two pairs of leaves averaged 18 cm. in stem length (range 10.7 to 28.4). The shortest long-shoot stem ob- served (on a herbarium specimen from Pennsylvania) had two leaf- nodes and was 7.4 cm. long. No shoots between 5.3 and 7.4 cm. were encountered. Rae ange cogs pone are exceptionally long in this species. Many at least one internode between 20 and 25 cm.; the long- est observed was 28.5 cm. Acer rubrum.— The annual shoots of red maple have smaller, more numerous leaves and shorter internodes than striped maple shoots. They terminate in vegetative buds; inflorescences are borne on separate non- leafy shoots in this species. As in striped maple, the long shoots are out- numbered by short shoots early in the life of the tree, and in the crowns of older trees they make up onl 4 . t (Wilson, 1966). P y 0 5 percent of the annual shoots jet) Le) so) Fac) — a rs¥) md fa) a | ba 5a oe a7) I 2 tS < bab) pa p ea oO co a ° nA Oo Qu. is") < &, ° z 3 ag = ad ° = , but axillary buds sometimes produce long nodes. In a sample of 48 long shoots from 0 to 54 cm. and bore 12 to 16 pairs of leaves. 1971] CRITCHFIELD, ACER 243 THE WINTER BUD AND ITS CONTENTS Acer pensylvanicum. — The dormant vegetative buds of striped maple are remarkably uniform in construction and contents, and they offer no indications of the type of shoot they will produce the following spring. Nearly all have two pairs of bud scales, with only the outer pair ex- posed. As in other maples, the scales are modified leaf bases. The ter- minal buds are 6 to 10 mm. long; axillary buds are smaller (2 to 5 mm.), and some are rudimentary. In both this species and red maple, the en- velope formed by the bud scales is much larger than the embryonic shoot it covers, leaving considerable space for the shoot to enlarge be- fore the bud begins to expand. With rare exceptions, the buds contain a single pair of embryonic leaves and a pair of primordia. In terminal buds collected at the end of the growing season, the embryonic leaves were 13 to 40 times as long as the primordia (TABLE 1). They were far advanced in development, hay- ing a short petiole and a blade with three nearly equal lobes and many teeth (FicurE 1). The primordia were either undifferentiated in shape, as in Ficure 1, or showed the beginnings of two lateral lobes. If a bud subsequently produces a short shoot, the embryonic leaves expand into the single pair of leaves on the shoot, and the primordia develop into the outer scales of the new winter bud. If the bud produces a long shoot, the primordia develop into the second pair of foliage leaves. During bud expansion in the spring, the outer scales separate and the inner scales elongate. They form a loose envelope around the growing shoot, held tightly together along their margins by abundant tangled hairs. By the time they separate, both pairs of leaves have increased greatly in size (Taste 1) and are densely pubescent. A new pair of primordia is initiated at the apex of most embryonic shoots during bud enlargement (TABLE 1). The distinction between future long and short shoots was usually apparent during the later stages of bud enlargement. The primordia at the second node of future short shoots began to develop into bud scales. They grew at a relatively slow rate, and the leaf bases increased dispro- portionately in size. The first leaf pair was 16 to 38 times as long as the second in putative short-shoot buds. In this category were most axillary buds and the terminal buds of all short shoots and a few (3 of 20) long shoots. In buds that produce long shoots, the growth rate of the second pair of leaves was much faster, and by the time the buds opened pair 1 was only 3.3 to 16 times as long as pair 2. Most long- shoot terminal buds (17 of 20) and a few axillary buds were in this category. A single terminal bud sampled at this stage (omitted from TABLE 1) differed from all others in having two pairs of well-developed embryonic leaves. The second pair was 40 percent as long as the first and 19 times as long as the pair of primordia at the apex. On the same tree, a bud like 244 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 TABLE 1. — Bud contents and previous season’s leaf production of Acer pensylvanicum and A. rubrum A. PENSYLVANICUM A. RUBRUM Terminal buds Terminal buds Lon Short Axil. on Short Axil. shoots shoots buds shoots shoots buds Buns pormMant * No. of buds 15 5 9 14 1 No. leaf-nodes on previous season’s shoot: Mean 41 -~ 78 2.0 a Range 2-8 _ 4-11 1-3 =: No. If. pairs in bud: Mean 2 3 2.4 2 Range ~ ~ ~ 2-3 = Range in If. length (mm.) at node: 2.6~4. 1.3-3.1 80~1.2 60—.90 60 2 .05~.30 02-11 50-85 .25-.70 30 = - 10~.35 .04—.20 a Ey BupDs SWELLING ” No. of buds 20 3 41 33 20 14 No. leaf-nodes on previous season’s shoot: Mean 4.0 1 = 7.9 2.0 = Range 2-8 - = 4-16 1-3 ae No. lf. pairs in bud: Mean 28 2.6 aa 4.0 3.6 3.3 2-3 2-3 3-5 3~4 2-4 Range y he Range in If. length (mm.) at node: i 2 30-3.1 40-1.0 07-2.6 2.0-6.5 1.2-5.0 37-3.5 3 01-.15 01-.10 01-.15 45-5 15~2.2 05-.70 : 01-1.5 01~.55 01-.10 .01-.35 ~ = “A. pensylvanicum collected | ; ; ate Sept. and -Oct. 1- lected mid-Feb. from 1 “at pt. and mid-Oct. from 2 trees. A. rubrum co 3 A. pensylvanicum collected mi i in id- to late April fr . ds opening. A. rubrum collected same period from 6 fees a aula aati ; this one apparent] lea y produced a sing] . ; 1 ves during th ingle long shoot with two pairs of early © ensuing growing season (see next section). buds contained two leaf pai i ’ . pairs. In buds with three pairs, pairs 2 and 3 were 67 and 22 percent as long as was pair 1 (ranges 1971] CRITCHFIELD, ACER 245 60 to 78, 11 to 30). The embryonic leaves at nodes 1 and 2, despite their small size, had well developed, mostly glabrous blades with three nearly equal lobes and several marginal teeth, but only the largest had distinct petioles. The leaves at node 3 were unlobed if they were less than .15 to .20 mm. long, but above this size they had the beginnings of two lateral lobes. In red maple, bud enlargement and the growth of the embryonic shoot was a gradual process that began long before the buds opened at the beginning of May. By mid-April, long-shoot terminal buds from the same tree sampled in February had nearly doubled in length, and the first pair of leaves had more than tripled (from 1.0 to 3.5 mm.). Pairs 2 and 3 had grown even faster, and were 79 and 37 percent as long as pair 1 (ranges 73 to 82, 28 to 42). Most red maple buds initiated an additional pair of primordia during enlargement (TasLe 1). A few at 402 2 40.1 ss mm Jo mm \ IL oL ASN | 2 NODE Ficure 1. Contents of a winter bud of A. pensylvanicum. Collected in mid- Gctober, this bud terminated a long shoot with seven pairs of leaves. One of the ab onic leaves at node 1 is omitted. The primordia at node 2 flank the dome-shaped apical te 246 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 expanding terminal buds of long shoots (3 of 33) had five pairs of leaves (TABLE 1), but it is uncertain whether they had four pairs in the dormant bud or initiated two pairs during bud swelling. Other species. — Expanding buds were collected in late April, a few days before bud opening, from a single tree of Acer saccharum Marsh. (sugar maple). This species lacks the specialized long and short shoots of réd and striped maples. The buds were terminal or axillary on annual shoots 1.0 to 21.8 cm. long with 2 to 3 pairs of leaf scars. The nine buds in the sample had 8 to 10 pairs of scales and contained 3 to 6 pairs of leaves, the uppermost a pair of primordia. If the latter were initiated during bud expansion, as in red and striped maples, the dormant buds con- tained 2 to 5 pairs of leaves. In Acer platanoides L., according to Moore (1909), the entire leaf complement of the next season’s shoot is present in the winter bud. She found 3 to 4 pairs of leaves, the uppermost pair small and morphologically undifferentiated. Kiister (1898) also reported three leaf pairs in the buds of this maple. The winter buds of another European species, A. pseudoplatanus, have 6 to 8 pairs of scales and 2 to 4 pairs of leaves (Schiiepp, 1929). GROWTH OF THE SHOOT ; Red and striped maples are among the many temperate woody plants in which the principal shoots elongate until midsummer or later. Stem extension of red maple has been reported to continue until late June to period in the spring and early summer. Acer Saccharum completes 90 to 100 percent of its stem extension in 17 to 35 days (Cook, 1941; Kienholz, sie Jacobs, 1965). It is widely assumed that shoot growth in this group, unlike the first, is restricted to leaves and internodes laid down in ie ee he inner bud scales separated at the end neha ‘s : en the buds were 2 to 5 cm. long, and were shed within a few the aa le i a of the embryonic leaves of the winter bud into oa aves of the shoot was well under way by the time the bud hee tet ee of the early leaves continued through most of May. €s rellexed and unfolded soon after emergence. The first pair 1971] CRITCHFIELD, ACER 247 100 re r T T T T T - 90} Leaves: ® 4 Vv early v late BOL : 7O- a 6O0r a £ o } ae | : = ® fon =< sat o 2 wn 30} i 20L . JOL A © 10 20.3010 20 30 10 20 30 10 May June pare July August g and one sprout (4). Two additional pairs of leaves matured on shoot 1 before 248 JOURNAL OF THE ARNOLD ARBORETUM [ vou. 52 days after the first pair (FicurE 2: shoot 1). The petioles of the early leaves lagged behind the blades, reaching 90 percent of their final length an average of 11.1 days later (range 7 to 16). The first internode, already 2 to 6 mm. long in elongating buds, grew rapidly after the shoot emerged. On both long and short shoots, it reached 90 percent of its final length in the latter half of May, 7.5 to 13 days before the pair of early leaves at its upper end. This internode elongated faster and reached much greater final lengths on long shoots: 5.2 to 14.5 cm. in the 37-shoot sample. The expansion of the leaves at the first node, the extension of the stem below the node, and the development of a new winter bud completes the growth of non-flowering short shoots. The bud and the short inter- node separating it from the leaf-node developed slowly throughout the summer, The internode below the bud ultimately reached a length of 2.5 to 8 mm. The “stalked” bud, with its single pair of exposed scales, is a characteristic feature of the subdivision of Acer to which A. pens ylva- nicum belongs (section MAcRANTHA Pax). On long shoots, the first pair of late leaves began its most rapid growth in length in the third week of May, when the early leaves were about two-thirds of their final size. The phyllochron of the first late leaves was the longest on most shoots, averaging 18.3 days (range 16 to 21). Succeeding pairs of late leaves, all initiated during the growing season, developed at intervals of 14.1 days (range 4.5 to 20). The petioles matured after the blades, but the lag was shorter (mean 7.6 days) and more variable (range 0 to 21) than in early leaves. The long internodes of the late leaves matured well before the cor- responding leaves. They reached 90 percent of their final length 11.5 days earlier (range 6.5 to 17). At the completion of extension growth, internode 2 was usually the longest on the shoot (Ficure 12c). The only exceptions in the 37-shoot ct Om at ter ath ny Sek eae Groce 7: chact 1) only shoot with two pairs of early leaves oti gs oh of internodes decreased steadily from the longest in- € shoot tip on jong ey with six or fewer leaf-nodes, but ; more nodes had a second peak in internode length aay tae 3 to 5 internodes above the first. Most long shoots wocarpa (Critchfield, 1960) and Ginkgo biloba (Critch- maple. The first smal ] pai ; when the sprout was pair of leaves (Ficurr 6) matured in late May, only 1 cm, long; the second pair, much larger than 1971] CRITCHFIELD, ACER 249 the first, did not reach 90 percent of their final length until more than a month later (FicurE 2). By the time shoot growth stopped at the end of July, three more pairs of leaves had matured at intervals averaging 10 days (range 7.5 to 12.5). Leaf pairs 2 to 5 were like the late leaves of long shoots in all respects. In a small defoliation experiment, the removal of the late leaves at an early stage of development drastically reduced the elongation of long- shoot stems. Three shoots on the same sprout clump were treated. Be- ginning in mid-May, leaf pairs 2 to 5 were cut off when they reached lengths of 1.2 to 3.0 cm. (Ficurr 3), and the cut surfaces were covered with lanolin. Stem extension, which is largely concentrated in internode 2 during late May and early June, continued almost normally for about 3 weeks. On two of the three shoots (FicurE 3: B, C) the elongation of this internode was hardly affected by the treatment; its final length was 83 to 84 percent of the control. On the third treated shoot (A), the final length of internode 2 was only 38 percent of the control. In early June, stem extension of all three treated shoots declined abruptly and permanently (Ficure 3). Later-formed internodes were only 8 to 12 percent as long as corresponding internodes of the control shoot. Two shoots (B, C) began to die back from the tip in early July, after elongation had almost stopped. Shoot A subsequently produced a sixth pair of leaves, but by the end of the season all three treated shoots had died back to the first node. Acer rubrum. — The buds opened in early May when they were 8 to 11 mm. long. The 2 to 3 pairs of early leaves on each shoot expanded rapid- ly through most of May. Early in development, the unfolding blade re- flexed, forming an acute angle with the petiole. The expansion of the first two leaf pairs was almost simultaneous. The phyllochron of pair 2 averaged only 1.0 days, ranging from —2.5 to + 4 days. (Negative phyllochrons were due to the maturation of pair 2 out of sequence.) If a third pair of early leaves was produced, it ma- tured 2.5 to 5.5 days later, with a mean phyllochron of 3.4 days. All early leaves on a total of 17 shoots measured in three seasons reached 90 percent of their final length between May 25 and June 13. The petioles of the early leaves matured a mean of 10.1 days after the blades, and their variation in final length was reflected in the range of this interval: 5 to 20 days. The internodes of the early leaves matured a week or more before the corresponding leaves. This interval could not be accurately estimated on short shoots, which rarely produce internodes more than 1 cm. long, but the longer internodes of early leaves on long shoots reached 90 per- cent of their final length 8.5 days before the leaves (range 5 to 13). The growth of most short shoots of red maple was almost completed by early June (Ficure 4: shoots 3, 6; Ficure 9). Some short shoots later produced a pair of small transitional or late leaves, but their ex- pansion was not accompanied by appreciable stem elongation. The new JOURNAL OF THE ARNOLD ARBORETUM [von. 52 250 70 Control 60} 50} E © ~ '40- £ _ mn 5 30 E @® oe — ” 20h lO j a y Y y y : Be eo a ae See 16 3 ST AY JUNE JULY AUGU . DATE Ficure 3. The effect of late-leaf removal on stem extension in - eat vanicum, pair of developing late leaves was removed from eac shoot (A-C) on dates indicated by arrows. Control is shoot 3 of FIGURE 2. terminal buds of the shor size in August. The production of a t shoots developed slowly, reaching their final ter second set of leaves on long shoots began af 1971] CRITCHFIELD, ACER 251 a distinct hiatus in the continuity of leaf production. At this stage the early leaves and their internodes had nearly completed their growth, but the first pair of late leaves was only a fraction of its final size and its internode was still elongating (Ficure 8). On the long shoots measured in 1957-58, the phyllochron of total length for the first pair of late leaves averaged 12.9 days (range 7.5 to 19); in 1959 the phyllochron of blade length for this pair averaged 8.8 days (Ficurr 4). On all shoots, this interval was 5 to 16 days longer than the preceding phyllochron and 1 to 9 days longer than the next. Successive pairs of late leaves matured at intervals averaging 6.9 days (range 2 to 16). Their petioles tended to be relatively shorter than those of early leaves, and the interval between blade and petiole maturation was slightly less (mean 8.2 days, range 2 to 16). The interval between internode and leaf maturation was also somewhat shortened on the late-leaf part of the shoot, and tended to decrease to- ward the shoot tip. Internodes matured an average of 5.7 days before the corresponding late leaves (range 0 to 12). In their early stages the blades of late leaves are red-pigmented and strongly reflexed. This gives the crowns of actively growing red maples a characteristic appearance in early summer, when late-leaf production is at its height. Each of the principal branch systems, comprising a peren- nial long shoot covered with short shoots of various ages, terminates in a distinct reddish cone of developing late leaves. When stem extension is completed, the internode separating the early and late leaves is usually the longest on the shoot. Most shoots from terminal buds have three pairs of early leaves, and internode 4 is usually the longest. On long shoots maturing from axillary buds, which more commonly have only two early-leaf nodes, internode 3 is usually the longest. In the 48-shoot sample, internode 4 was longest on 23 of 32 shoots from terminal buds, and internode 3 was longest on 12 of 16 shoots from axillary buds. Most of the exceptions were shoots with two peaks of internode length, on which internode 3 or 4 was the first peak and the second peak was the longest internode on the shoot. As in striped maple, the incidence of two peaks in the final length of internodes along the shoot depends partly on number of internodes. In red maple it also varies from tree to tree, with two-peaked shoots uncom- mon or absent on some trees. In the 48-shoot sample, few (3 of 29) Shoots with 4 to 8 internodes had two peaks, but 9 of 17 shoots with more than 8 internodes had two peaks 2 to 5 internodes apart (mean 3.5). Other species. — The timing of leaf expansion was measured on five Shoots of a single sugar maple growing at the Harvard Forest. The Shoots developed from terminal buds of shoots that had produced 2 to 5 Pairs of leaves the previous year. The buds opened in late April, and 2 to 4 pairs of leaves reached 90 percent of their final length during the last half of May at mean phyllochrons of 1.0, 1.2, and 5.8 days. One 252 JOURNAL OF THE ARNOLD ARBORETUM [vou. 52 Leaves: vearl v late Year: 2/958 01959 i Length- cm ° ra) Stem 0 FIGURE 4, The timing of stem elonga Ps Y transitional 1 20 30. W...20 30 10 JUNE JULY AUGUST DATE i : r m tion and leaf maturation on A. rubru 1971] CRITCHFIELD, ACER 253 shoot produced a fifth and smaller pair of leaves which matured about 2 weeks after pair 4. The development of a single heterophyllous long shoot of A. pseudo- platanus was described in detail by Schiiepp (1929). The rapid ex- pansion of the first three pairs of leaves after the bud opened was fol- lowed by a long pause before the second set of leaves began to mature. Phyllochrons of blade length (estimated from Schiiepp’s Figure 4) were 1 and 3.5 days for pairs 2 and 3, but pair 4 (the first late leaves, in the terminology used here) did not mature until 20.5 days later. Blades of the next six pairs of leaves matured at intervals of 7 to 12.5 days (mean 8.7). The petioles of all leaves matured 14 to 27 days after the blades. Stem elongation was continuous, and by the time it ceased in early August the shoot had 14 pairs of leaves. Its development differed from that of a comparable red maple shoot primarily in the longer interval between blade and petiole maturation, and the much longer phyllochron separating the early and late leaves. VARIATION IN LEAF FORM The early leaves of red and striped maples, developing from embryonic leaves in the winter bud, predominate in the crowns of all but the young- est trees. From the data of Wilson (1966) on red maple, it is estimated that they make up about 83 percent of the photosynthetic area of the crown of 5-year-old trees, and this proportion increases to 97 percent in trees 60 years old or more. The early leaves are usually the only leaves associated with the reproductive structures of maples. These are the leaves that have been utilized in the classification of the maples, since most plant collections sample parts of the shoot system bearing flowers and fruits. In taxonomic descriptions the early leaves are sometimes re- ferred to as “typical,” “ordinary,” or “adult” leaves. Leaves that are not preformed in buds are produced throughout the life of red and striped maple trees, although their relative number dimin- ishes with age. In this category are such diverse types as the small leaves of first-year seedlings, the large leaves of vigorous sprouts, and the late leaves of long shoots. Also included are leaves of the lammas and disguised lammas shoots? regularly produced by young trees of Acer platanoides, A. pseudoplatanus, and other maples (Spath, 1912; Burger, may be most common in species that do not routinely produce long and short shoots; 1 shoots. Triangles and V’s indicate dates on which leaves (1958) or blades (1959) reached 90 percent of final length. Shoots 3 and 6 are short shoots, the others long shoots. Three additional pairs of leaves on shoots 1 and 2 were 254 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 1926). All of these kinds of leaves develop epigenetically, and this com- mon mode of ontogeny appears to be responsible for the tendency of leaves on these otherwise dissimilar kinds of shoots to resemble each other in shape. In species with lobed leaves, including red and striped maples and many others, the lateral lobes of epigenetic leaves tend to be displaced down- ward on the blade, to project outward from the midrib at a greater angle, and often to be reduced in size. These leaves, borne mostly on “sterile” shoots, are poorly represented in herbaria. In taxonomic accounts they are sometimes referred to as “atypical,” “extraordinary,” or “juvenile” leaves. Acer pensylvanicum. — The early leaves of striped maple are among the largest in the genus, They are bigger than any of the late leaves on the same shoot, but the two kinds of leaves overlap in size on different shoots. In a sample of both shoot types, the blade area of early leaves was 87-388 cm.*; that of late leaves 14-179 cm2 On the most vigor- ous long shoots (5 to 9 leaf-nodes) the biggest late leaves were 46 to 64 percent as large as the early leaves, but on long shoots with only two nodes the late leaves were much smaller (9 to 35 percent). The two kinds of leaves also differ in shape. The more uniform early leaves terminate in three nearly equal lobes (Ficurr 5). The principal lateral lobes of the late leaves are shorter than the middle lobe and dis- placed basally (Ficure 12a, e). The veins forming the axes of the lateral lobes were 79 to 98 percent as long as the midrib in the early leaves of the sampled shoots, 49 to 76 percent in the late leaves. The middle lobe of early leaves was 35 to 45 percent of the blade length; that of late leaves was 49 to 62 percent, The lateral lobes are Most conspicuously reduced at the first late-leaf pris ne leaves of sprouts are small, unlobed, and coarsely toothed, but asl e first node the leaves are indistinguishable from late leaves of is a. ce in Size and shape (Ficurr 6). This sequence in leaf shape fe ia ae! on elongate sylleptic shoots — branches that develop di- (Spith, 191 2). ary meristems during the same season as the parent axis Th : io ric and often the only, pair of leaves on first-year seedlings of ite . e = serrate but not lobed. Their similarity in shape to the much sen ‘ale * i trees was pointed out by Deane (1896). They are , wever 3 areas of only 2.6-6.7 Sie few measured seedling leaves having blade cer rubrum, — SSO Wie i Although red maple has an extensive infraspecific no- Primarily on jts variability in leaf shape, the differences 1971] CRITCHFIELD, ACER 255 GURE 5. A ape age long shoot of A. Ad igh Collected at the Harvard Forest in early June, this shoot ao one r of nearly mature early wages and two ‘pte ie pairs of late leav RE ure leaves of a sprout of 7 splehlging ert (1 leaf per node). Except for ‘leaf ir node 1 (lower left), they are indistinguishable fon! late leaves of long shoots. Growth of this sprout is shown in Fics RE 2 (shoot 256 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 between early and late leaves of the same shoot are not as pronounced as in striped maple, and transitions are common. The early leaves, like those of striped maple, generally have larger blades (Wilson, 1966). Differences in leaf shape also parallel those of striped maple (FicuRE 12a, e). Early-leaf blades have three subequal lobes (FicuREs 8—10), in contrast to the elongate median lobe of late leaves (FicuRE 10). In a sample of long shoots, the midlobe was 46 to 56 percent of the blade length in early leaves, 60 to 70 percent in late leaves. The main lateral veins of the first two pairs of early leaves were 82 to 91 percent as long as the blade, but the third pair, which often has narrower blades (FIGURES 8, 10), sometimes fell within the range of the late leaves (65 to 79 per- cent). The basally displaced and shortened lateral lobes of the late leaves tend to project outward from the midrib. The angle formed by their axes was 95-117°, compared to 73-86° in early leaves. This feature provided a consistent distinction between the two kinds of leaves in this species but not in striped maple. The relative length of petioles tends to decrease from node to node, stabilizing somewhat in the late-leaf series (Ficures 10, 12b). The petioles of early leaves were 60 to 114 percent as long as the blades; late-leaf petioles were 30 to 80 percent. As Wilson (1966) has pointed out, variations along the shoot in petiole length, together with changes in blade size and internode length, tend to minimize the mutual shading of leaves, The leaves of red maple sprouts and sylleptic shoots resemble the late leaves of long shoots in their elongate median lobes, shortened and pro- Jecting lateral lobes, and short petioles (Jackson, 1899: pp. 99, 100, Fig- ure 27). These tendencies reach their extreme expression in the leaves of first-year seedlings. The lateral lobes of seedling leaves are usually much smaller than the elongate median lobe (Ficurr 7; Jackson’s Fig- ped “ts ), and are sometimes reduced to teeth (Brayshaw, 1959: photo. on p. mented by Published accounts (Brayshaw, 1959; Elwes and Henry, 1908; : ade ei has not been comprehensively mono- i » and the partial classification of Rehder (1940), slightly modified by Brizicky (1963), is followed here. 1971] CRITCHFIELD, ACER 257 cist et, _. Se as e i SUT UE Ey” Hitnahin iain URE 7. First-year seedling of A. rubrum collected at Harvard Forest in ne Ficures 8, same a (late May) from same ee On the long shoot (FIGURE 8) three pairs of early leaves are nearly mature and two pairs of late leaves have un- folded. The growth of the short shout (Ficure 9), with two pairs of early leaves, is nearly complete. 258 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 (1) Species in this category, which possibly includes the majority in the genus, are like red or striped maple in the organization of the shoot system and the regular production of two kinds of leaves. Included here are the other species in sect. MACRANTHA, all closely similar to striped maple. More widespread in the genus is the red maple pattern: less sharply distinct long and short shoots, more than one early-leaf node on most shoots, and less abrupt changes in leaf shape on long shoots. This pattern is represented in sections CAMPESTRIA Pax, ARGUTA Rehd., and S Japanese maple in sect. RuBra which closely resembles red maple in most of its characteristics. Members of group (1) are considered in more detail below. (2) This group has the same kind of shoot organization as (1), but epigenetic leaves of long and other types of shoots do not differ as con- sistently in size and shape from leaves preformed in buds. Included here are maples in sections Rusra (A. saccharinum L.) and Acer (A. spicatum Lam., A. caudatum Wall.). (3) In this group the shoot system is not organized into long and short shoots, and the elongate shoots of the crown are known or inferred to be mostly preformed in the bud. But leaves that develop epigenetically, when they are produced at all, differ in form and size from preformed leaves. All of the species in section PLATANOIDFA Pax are included here. Representative of this group are Acer platanoides and A. miyabet Maxim. Many elongate shoots of these maples bear only 3 to 4 pairs of leaves, but the most vigorous produce 1 to 3 additional pairs of smaller, differently shaped leaves crowded at the shoot tip. On such shoots of A. miyabei there was a distinct break in size and form between nodes 4 and 5 (Ficure 12d, e). The upper leaves were smaller, with an elongate prin lobe and reduced lateral lobes projecting at right angles from the Acer platanoides shows the same tendencies in leaves at the tips of ment. (4) The shoot system of thi the differences in size and sha iy ao o Qu °o Lae? lon | sa] p re S g a c — & Q. jae — 1971] CRITCHFIELD, ACER 259 Of the many maples in category (1), the largest aggregation of related species is sect. Macrantua, which includes 15 to 20 Asian species in addition to Acer pensylvanicum. Their classification is based almost en- tirely on differences in size and shape of early leaves; the reproductive structures are rather uniform and do not provide many useful distinctions (Rehder, 1933). The Asian species are nearly identical to striped maple in the organiza- tion of their shoot systems. Throughout the section, the short shoots have a single pair of early leaves which vary in shape from 3- or 5-lobed to unlobed in different species. The long shoots bear, in addition, one or more pairs of smaller late leaves, and internode 2 is the longest on most shoots (FicuRE 12c). In species with lobed leaves, the differences in shape between the two kinds of leaves are much like those in striped maple: the late leaves have relatively elongate median lobes and shortened, basally displaced lateral lobes (FicuRE 11b: A. rufinerve Sieb. & Zucc.; Ficure 12a, e: A. tegmentosum Maxim., A. capillipes Maxim.). In some species the angle formed by the lateral veins is consistently larger in late leaves (e.g., A. grosseri Pax: 65~82° in early leaves vs. 92~-106° in late leaves). The trend toward reduced lobing is sometimes reversed in species with unlobed or obscurely lobed early leaves, and the late leaves may have small lateral lobes at the base of the blade. Throughout this section, the marginal serrations of late leaves tend to be fewer, coarser, and blunter. Apart from leaf shape, the other maples in section Macrantna differ from striped maple primarily in their smaller vegetative structures. Only A, tegmentosum approaches it in leaf, internode, and bud size. In the other species, both kinds of leaves are smaller, the vegetative short shoots average 1 to 1.5 cm. and rarely exceed 3 cm. in length, and the longest internodes of most long shoots are only 8 to 12 cm. Also included in group (1) are most or all of the maples in sect. CAM- PESTRIA Pax. (Possible exceptions are A. opalus Mill. and A. kyrcanum Fisch. & Mey., for which adequate material was not available). The or- ganization of the shoot system is much like that of red maple, with 1 to 3 Pairs of early leaves on all shoots originating from buds. Long shoots are regularly produced, and heterophylly is pronounced (Ficure 1Ic, e, f), but the differences between early and late leaves vary greatly between Species. Acer campestre L., like many other maples, has late leaves with short petioles (FicurE 12b) and basally displaced lateral lobes (12e) which project from the midrib at large angles (12d). Unlike the late leaves of most other maples, however, they often have more lobes and blunt teeth than the early leaves. This tendency is more conspicuous in A, mons pessulanum L., in which the more numerous lobes and teeth of the late leaves are the most reliable distinction between the two kinds of leaves (Ficurr 11e). The opposite is true of A. orientale, a Mediterranean Species with small leathery leaves. The lateral lobes of its late leaves are much reduced in size (Ficure 11f), and the first pair on long shoots may lack lobes or teeth, aeT : “4 260 JOURNAL OF THE ARNOLD ARBORETUM VOL: 152 Z20( Ni 10 [ FIGURE 10. Mature leaves of a long shoot of A. rubrum (one leaf per node). Early leaves in upper row, node 1 at left. Figure 11. Early and late leaves of Acer species: . 5), b. A. rufinerve (1, 3). c. 4. campestre var. leiocarpum (2, 5), d. « * oa gerianum (2,7), e. A. monspessulanum (1, 8), f. A. orientale (1, 4). Leaves a, c, and d from shoots illustrated in FIGURE 12. ‘ oe a. A. tataricum (nodes 2, : buer- The species in section Arcuta also conform closely to the red maple pattern of shoot organization. They have a maximum of two or three -— leaves per shoot. depending on the species. The expression of heterophy ; ly in A. barbinerve Maxim. is typical of this section and similar to ea! other maples. Its late leaves have shorter petioles (Ficure 12b) ant smaller, basally displaced, projecting lateral lobes (Ficure 12a, d, e). ; Acer section ACER (Spicata), the most heterogeneous of Rehder's sec 1971] CRITCHFIELD, ACER 26 1 ‘ i ee ees eee een oe T T r T T 100 @ MEAN LENGTH OF PRINCIPAL LATERAL VEINS “| @RELATIVE PETIOLE LENGTH : (% of midrib) 5 (% of blade length) 7 150 4 80. a ; aS 12s- y % R % PS | tool 4 6oL i 7 735/ 4 BA .c 7 1 3 5 CL So, 4oL R SS a Ss ee eee L 4 i i n i 1 Paras i 257 I 3 5 ta 9 4 shed 4 NODE ; ee NODE I 3 SE 7 3 ee © INTERNODE 4 ae eae Hae LENGTH 150 @ ANGLE BETWEEN LATERAL VEINS 4 20, T 125 15, cm degrees 3 ay = M P = a 75 c 50 L a i ' wove> 5 ! 3 wove © . : 3 nove © t SL SSL Soa iia Rare | Fy ae eae ee Ra See at 90 M 4 @LENGTH OF MIDDLE LOBE TA gol. (% of blade length) BA ; cL eu TOL af - c v 60] R 1 * P sol SS 7 7 4o0L 7 30 ee es SE Sa 3 5 ' Nope 3 5 1 3 wove © T 1 NODE FIGURE 12. Variation in leaves and internodes on long shoots of Acer species: A, barbinerve (BA), A. buergerianum (BU), A. campestre var. leiocarpum (CL), j P pseudoplatanus ’ . black symbols. Uppermost leaf-nodes (not shown) were immature in A. capil- hpes, damaged by insects in A, rubrum. tions, includes several taxa that conform in varying degrees to the red maple pattern: A. pseudoplatanus, A. buergerianum Miq., and the shrubby Species-complex made up of A. tataricum L. and A. ginnala Maxim. o these, A. pseudoplatanus appears to be least regular in the production of long shoots: extension growth of young trees is commonly prolonged by lammas and disguised lammas shoots (Spath, 1912; Burger, 1926). Long 262 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 shoots are most abundantly produced by the A. tataricum-A. ginnala complex, and this abundance is a factor in their shrubby habit. The expression of heterophylly in these members of section SPIcATA follows the familiar maple pattern. Pearsall and Hanby (1926) observed that A. pseudoplatanus leaves produced during the summer were smaller and more dissected, with reduced basal lobes. They associated these morphological differences with increased water stresses. In the same species, the reduced lateral lobes of a long-shoot late leaf and a leaf of a sylleptic shoot were illustrated by Schiiepp (1929: Figure 16). His data on a single long shoot shows a pronounced break between nodes 3 and 4 in the relative length of lateral lobes and petioles (FicurE 12a, b). Schiiepp’s comparative observations of preformed and epigenetic leaf ontogeny make it clear that the lateral lobes of the preformed leaves ap- pear earlier and are more nearly equal in size to the median lobe at sub- sequent stages of development. First-year seedling leaves of A. pseudo- platanus are either unlobed or have reduced lateral lobes (Elwes & Henry, 1908: p. 642). Heterophylly is more pronounced in A. buergerianum and A. tataricum (Ficure lla, d). Both have late leaves with elongate median lobes and are smaller than succeeding leaves, often unlobed, and occasionally al- most devoid of serrations (Jackson, 1899: Figure 33 Je DISCUSSION ganization of the shoot system and in the ways in which extension growth is prolonged after the expansion of the leaves and internodes laid down in the winter bud. In type of shoot development least temporarily, after the winter bud. Prolonged continuous, through th red and stri : maples with specialized shoots, ile maples, and by inference other of Populus trichocarpa (Critchfiel this type. Stem elongation 1971] CRITCHFIELD, ACER 263 the second set of leaves on long shoots sometimes consists solely of a single pair of late leaves originating from such primordia. More commonly, most of the leaves produced during the second phase of long-shoot de- velopment are initiated and expand during the same season. In some woody plants with long and short shoots, this causal rela- tionship between the production of a second set of leaves by some shoots and the elongation of those shoots appears to be mediated by auxin or auxin precursors produced by the developing leaves. The evidence for this interpretation, including the relative auxin production of long and short shoots in Ginkgo biloba (Gunckel & Thimann, 1949) and Cer- cidiphyllum japonicum (Titman & Wetmore, 1955), has been reviewed elsewhere (Critchfield, 1960). The work of Zimmermann (1936) on the production of auxin by Acer pseudoplatanus shoots supports the extension of this interpretation to maples with specialized shoot systems. Zimmer- mann did not discriminate between long and short shoots, but his data and photographs show that long shoots with about eight leaf-nodes, sampled in late June, had much higher levels of auxin than any shoots sampled earlier in the season. Zimmermann also found a fairly close re- lationship between internode length and auxin content in developing long shoots of this maple. The stems of maple short shoots usually exhibit a limited but measurable amount of elongation, unlike those of some woody plants with specialized shoots. Ginkgo biloba, for example, produces short shoots only 1 to 2 mm. long. The leaf complement of red maple short shoots ranges from one to three pairs, and the presence of each additional pair is associated with small but consistent increases in stem length. These increments of stem growth are of a different order of magnitude, however, than those associated with late-leaf production. This is best illustrated by striped maple, in which the distinction between short and long shoots is reduced to the simplest possible terms. If a shoot produces only a single leaf-node it elongates an average of only 2 cm., despite the large size of the leaves. But the expansion of a single pair of late leaves, although they are only 1/10 to 1/3 the size of the early leaves, produces a stem averaging nine times as long. This great discrepancy between leaf size and stem length — 30- to 90-fold — illustrates one of the principal questions concerning plants with specialized shoot systems: why the elongation of short-shoot internodes is either limited (striped maple) or nonexistent (Ginkgo). Red and striped maples, despite their similarity in shoot organization, differ greatly in potential leaf production per shoot. The greater potential of red maple is due partly to the nearly synchronous expansion of 2 to 3 pairs of preformed leaves, but a more important factor is its short late- leaf phyllochron — less than half of the 2-week interval of striped maple growing in the same environment. At an average rate of leaf development, a long shoot of red maple in the Harvard Forest could produce 17 leaf pairs by mid-September. An average shoot of striped maple could pro- duce only eight pairs in the same period. However, because striped maple has larger leaves and longer internodes, it is estimated that its 8- 264 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 52 node shoot would exceed the 17-node shoot of red maple by at least 50 percent in both stem length and photosynthetic area. In addition to the foliage leaves, maple shoots periodically produce another type of appendage, the bud scales. From comparisons of foliage- leaf and bud-scale ontogeny in Acer platanoides and other woody plants Goebel (1905) concluded “. . . the path of development is originally the same for all leaves, but in many leaves at an earlier or later period the development may proceed along different paths.” His interpretation is equally applicable to the early divergence in development of preformed and epigenetic leaves of Acer, as Schiiepp (1929) showed in A. pseudo- platanus. Of the ontogenetic alternatives represented by early and late leaves, the development of leaves at the tip of a growing shoot is much more general. Leaves preformed in winter buds are produced only during the sharply discontinuous, unlike that of epigenetic leaves. The first phase Is a prolonged, well-defined period of what Sachs (1893) called ‘“mor- phologische Ausgestaltung” (putting-into-shape), during which the form of the leaf is determined. The shape of the first leaves in the winter buds of Acer palmatum Thunb. is fully blocked out by the time they are 5 mm. long, and undergoes little change during later development (Meijknecht, 1955). The embryonic phase of preformed leaves, as Sachs pointed out, is separated from their expansion phase by a sharp boundary imposed by winter dormancy, The degree of heterophylly associated with these two developmental patterns varies greatly between species of Acer, although the same mor- reed shoot. In most maples, however, the two kinds of leaves are distinct ki a _ shap ¢, and maples with specialized shoot systems have both inds present in the crowns of most trees. possible to identify the occa , 1971] CRITCHFIELD, ACER 265 the value of taxonomic information in such related fields as ecology and paleobotany, which necessarily deal with whole-plant variability, LITERATURE CITED Bonn, T. E. T. Studies in the vegetative growth and anatomy of the tea plant (Camellia thea Link.) with special reference to the phloem II. Further analysis of flushing behaviour. Ann. Bot. II, 9: 183-216. 1945. BraysHaw, T. C. Tree seedlings of eastern Canada. Canada Forestry Branch Bull. 122. 38 pp. 1959. BRIZICKY, G. K. The genera of Sapindales in the southeastern United States. Jour. Arnold Arb. 44: 462-S01. 1963. Burcer, H. Untersuchungen iiber das Héhenwachstum verschiedener Holzar- ten. Mitt. Schweiz. Centralanst. Forstl. Versuchswesen. 14; 29-158. 1926. CLausen, J. J . & T. T. Koztowsxt. Heterophyllous shoots in Betula papyrifera. Nature 205: 1030-31. 1965. °K, D. B. The period of growth in some northeastern trees. Jour. Forest. 39: 956-959, 1941, CRITCHFIELD, W. B. Leaf dimorphism in Populus trichocarpa. Am. Jour. Bot. 7: 699-711, 1960, ~———-. Shoot growth and heterophylly in Ginkgo biloba. Bot. Gaz. 131: 150- 162. 1970a, : "————. Shoot growth and leaf dimorphism in Boston ivy (Parthenocissus tri- cuspidata). Am. Jour. Bot. 57: 535-542. 1970b. FANE, W. Notes from my herbarium V. My seedling collection. Bot. Gaz. 21: 210-214. 1896. Etwes, H. J., & A. Henry. The trees of Great Britain and Ireland, vol. 3. Privately printed, Edinburgh. 1908. GorseL, K. Organography of plants, Parts I & IJ. English ed. Clarendon Press, Oxford. 270 & 707 pp. 1900 & 1905. GUNCKEL, J. E., & K. V. THIMANN. Studies of development in long shoots and short shoots of Ginkgo biloba L. III. Auxin production in shoot growth. - Jour. Bot. 36: 145-151, 1949, : Jackson, R. T. Localized stages in development in plants and animals. Mem. Boston Soc. Nat. Hist. 5: 89-153. 1899. ; Jacoss, R. D, Seasonal height growth patterns of sugar maple, yellow birch, and red maple seedlings in Upper Michigan. U.S. Dep. Agr. Forest Serv. Lake States Forest Exp. Sta. Res. Note LS-57. 4 pp. 1965. : Krennorz, - Seasonal course of height growth in some hardwoods in Con- necticut. Ecology 22: 249-258. 1941. 266 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Konprat’EVA-MELVILLE, H. A. On the heterophylly in the seedlings of some woody plants. ag Russian; English summary) Bot. Zhur. 50: 605-613. 1965, Kuster, E. Uber das Wachstum der Knospen wahrend des Winters. Beitr, Wiss, Bot. 2: 401-413. 1898. MEIJKNECHT, J. G. On the ideal value of varying characters. Acta Bot. Neerl. 4: 273-320. 1955. Moore, E, The study of winter buds with reference to their growth and leaf content. Bull. Torrey Bot. Club 36: 117-145. 1909. PearsALL, W. H., & A. M. Hansy. Growth studies V. Factors noi i: the de- velopment and form of leaves. Ann. Bot. 40: 85-103. 1926 Renper, A. New species, varieties and combinations from the herbarium and the bibeceies of the Arnold Arboretum. Jour. Arnold Arb. 14: 200-222. 1933 ————. Manual of cultivated trees and shrubs. ed. 2. 996 pp. New York. 1940. Sacus, J. Mee ilatache Notizen iin Ueber Wachstumsperioden und Bil- dungsreize. Flora 77: 217-253. ScHUepp, O, Untersuchungen zur Be bente und experimentellen Sapir lungsgeschichte von Acer Pseudoplatanus L. Jahrb. Wiss. Bot. 70: 804. 192 nae ae R. a The leaf dimorphism of Liguidambar Sstyraciflua. Am. Midl. Nat. 42-50. 1967, SpAtH, H. “i ie Johannistrieb. 91 pp. Paul Parey Verlags. Berlin. 1912 TITMAN, P. W., & R. H. Wetmore. The growth of : and short shoots in Cercidiphylium. Am. Jour. Bot. 42: 364-372. ILS Bore Pap, 14. 21 pp. Harvard Forest, Petersham, Massachusetts: 1966, ZIMMERMANN, W. A. Unt tersuchungen iiber die riumliche und zeitliche Ver- teilung des Wuchsstoffes bei Baumen. Zeitschr. Bot. 30: 209-252. 1936. Pactric SOUTHWEST ForEST AND RANGE steele tapas Forest SERVICE, U.S. DepartM MENT OF AGRICULTU P.O. Box 245, BERKELEY, CALIFORNIA 94701 1971] MILLER, POLYGALACEAE 267 THE POLYGALACEAE IN THE SOUTHEASTERN UNITED STATES Norton G. MILiLer POLYGALACEAE R. Brown in Flinders, pce Terra Austr. 2: 542. 1814, “Polygaleae,”’ nom. cons. (MiLKwort FamIty) Annual or perennial herbs [spiny shrubs, sometimes small trees or climbers] with taproots, rarely rhizomes. Leave es alternate or whorled, simple, scalelike in some, entire, exstipulate [petioles biglandular in a Inflorescences axillary, termin al [or extra-axillary| racemes and o = | mous types infrequently present. Sepals [4] 5, persistent in fruit or caducous, free, or the abaxial pair fused [or the calyx + synsepalous], the inner (lateral) segments (wings) petaloid and larger than [or + equal] the others. Petals 3 [5, or if 3, 2 additional rudimentary petals often present|, in most incompletely fused to form a tube in which the adaxial margins of the upper two segments overlap, the abaxial petal (keel) fimbriate-crested or not, sometimes lobed, the corolla caducous. Stamens 8, rarely [4 or 5] 6 [7 or 8-10], filaments mostly united, the resulting sheath connate with the corolla below, but split above; anther sacs 2, dehiscence usually + porocidal [sometimes by longitudinal slits |, introrse, the opening generally subterminal, oblique; pollen polycolporate, *Prepared for a generic flora of the southeastern United States, a project of the Arnold Arboretum and the Gray Herbarium of Harvard University made pos- : d initiated in the first contribution of the series (Jour. Arnold Arb. 39: 296-346. 1958). The area dealt with includes North and South Carolina, Georgia, Florida, Tennessee, Alabama, Mississippi, Arkansas, and Louisiana. Descriptions are based on species oc- Racy in this region, with additional gr beanrie from extra- regional taxa in are those which I have not age able to check for accuracy of citation. My thanks are extended t for his continuing assistance and advice, to Mrs. Nancy who helped with bibliographic problems and prepared the final typescript, ne Dr. Lily Perry for providing some much appreciated help in the library, and Dr. R ong for sending a c of the treatment of the late H. F. L. Rock, S. J. Smith, and C. E. Wood, ies 268 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 often synorate, exine stratification obscure, polar axis 25-62 p». Ovary superior, sessile [rarely stipitate], [1]2[3, 4, 5, or 7 or 8]-locular, hypog- ynous nectary present [or absent, sometimes annular]; style one, central, apical, usually curved, the apex 2-lobed [simple, or otherwise], one lobe stigmatic, the other often ending in a tuft of hairs; ovules one [rarely more| in each locule, pendulous from the ovary apex near the septum, anatropous to hemianatropous, crassinucellar, bitegmic, both integu- ments forming the micropyle. Fruit a loculicidal capsule [or drupe, nut, or samara]. Seeds often hairy [sometimes comose], generally with a fleshy micropylar aril or a harder outgrowth; endosperm usually present, fleshy; embryo straight, cotyledons orbicular, linear {or foliaceous]. Em- bryo sac development usually of the Polygonum type; embryogeny of the Asterad type. TyrE GENUs: Polygala L. Separate but allied family. Perhaps the most advanced tribe of the family, the Moutabeae Chodat (calyx and corolla united, filaments fused into a split sheath, ovary 2-8-locular, placentation axile), contains only Moutabea Aubl., a small genus of tropical South American shrubs. In addition to Polygala, two other genera of tribe Polygaleae (calyx and co- rolla separate, filaments fused into a split sheath, ovary 2-locular, pla- centation axile) are found in North America: Securidaca L., fruits inde- hiscent, dry, ae chromosome numbers for nm = 10, 14, 16, 18, 20, 22 32 Families consid sient ; meriaceae, these families comprise the St, which are regarded as an offshoot of the Sapin- 1971] MILLER, POLYGALACEAE 269 dales. Other authors suggest relationships to certain families or groups of families for the most part originally included in the Geraniales of Eng- ler & Prantl. Apart from certain species of Comesperma Labill., Securidaca, and Polygala that are sometimes cultivated as ornamentals, few other mem- bers of the family are of economic importance. Oil from seeds of Xantho- phyllum lanceatum (Mig. ) J. J. Sm., the siur tree of Sumatra, is used in native cooking and in Europe in the manufacture of candles and soap (Hutchinson). REFERENCES: BarLton, H. Polygalacées. Hist. Pl. 5: 71-92. 1874. English transl. M. M. HArroc. The Natural History of Plants 5: 71-92. 1878. [Including Kra- meria (Krameriaceae). BenTHaM, G., & J. D. Hooker. Polygaleae. Gen. Pl. 1: Rc 1862. [15 genera including Krameria and Trigoniastrum (Trigoniaceae). | BLAKE, S. F. Polygalaceae. N. Am. Fl. 25: 305-379. 1924. [ Monnina (11 spp.), Polygala (179 spp.), Securidaca (as Elsota, 8 spp.).] BOLKHOvsKIKH, Z., V. Grir, T. Matvejeva, & O. ZAKHARYEVA. Chromosome numbers of flowering plants. (Russian and English prefaces.) 926 pp. mie, 1969. [Polygalaceae, 576, 577.] CANDOLLE, A. P. pe. Polygaleae. Prodr. 1: 321-342. 1824. a. R. Sur la distribution et l’origine de l’espéce et des groupes chez les Polygalacées. Bibliot. Univ. Arch. Sci. Phys. Nat. III. 25: 695-714. 1891a. Monographia Polygalacearum. I. Mém. Soc. Phys. Hist. an Genéve Suppl. 1890(7). 143 pp. pls. 1-12. 1891b; II. Ibid. 31(2). xii + 500 pp. bls. 13-35. 1893. [405 spp., over half New World, the fea pe ie African, Asiatic, or Australian. ] — rt, Polysalaceae. Nat. Pflanzenfam. III. 4: 323-345. 1896 CLute, W. N. apes § of plant names LXXIII. Rues and milkworts. Am. Bot. 44: 10-17. 1938 Cronquisr, A. The evolution and classification of flowering plants. xii + 396 pp. Boston. 1968. [Polygalales, 274-277.] Davis, G. L. Systematic embryology of the angiosperms. x + 528 pp. New York. London, Sydney. 1966. [Polygalaceae, 216, 217 Dupe, V. P. Mo rphological and anatomical studies in Polygalaceae and its allied families. Agra Univ. Jour. Res. Sci. 11: 109-112. 1962.* Duke, J. A. On tropical tree seedlings I. Seeds, seedlings, systems, and sys- tematics. Ann. Missouri Bot. Gard. 56: 125-161. 1969. [Polygalaceae, 152. EIcHLer, A. W. Polygalaceae. Bliithendiagramme 2: 356-359. 1878 ERDTMAN, G. The systematic position of the genus Diclidanthera Mart. Bot. Not. 1944: 80-84. 1944. [ Belongs to Paleiice on the basis of pollen, stomatal apparatus, stone cells in pith, and vesture. ] ——. Pollen morphology and plant taxonomy. Angiosperms. Corrected re- print and new addendum. xiv + 553 pp. New York & London. vo [Polygalaceae, 332, 333.] 8 Farcrr, K., & L. vAN DER Pryt. The principles of pollination ecology. x + 24 pp. Toronto & Oxford. 1966. [Polygala comosa, 47; P. Chamaebuxus, 157-159,] 270 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Gite, E., & P. N. ScHtrworr. Die systematische Bedeutung des Vorkommens von Saponinen fiir Polygaceen-Gattungen. Arch. Pharm. 270: 276-283. Seber Polygalaceae. Gen. Pl. U. S. 2: 221-224. pls. 183, 184. 1849. [ Polyg- ala cruciata, P. paucifolia, P. Senega. Hecr, G. Polygalaceae. Jn: G. Hecr, Illus. Fl. Mittel-Europa. 5: 86-113. pl. 176. 1924. [Polygala (10 spp.), 89-113.] HEGNAUER, R. Chemotaxonomie der Pflanzen. Band 5. Dicotyledoneae: Mag- grouping. Lilloa 8: 83-198. 17 pls. 1942. [ Polygalaceae, 130-132. ] Hurcuinson, J. The genera of flowering plants. Vol. 2. xi + 659 pp. Lon- don. 1967. [Polygalaceae, 336-344. ] ; Jaucu, B. Quelques points de l’anatomie et de la biologie des Polygalacées. Bull. Soc. Bot. Genéve II. 10: 47-84. 1918. Knuru, P. Polygalaceae. Handb. Bliithenbiol. 2(1): 150-153. 1898. [Polyg- ala only.]; Ibid. 3(1): 449-452. 1904. [Monnina, Muraltia, Nylandtia, Polygala (4 spp., including P. polygama & P. paucifolia), Salomonia, Se- curidaca. | - Handbook of flower pollination. (Transl. J. R. A. Davis.) Vol. 2. viii + 703 pp. Oxford. 1908. [Polygala, 146-149; transl. from Vol. 2 of German edition only, see also Knuti (1904). LEvyNs, M. R. The floral morphology of some South African members of Polygalaceae. Jour. S. Afr. Bot. 15: 79-92. 1949. [Nylandtia (as Mundia, sp.), Muraltia (5 spp.), Polygala (2 spp.). | - The genus Muraltia, Jour. S. Afr. Bot. Suppl. 2. 247 pp. 1954. [115. ra. | Lewis, W. H., & O. HeRRERA-MacBrype. Polygalaceae. In: R. E. WoovsoN, R. W. Scurry, et al., Flora of Panama. Ann. Missouri Bot. Gard. 56: 9- 28. 1969. [Monnina, Moutabea, Polygala (13 spp. incl. P. leptocaulis), Securidaca,] Luspock, J. A contribution to our knowledge of seedlings. Vol. 1. viii + 608 Pp. London & New York. 1892. [Polygaleae, 205-207. Martin, A.C. The comparative internal morphology of seeds. Am. Midl. Nat. 36: 513-660. 1946. [ Polygalaceae, 614, 615.] METcaLre, C, R., & L. Cuarx, Polygalaceae. Anat. Dicot. 1: 133-138. 1950. Pryt, L. VAN DER. Principles of dispe lin, Heidelberg, New York. 1969, Rao, A. N. An embryological sty Phytol. 63: 281-288. pl. 17. 196 pollen nin anthers which be . . . i dy of Salomonia cantoniensis Lour. New come firmly attached to the viscid stigma ri vile ild flowers of the United States. Vol. 2. The Southeastern ppt sitt 2. New York. 1966, [Polygala, 528-530, pls. 195 p.p., 196, Ritey, H. N . The dispersal of plants throu . 22 ghout the world. xx + 744 pp Ron hE 50 Taba a ) » ». 4. Polygalaceae, In: A. . Am. 1(1): ; 448-460, 1897 m: A. Gray, et all, Synop. Fl. N. Am CHOLZ, H. Polygalaceae. In: } M , Pflanzen- familie. od 47. © sy a Engler’s Syllabus der 1971] MILLER, POLYGALACEAE 271 SMALL, ~ K. Polygalales. N. Am. Fl. 25: 299. anual of the southeastern flora. xxii : air pp. New York. 1933. oie spa 765~774. | paren fs Flowering plants, origin and dispersal. (Transl. C. JEFFREY.) x + 310 pp. plese eae D.C. 1969. [Polygalales, 226, 227 THORNE, nS F. Synopsis of a putatively phylogenetic classification ‘a the flow- ering plants. Aliso 6: 57-66. 1968. | Malpighiaceae, Polygalaceae, Kra- meriaceae, hs ape and Vochysiaceae comprising suborder Polyg- TurRNER, B. L. cee numbers in the genus Krameria: Evidence for familial status. Rhodora 60: 101-106. 1958. [22 = 12 in 3 spp. 7 some evidence seems to negate any relationship with the subfam. pono ria (Leguminosae) in which Krameria has been given tribal status by som Upnor, J. C. T. Leen: flowers. Bot. Rev. 4: 21-49. 1938. [Polygala- a 8 > d = & Q bo - cS) 38, 39.] VALENTINE, D. H., & D. A. Wess, eds. Polygalaceae. Jn: T. G. Tutty, V. H. HeEywoop, ef al., eds. Fl. Europaea 2: 231-236. 1968. Dedetinbee of Polygala by J. MCNEILL; 33 spp. in 3 subg.] Polygala Linnaeus, Sp. Pl. 2: 701. 1753; Gen. Pl. ed. 5. 315. 1754. Single- to many-stemmed, erect, ascending, or rarely creeping herbs [or unarmed to spiny shrubs| of open, dry or wet sands, moist peats, and, less frequently, damp, wooded sites. Roots usually aromatic, sparse in sometimes organized into a basal rosette, petioles short, blades generally ovate or lanceolate, or reduced to scales in some, entire, pubescent or glabrous, and often with glandular dots. Inflorescences loose to tight, terminal or axillary [rarely extra-axillary] racemes or corymbose panicles, flowers often aggregated into spikes or heads, rarely 1-4 on short, sub- terminal, axillary branches. Chasmogamous flowers zygomorphic, pedi- cellate, subtended by a persisting or caducous bract and two bracteoles; flower color white, pink, yellow, or orange, yellow flowers of some becom- ing blue-green upon drying. Sepals 5, quincuncial in bud, the wings large and petaloid, abaxial sepals 2, small, greenish, free or fused laterally for most of their length, adaxial sepal greenish, generally larger than either of the two abaxial segments, hooded in some; margins of abaxial and adaxial sepals often ciliate, sometimes bearing glandular hairs. Petals 3, the upper two fused to the abaxial petal (keel) along one margin, but free and overlapping above, or the lower margins of the upper two petals fused to the stamen sheath, which is connate basally with the keel; corolla often changing color with age. Stamens 8, rarely 6 in two groups of 3 each; anther sacs laterally confluent at anthesis, dehiscence Subapical, + porocidal, or by a V-shaped slit on the inner face of the anther wall. Ovary bicarpellate, compressed contrary to the septum, an 272 JOURNAL OF THE ARNOLD ARBORETUM [VoL, 52 SEF News Bie Ss, a‘ = wy om NS Guyym NSS P LBM yh iN DGIEINS ygala subg. Chamaebuxus la: a, plant with two chasmogamous flo ; cha . f of corolla removed to show androecium a ; > G, same, style removed, 3; e, anther shedding pollen, Fo f, empty anther, 20; g, ovary in longitudinal section, one ovule per heeee style mostly removed (semidiagrammatic), %_8; h, ovule and immature ‘ae lobed aril, funiculus slightly exaggerated, 15; i, mature seed with oben aril ; J, mature seed in longitudinal section, endosperm stippled, 1971] MILLER, POLYGALACEAE 273 abaxial or adaxial nectariferous gland often present; style usually 2- lobed, stigmatic lobe generally subapical, thin membranes borne on both sides of the style between or at the junction of the two lobes. Cleistoga- mous flowers of + similar morphology borne on basal, rarely aérial, branches in a few. Fruit a 2-loculate, glabrous [or pilose], thin-walled, loculicidal [sometimes winged] capsule, often accompanied by the per- sistent calyx. Seeds ovoid, globose, or ampulliform, pubescent, rarely glabrous and finely tuberculate, usually bearing a 2- or 3-lobed aril at the micropylar end [a hilar appendage present in some]; cotyledons ovate or linear, endosperm continuous, fleshy. (Including Asemeia Raf., Galypola Nieuwl., Pilostaxis Raf., Senega Spach, Triclisperma Raf.). LecToTyPE SPECIES: P. vulgaris L., see Britton & Brown, Illus. Fl. No. U. S. ed. 2. 2: 446. 1913. (A pre-Linnaean name from Greek, polys, much, and, gala, milk, in reference to a plant which was thought to increase lactation.) — MILKWort A large (about 500 species), widely distributed genus (throughout Temperate and Tropical Zones, but absent from New Zealand, Polynesia, and the Arctic regions), well known for its small, usually brightly col- ored flowers. In North America the genus is best represented in the eastern part of the continent and in the southwestern United States and Mexico. It is largely absent from the dry, mid-continent region and the Pacific Northwest (Polygala californica Nutt. ex Torr. & Gray from cen- tral, coastal California to southwest Oregon only). Blake (1924) credits 31 species to the southeastern United States; Small (1933) accepts 37 species (in five genera) for the same area. Many are endemic to parts of this region; others occur northward or in the Caribbean region where the genus is also well represented. In his Monographia Polygalacearum, Chodat (1893) proposed sete Polygala into ten sections.2 One of these has been renamed sect. DO- SEMEIOCARDIUM Adema, a change necessitated by the erate that the type of sect. SEMEIocARpIuM (Zoll.) Chodat (Semeiocardium Ar- riensii Zoll., which occurs on Madura and Kangean islands off northeast Java) is actually a member of the Balsaminaceae (see C. A. Backer, * While unambiguously using the rank sectio in the monograph, Chodat refers to freely interchanged on pp. 154 and a in ee 1913 paper, they seem to have been employed as equivalent categories, as Candolle and others had used them earlier (see G. K. Brizicky, Taxon 18: tue bie 9). layer of seed coat hatched tiger ore et = Ai : branch bearing two young capsules developed from cleistogamous flow x 2; 1, cleistogamous ower, perianth and pay pulled away from "style (see 9/9 x aie m-p, ow 274 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Gard. Bull. Singapore 9: 70-72. 1 pl. 1935). Blake’s classification (1916) in which many of Chodat’s sections are treated as subgenera expresses the morphological diversity of the genus better. In the region covered by the North American Flora, Blake (1924) later recognized 13 infra- generic categories (including Badiera DC., subshrubs with subequal sepals, predominantly of West Indian distribution) whose rank, unfortunately, was not indicated. Species belonging to three subgenera are found in the southeastern United States. Only one member of subg. CHAMAEBUXUS (DC.) Blake (keel cristate, calyx caducous, abaxial sepals separate), Polygala pauci- folia Willd., fringed polygala, flowering wintergreen, occurs in our area or elsewhere in eastern North America, although species assigned to this subgenus are known from the southwestern United States. A plant of wide but somewhat sporadic distribution, typically found in moist, decid- uous forests, this species ranges from Quebec (Anticosti Island and Gaspé County), southward through parts of New England, New York, and Pennsylvania and along the mountains to northern Georgia (includ- ing a few stations in western North and South Carolina and eastern Ten- nessee). It extends westward across southern Ontario (as far north as James Bay), Michigan, northern Wisconsin, and Minnesota to southern Manitoba and central Saskatchewan. _ Differing from our other species of Polygala in having six rather than eight stamens, an obscurely bilobed stigma lacking a tuft of hairs, an adaxial nectariferous gland, and seeds with three-lobed arils, P. pauci- folia appears to be closely related to certain other members of the sub- Senus, particularly the European P. Chamaebuxus L. In P. paucifolia, minute cleistogamous flowers are produced on short, erect branches that originate near the base of the upright leafy shoot (see Ficure 1, a, k, | and less frequently, developed from the keel, and have their stamens fused into a unit borne on either side of the ovary and af-bearing branches of P. paucifolia overwinter and are replaced in the spri : ; ape pring with fol hoot produced from the shallow baa oliage developing on a new s ty esse ppb (Chodat) Blake (keel ecristate, calyx persistent flora Walt 3s Meh connate) is represented with us by Polygala grandi- » <” = 28, which ranges along the Coastal Plain (rarely at southeastern North Carolina to Florida and shape, flower color, and y in this region (see Blake, 1971] MILLER, POLYGALACEAE 275 with linear or narrowly linear-lanceolate leaves (vs. narrowly lanceolate to elliptic in var. grandiflora, which probably includes P. cumulicola Small and P. miamiensis Small, both described from collections made in Dade County, Florida) and var. leiodes Blake, with glabrous or sparsely pubescent, linear leaves and sepals lacking nonglandular hairs. While ac- cepting these varieties, Long stresses dark purple wings (vs. greenish or purplish in the others) as the distinguishing character of var. /eiodes. These characters are often difficult to correlate in herbarium specimens, and the P. grandiflora complex in Florida remains perplexing and in need of further study. Our other species, about 30 in number, belong to subg. PoLyGALa (subg. Orthopolygala (Chodat) Blake, keel cristate, calyx persistent in fruit, abaxial sepals separate), which contains at least three-quarters of the species in the entire genus. Although Chodat (1893) provides a lengthy classification of the species in his sect. ORTHOPOLYGALA, some of the categories adopted appear unnatural (e.g., ‘‘Senegae,” containing Polygala Senega L. and P. polygama Walt.), while others contain clearly related species (e.g., ‘““Decurrentes,” comprised of P. lutea L., P. Rugelit Shuttlw. ex Chapm., P. nana (Michx.) DC., P. cymosa Walt., P. ramosa Ell., and P. Balduinii Nutt.). Small (1933), treating only those species found in our area, recognized a monotypic genus, Galypola Nieuwl., for P. incarnata L., Pilostaxis Raf. for species in the Decurrentes, and kept the residuum in Polygala in which seven informal, although named, cate- gories were presented. The characters used to distinguish the two genera are minor and seem better utilized at some infrageneric rank. Eight, possibly more, endemic species belonging to subg. POLYGALA occupy portions of the Southeast. Known only from Florida are Polygala Lewtonii Small, with cleistogamous flowers, according to Blake (1924), reported from the sandy Pinus clausa scrub areas in the central part of the state (Highlands, Lake, Marion, Orange, and Polk counties) and os Rugelii, 2n = 68, which grows on wet, sometimes peaty sands through- out the Florida peninsula. Others, less restricted in distribution, and found mainly at places along the Coastal Plain, include P. Balduinii Nutt. (possibly conspecific with P. Carteri Small, from southern Florida, but this problem needs further study), P. setacea Michx., P. crenata James, with cleistogamous flowers borne on basal, leafless branches, P. Chapmanii Torr. & Gray, 2n = 72, P. leptostachys Shuttlew. ex A. Gray, P. Boykinii Nutt. var. Boykinii, 2n = ca. 28, and from south Florida only, r. Boy- Rinii var. sparsifolia Wheelock (which probably includes P. flagellaris Small and is presumably the same as P. praetervisa Chodat, 2n = 96, since the types cited by Wheelock and Chodat are portions of the same collection). JOURNAL OF THE ARNOLD ARBORETUM | a <= —~ SS sabal “J oS an VS all cai , b, P. Senega: a, habit of plant in flower,. hash a i. poe removed, X 1/3; b, tes Pati flower at an- ewand Ge ein : AEE me pil C; inflorescence, flowers near the apex, fruits inflorescence, flowers near. me seed with two-lobed aril, & 15. e—n, P. lutea: e, x 8: ; € apex, fruits toward the base FIGURE 2. Polygala subg. Polygala. a c which collects ; rt db ie the pollen, x 25; j, ovary in slightly. oblique longitudinal sec- 1971] MILLER, POLYGALACEAE 277 places in the West Indies, and P. alba Nutt., 2n = 24, ca. 72, and 104— 108, which Blake (1924) credits to Louisiana and Wheelock to Ar- kansas, although specimens have been seen only from the region to the west (Puebla, Mexico, to Texas and Arizona, and northward to south- ern Saskatchewan). Other species of less restricted distribution in the Southeast, but which nevertheless occur mainly beyond our area, are P. Senega L., known primarily from calcareous sites from northern Maine across Ontario (north to the James Bay region) to western Alberta, and south to the Carolinas, Georgia, Tennessee, Arkansas, and South Dakota; ‘P. sanguinea L., from Nova Scotia to southern Ontario and Minnesota, south to North Carolina, Tennessee, Alabama, Louisiana, Mississippi, and eastern Texas; and P. verticillata L. var. verticillata (P. Pretzii Pennell) ,* which ranges widely throughout the eastern United States and southern Canada south to Tennessee (Pennell, 1931) and Louisiana. Several varieties (some often treated as species) are usually recognized as dis- tinct from var. verticillata. Those represented in our region (and also to the north) include P. verticillata var. isocycla Fern. (P. verticillata L. var. verticillata sensu Pennell), 2n = 34, a plant of dry soils throughout the Southeast, P. verticillata var. ambigua (Nutt.) Wood, from North Carolina to Arkansas, at stations mostly inland from the Coastal Plain, and P. verticillata var. dolichoptera Fern., from Arkansas. A second group of species, .growing chiefly in savannah often peaty sands, ranges widely across the Gulf and Atlantic ants plains, sometimes occurring west to eastern Texas and north to coastal - New England. These are P. Hookeri Torr. & Gray, from North and South Carolina, Florida, and Louisiana (to Texas, Blake, 1924); P. brevi- folia Nutt., from New Jersey, North Carolina, Florida, Alabama, and Mississippi: P. nana (including P. arenicola Small), 2n = 68, tro South Carolina (Lexington Co.) and Tennessee (Rhea Co.), Georgia, Florida (widespread), west to Louisiana, eastern Texas (and Arkansas, Blake, 19 924); P. ramosa, 2n.= 68, from New Jersey (Small, 1933) and Delaware to Florida, west to eastern Texas; P. cymosa, from Delaware to Florida, west to eastern Louisiana; and P. lutea (with orange, rarely yellow flowers), 2n = 68, from Long Island to Florida, west to eastern Louisiana. The third group, species found both at Coastal Plain Loilittes and more or less widely in the interior of the continent, includes the following: P. polygama Walt. (including P. aboriginum Small, according to James), *See Fe mald (1938) and Pennell (1933 & #04 for two different views on the interpretation of the Linnaean type of this speci en Arch a ne See A riser aren eet A FS tion, at left, ovule with immature aril, at right, ovule showing funiculus yoo diagrammatic , X 20; k, ovule with immature two-lobed aril, X 25; 1, fruit an Persistent calyx, corolla fugacious, X 6; m, seed with two-lobed aril, X 25; 2, hatched (semidiagrammatic), X 25. 0, P. nana, arillate seed, X 25. mosa, arillate seed, 25. q, P. cymosa, seed, aril minute or wanting, NX 25. 278 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 2n = 56, with aérial (Robinson, Shaw) or subterranean racemes of cleis- togamous flowers, distributed from Nova Scotia to Florida, west to Texas, and at scattered localities inland on dry sandy soils from southern On- tario to Minnesota, south to Arkansas and northern Georgia (northern and inland material often referable to var. obtusata Chodat, 2n = 56); P. incarnata L. (Galypola incarnata (L.) Nieuwl., our only species with the corolla at least twice as long as the wings), from Long Island to Florida, west to Texas, and inland sporadically to Wisconsin, Iowa, Kan- sas, and Oklahoma; P. cruciata L. (including P. ramosior (Nash ex Robinson) Small), 2n = 36, with a distribution pattern similar to the pre- ceding species, except inland (lake margins, meadows, peaty soils) to only the Upper Great Lakes region (Minnesota, Wisconsin, Michigan), and southward at scattered places in Illinois, Indiana, Ohio, Kentucky, Tennessee, and Alabama (most collections from the north, both coastal and inland, are referable to var. aquilonia Fern. & Schub.); P. Curtissii A. Gray, 2n = 40, from Delaware to Georgia and Alabama, but less fre- quent southward along the Coastal Plain, although occurring inland from northern Alabama and Georgia, north to Ohio and West Virginia; P. Nut- tallii Torr. & Gray, 2n = 46, from Massachusetts south to North Caro- lina and Georgia, inland to Tennessee and Kentucky (also Alabama, Mis- SISSIPP1, and Arkansas, Blake, 1924, Small, 1933); and P. mariana Mill. (possibly including P. Harperi Small, which according to Fernald, Gray’s Man. Bot. ed. 8. 956, 1950, ranges from southeastern Virginia to Florida, west to eastern Texas), 2n = 34, from New Jersey to Florida, west to the inner tangential wall, extend along he outer tangential wall. Shrinkage of her wall in a predetermined zone on its currence of partially or entirely tetraspo- here flowers of certain species indicates that of Po . : abaxial microsporangia. ygala are derived by suppression of the two The pollination mechanism ; : ; little studied prey it so mM species of Polygala is an interesting but enkatesh’ Sails commonly occur. In the Indian : work indicates that autogamy may 1971] MILLER, POLYGALACEAE 279 However, whether these species are proterogynous or proterandrous, and therefore suited to cross-pollination as well, was not mentioned. In many of our species self-pollination seems to occur also but by a somewhat dif- ferent mechanism. For example in P. Jutea, the apical stigmatic lobe ends in a tuft of hairs which catches pollen shed from the eight tightly surrounding anthers (Ficurre 2, g). At this stage the stigmatic lobe is bent away from the pollen mass and toward the base of the style (FIGURE 2, h). Later, as the flower ages, the stigmatic lobe is reoriented so that the stigmatic surface is presented to the accumulated pollen (FicurRE 2, i). Within a single flower, however, the relationship between the time of anther dehiscence and the period of stigma receptivity is not known pre- cisely, so it is possible that the stigma is receptive at the time pollen is shed, and autogamy occurs only if cross-pollination by insects has not. Observations on living plants need to be made. The relatively large flowers of Polygala paucifolia, in contrast, seem especially suited to bee pollination, which has been described for the re- lated P. Chamaebuxus (Faegri & van der Pijl). The two lateral, petaloid sepals and the keel of these species (and other members of the Polygala- ceae) are analogous to the wings (alae) and standard (vexillum) of papilionaceous legumes, and the pollination mechanism in certain members of both families has much in common. A bee in search of nectar pro- duced at the base of the ovary contacts the rigid style after forcing down the hinged, bowl-like apical appendage of the keel. In P. paucifolia this is crested and perhaps acts as a landing pad. Pollen accumulated on the horizontal surface at the end of the style is deposited on the underside of the insect, which leaves behind foreign pollen obtained from previously visited flowers. Outgrowths at the micropylar end of the ovule, usually termed “arils,” but perhaps more properly called arillodes (see L. van der Pijl, Acta Bot. Neerl. 6: 618-641, 1957), since they develop from tissues of the outer integument (Bresinsky), not from either the funiculus or the hilum, may be of diagnostic size and shape in certain species. These two- or three- lobed structures, often composed of large, hyaline cells, are said to be im- portant in local dispersal by ants that utilize their contents as a food the stages between prometaphase I and telophase I and of the phases of meiosis II (Lewis & Davis). Speciation through polyploidy, aneuploidy, and perhaps hybridization is suggested by the reported chromosome num- n = 14, 16, 18, 24, 28, 30, 32, 34, 36, 38, 40, 42, 44, 46, 48, 52, 56, 68, 72, ca. 84, 96, and 104-108, with 27 = 34 being the most fre- quently obtained count (9 spp. ). aes The genus is of little commercial importance. Several species, includ- 280 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 ing Polygala myrtifolia L., from South Africa, P. paucifolia, and fr Chamaebuxus, are sometimes cultivated as ornamentals, and preparations containing saponins from the thick, fleshy root of P. Senega are used medicinally as expectorants for treatment of various forms of bronchitis and asthma. Until recently, most senega root came from Manitoba and Saskatchewan, Canada (Gillett). The species is cultivated in Japan and India for its roots; elsewhere naturally occurring stands are harvested. Polygala butyracea Heck., from tropical Africa, is reported to be a source of fiber used locally for making cloth and other items (Hutchinson). REFERENCES: Under family references see BLAKE, BOLKHOVSKIKH, CHopAT (1891b & 1893), Farcrt & VAN DER Prjt, Gray, KnutH (1904), Lewis & HEerrERA-MACBRYDE, Ricker, Rostnson, SMALL (1933), Torrey & CRAY (1838 & 1840), and VALENTINE & WEBB, ApeMA, F. A review of the herbaceous species of Polygala in Malesia (Polyg- alaceae). Blumea 14: 253-276. 1966 [1967]. [14 spp. in 2 sections; sect. PSEUDOSEMEIOCARDIUM, sect. nov. | AHLEs, H. E., & A. E. Raprorp. Species new to the flora of North Carolina. Jour. Elisha Mitchell Sci. Soc. 75: 140-147. 1959. [P. grandiflora, 143.] Baker, M. F. Polygalas in Florida. Am. Bot. 39: 100-104, 1933. [12 spp.] BaLansarp, J., & F. PELLISSIER. Les saponines de Quillaya [sic] et Polygala agents de proliferation cellulaire. Compt. Rend. Soc. Biol. Paris 137: 523, 524. 1943. [Experiments with Hedera helix. | BENNETT, A. W. Conspectus Polygalarum Europaearum. Jour. Bot. 16: 241- 246, 266-282. 1878. BLAKE, S. F. A new form of Polygala polygama. Rhodora 17: 201, 202. 1915. [P. polygama f. obovata from Florida and Louisiana; considered distinct and renamed P. crenata by James. ] ——. A revision of the genus Polygala in Mexico, Central America, and the . Gray Herb. 47. 122 pp. 2 pls. 1916. i BRAGA, R. Notas palinologicas, IIT. Contribuigéo ao estudo da morfologia polinica do género Polygala L. Nota previa. Tribuna Farmacéut. 29: 54, 55. 1961.* [28 Spp., pollen either equiaxial or longaxial. | BRESINsRY, A. Bau, Entwicklungsgeschichte und Inhaltsstoffe der Elaiosomen. Bibliot. Bot. 126; 54 Pp. 1963. [Scattered references to Polygala through- out, figs. 28, 33,] Carman, E. S. The flowering wi i den 11: 597. 1800. [P poles mee or fringed polygala. Am. Gar CatrorINt, P. E. Senega (drug) or Polygala virginiana (Senega root), adultera- tions and substitutions, 1. (In Italian.) Fitoterapia II. 26: 372-387. 1955*; 2. Ibid. 609-614 * [Includ gy.] CHopar, Rg Notice sur les Polygalacées & synopsis des Polygala d’Europe et Orient. Bibliot. Univ. Arch. Sci, Phys. Nat. III. 18: 281-299. 1887. ape et critique des Polygala Suisses. Bull. Trav. Soc. Bot. Geneve > 123-185. 7 pl. 1880. [8 spp. in two “sous-genres.”” ; ares — Sur quelques Polygala espagnols. Bol. Soc. Aragon. Ci. at. hae 154-166. 1013, fs spp. in three sections or “sous-genres.”’] 1971] MILLER, POLYGALACEAE 281 ——. Die geographische sapere der Polygala-Arten in Afrika. Bot. Jahrb. fares 111-123. CorNER, J. J., J. B. HARBORNE, = . Humpuri 1ES, & W. D. Ottis. Plant poly- phenols. on ae hydroxy cinnamoyl esters of Polygala Senega root. Phy- tochemistry 1: 73— sk 1962. CorrELL, D. S., & M. C. JoHNston. Manual of the vascular plants of Texas. xv + 1881 pp. Frontispiece + 1 map, Renner, Texas. 1970. [Polygala, 25 spp., 915-923. ] Duncan, W. H. Preliminary reports on the flora of Georgia — 4. Notes on the distribution of flowering plants erie species new to the state. Canahie 15: 145-159. 1950. [P. Senega, 1 Eaton, R. J. A new color form of Polygala paucifolia. Rhodora 54: 27. 1952. f. caerulea, f. nov., corolla caerulean blue. ; Persistence of a forms of Polygala paucifolia. Ibid, 27, 28. [White and blue color form FERNALD, M. L. The Sale snakeroot in Maine. Rhodora 4: 133, 134. 1902. EF Senega. ] W species, varieties and transfers. bid. 40: 331-358. pls. 497-507. 1938. (Reprinted in Contr. Gray Herb. 122.) [General notes on P. ver- ticillata and its typification, incl. discussion of var. isocycla, var. nov., 334- 338, pl. 501.] - Noteworthy plants of southeastern Virginia. [bid. 364-424, 434-459, 467-485. pls. 509-535. 1938. (Reprinted as Contr. Gray Herb. 123.) [Notes on P. verticillata, including var. dolichoptera, var. nov., and var. ambigua, 439, 440, pl. 525. ] . G, ScHuBErT. Studies of American types in British herbaria. Part Il. Some Linnaean species. Jbid. 50: 154-172. pls. 1097-1102. 1948. (Reprinted as Contr. ida ao 167.) [P. cruciata var. aquilonia, var. nov., 163-167, pls. 1099, 1100.] Fox, W. B., R. K. Goprrey, rie: L. Biomeuist. Notes on distribution of North Carolina plants — II. Rhodora 52: 253-271. 1950. [P. Curtissii, 263. FREEMAN, O. M. Notes on the flora of Polk County, North Carolina. Castanea 20: 37- 37. 1955, [P. Curtissia, P. paucifolia, P. Senega, 49. otes on some plant associations in Greenville and Pickens counties, South Carolina. /bid, 23: 46-48. 1958. [P. cruciata, 47.] ae - M. The milkworts of Canada. Res. Branch Canada Dep. Agr. gr. 5. 24 pp. 1968 [7 spp ae Db: Ry. La Sie he de Polygala Chamaebuxus L. Bull. Soc. Neuchateloise Sci. Nat. III. 78: 161-167. 1955. Cytology of Polygala. Nature 188: 604, 605. 1960. [Counts for 11 Spp. Gronpona, E. M. Las especies del eee vine en la provincia de Buenos Aires. Darwiniana 6: 83-121. 1942. [13 a Las especies Argentinas te género Pobnile Ibid. 8: 279-405. 1948. 4 spp.] ~———. Nuevas especies de Polygala del Brasil. Jbid. 9: 24-39. 1949. [7 spp.] Hanoi, J. W. Color variation in Polygala lutea. Castanea 26: 102. 1961. [Plants with lemon-yellow flowers (typical color orange) found in Bruns- wick Co., N.C. Hox, T. Medicinal ae - North America: 4. Polygala rg L. Merck’s Rep. 16: 155-157. 1907 282 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 ——. Morphology of North American species of Polygala. Bot. Gaz. 88: 167— 185. 1929. [P. Curtissii, P. incarnata, P. lutea, P. mariana, P. Nuttalli, P. paucifolia, P. polygama, P. sanguinea, P. Senega, P. verticillata var. am- igua. Houzner, G. Die Caruncula der Samen von Polygala. Flora 90: 343, 344. 1902. [P. vulgaris. | Hooker, W. J. Polygala paucifolia. Bot. Mag. 55: pl. 2852. 1828. Howett, J. T., & D. M. Porter. The plant genus Polygala in the Galapagos Islands. Proc. Calif. Acad. Sci. IV. 32: 581-586. 1968. [3 spp., 2 vars., all endemic. | James, C. W. Notes on the cleistogamous species of Polygala in southeastern United States. Rhodora 59: 51-56. 1957. [P. arboriginum = P. polygama, P. crenata replaces P. polygama f. obovata, and notes on P. Lewtonii.] Karsmark, K, A, Om gyneciemorfologien och embryologien hos Polygala co- mosa, Farmacevt. Revy 12: 12. 1933.* Ketett, A. Polygala paucifolia (Willd.). New FI. Silva 4: 201, 202. 1932. [See also R. B. Smith, Polygala pauciflora [sic.]. Zbid. 131. pl. 37.] Laxketa, O., & F. C. CratcHeap, Annotated checklist of the vascular plants of Collier, Dade, and Monroe counties, Florida. viii + 95 pp. Fairchild Trop. Gard, & Univ. Miami Press, Coral Gables. 1965. [Polygala, 13 spp., 51, 52.] Larsen, K. On the cytological pattern of the genus Polygala. Bot. Not. 112: 369-371. 1959, Larson, D. A., & J. J. SkvarLA, The morphology and fine structure of pollen of Polygala alba Nutt. and P. incarnata L. Pollen Spores 3: 21-32. 1961. [Also notes on Krameria lanceolata pollen in which the exine is stratified m contrast to the homogeneous exine of the Polygala species studied. | Levyns, M. R. The species of Polygala in the southwestern Cape Province. Jour. S. Afr. Bot. 21: 9-50. 1955. [26 spp. | ee W. H., & S. A. Davis. Cytological observations of Polygala in eastern [P. Boykinii var. Boykinii, vat. : nov., and var. sparsifolia Wheelock; P. grandiflora Dar var. angustifolia Torr. & Gray, and var. leiodes Blake, McNEILL, J. Taxonomic and a 23-34 in V. H. Heyvw es. E ; ape Mrranpa, F., & % ped Monogr. 13: 119-166. 1943. [P. Curtissii perate regions of eas ta in Veracruz, MUKHERJEE 2: 58-76, ee ee of Polygala, Bull. Bot. Soc. Coll. Sci. Nagpur Bhar a Saag notes on new and old genera of plants—I. Am. ala, Sener, Fri wntnt: 1914. [Anthalogea Raf., Galypola, gen. nov., Polys- Osor 4 Semese Trichlisperma [sic], 178-181.] iopetes a Seni JR., et al. The dispensatory of the United States of 1225, 1226] * Xvi + 2139 pp. Philadelphia & Montreal, 1955. [Senega, 1971] MILLER, POLYGALACEAE 283 PeaTTIE, D. C. The Atlantic coastal plain element in the flora of the Great kes. Rhodora 24: 57-70, 80-88. 1922. [P. cruciata from Grand Rapids, Michigan, and from the region adjacent to Lakes Erie and Michigan, 88.] PENNELL, F. W. “Polygala verticillata’ in eastern North America, Bartonia 13: 7-17. pls. 2, 3. 1931 [1932]. [P. ambigua, P. Pretzii sp. nov., P. ver- ticillata. | . Polygala verticillata and the problem of typifying Linnean species. Ibid. se 38-45 ——. On 6 typification of Linnean species as illustrated by Polygala ver- ticillata, bc dore 41: 378-384. 1939. [P. ambigua, P. Pretzii, P. verticil- lata. Reeves, J. H., Jk. Megasporogenesis and a in ae ec polyg- ama. = Coop. Wildlife Res. Unit Release 57. 3 i sporogenesis and megagametogenesis in sense polygama. (Abstr “Virginia Jour. Sci. IT. 8: 303. 1957. [Polygonum type of embryo sac in both cleistogamous and chasmogamous flowers. | Roginson, B. L. Polygala polygama var. abortiva sao an autumnal state. Rhodora 2: 242, 243. 1900. [Racemes of cleistogamous flowers borne in upper leaf axils during fall. SHa4H, C. S., & P. N. KHANNA. Pharmacognostic comparison of Indian, Pakistan & Delhi “senegas.” Jour. Sci. Indus. Res. 18C(7): 121-126. 1959.* [P. ga. & M. K. PANDHE. a seo of Polygala abyssinica R. Br. ore Naa Pharm. 22(3): 66-68. 1960. L. S. Vyas. Pharmacognosy of eae (ont Linn. and compari- son with eins I, P. Ibid. 19(10): 224-227. 1957.* SHaw, C. H. The ee structure of the flowers in Polygala polygama and P. paucifilora aoe a review of cleistogamy. . Bot. Soc. Penn. 1: 122-149. dg: 2 17. 1901. [Also issued as Contr. Bot. Lab. Univ. Penn. 2: 122-149. pls. 16, 17. 1901.] Sims, J. rene Senega. Bot. Mag. 26: pl. 1051. 1807. SMALL, J. K. A new candy-root from Florida. Torreya 26: 92, 93. 1926. [P. aboriginum, sp. nov., considered conspecific with P. polygama by James. | Sougces, R. Embryogénie des Polygalacées. Développement de l’embryon ches le ac see L. Compt. Rend. Acad. Sci. Paris 213: 446-448. 1941 “Se D. A contribution to the esata od of the genus Polygala. Proc. Natl. Acad. Sci. India 12: 1-6. ene = Beitrage zur or oe reine Senega L. Ber. Deutsch. . Ges. 30: 43-116. 1920. Daven, L. A. Plants used as oe by ants southeastern tribes. xi + 88 pp. Cambridge, Mass. 1940. [P. lutea, THorNE, R. F. Vascular plants meatier ie from Georgia. Cas- tanea 16: 29-48. 1951. [P. leptostachys, P. Nuttallii, 39.] The vascular plants of southwestern Georgia. Am. Midl. Nat. 52 257-327. 1954. [Polygala, 14 spp., 300. VENKATESH, C. S. The poral dehiscence of hae in Polygala. (Abstr.) Proc. Indian Sei. Congr. Assoc. 42(3): 238. oe lee special mode of dehiscence - anthers of Polygala and its sig- nificance in autogamy. Bull. Torrey Bot. Club 83: 19-26. 1956. [4 SPP-, n. | all Asia 284 JOURNAL OF THE ARNOLD ARBORETUM [voL, 52 WuHeE Lock, W. E. The genus gee in North America, Mem. Torrey Bot. Club 2: 109-152. 1891. [38 sp Wuerry, E. T. The presence of ae methyl salicylate in some American species of Polygala. Jour. Wash. Acad. Sci. 17: 191-194. 1927. ARNOLD ARBORETUM Present address: RVARD UNIVERSITY DEPARTMENT OF BOTANY UNIVERSITY OF NORTH CAROLINA CHAPEL HI, N.C. 27514 1971] BHANDARI, MAGNOLIALES 285 EMBRYOLOGY OF THE MAGNOLIALES AND COMMENTS ON THEIR RELATIONSHIPS ! N. N. BHANDARI COMMENTS ON FAMILIES Austrobaileyaceae. The genus Austrobaileya has been included in Magnoliaceae (White, 1933), in Austrobaileyae, as a subfamily of the Dilleniaceae (Croizat, 1940), Monimiaceae (see Bailey & Swamy, 1949), or as a separate family Austrobaileyaceae (Croizat, 1943). According to Bailey and Swamy (1949) the presence of monocolpate pollen, ethereal oil cells, and absence of raphides negate any relationship with the Dillenia- ceae. Since this genus has unilacunar nodes, it falls in category A of di- cotyledonous families of Bailey and Swamy (1950) and thus has no close affinities with the Magnoliaceae. Bailey and Swamy (1949) remarked that the totality of evidence provides no justification for excluding Austro- baileya from Monimiaceae unless the concept of the family is narrowed to exclude such genera as Trimenia, Piptocalyx, and Amborella (see also Bailey & Swamy, 1948). However, in a subsequent paper discussing the relationships of the Monimiaceae (Bailey & Swamy, 1950) they include this genus in a distinct family, the Austrobaileyaceae, closely related to Monimiaceae, Magnoliaceae, Degeneriaceae, and Annonaceae. Previously the Magnoliaceae included a number of genera of doubtful affinities like Drimys, Schisandra, Iilicium, Trochodendron, Tetracentron, and Euptelea (see Bentham & Hooker, 1862-67; Engler & Prantl, 1887-1909; Hutch- inson, 1959; Rendle, 1952). Dandy (1927) has circumscribed the fam- ily to include ten genera. Kapil and Bhandari (1964) have compared morphological and embryological characters of Magnoliaceae, Schisandra, and Illiciaceae and supported the removal of Schisandra and its allies to Schisandraceae, and Jllicium to Mliciaceae (see also Bailey & Nast, 1948; Gifford, 1950; Lemesle, 1955; Ozenda, 1946; Smith, 1947). Drimys has also been rightly removed to a separate family, the Winteraceae (Bhan- dari, 1963; Bhandari & Venkataraman, 1968; Dandy, 1933; Smith, 1942, 1943; Van Tieghem, 1900). Similarly the removal of Trochodendron and Tetracentron (Bailey & Nast, 1945; Croizat, 1947; Nast & Bailey, 1945, 1946; Smith, 1945) and Euptelea (Lemesle, 1946; Nast & Bailey, 1946; Smith, 1946) to their representative family has been amply justified. The ee conclusions are also corroborated by pollen morphology (Erdtman, woe). The family Degeneriaceae was established by Bailey and Smith (1942) and has been recognized as a distinct family of the magnolian stock (see *Continued from volume 52, p. 39. 286 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Eames, 1961), closely related to the Magnoliaceae and Himantandraceae (Bailey, Nast, & Smith, 1943; Swamy, 1949). Hutchinson (1959), how- ever, considers that Degeneria is closely related to Exospermum and Zygo- gynum and should, therefore, be included in the Winteraceae. Bhandari (1963) has compared the morphological and embryological features of the Winteraceae and Degeneriaceae, and remarked “‘. . . Degeneria dif- fers from the Winteraceae in many important features like the perianth, stamens, pollen grains, endosperm, embryo, seed coat and floral and vegetative anatomy, and is therefore, rightly placed in a separate mono- generic family, the Degeneriaceae (Bailey and Smith, 1942).” Swamy (1949) has concluded that Degeneriaceae, Himantandraceae, and Mag- noliaceae are distinct but closely related families. re : y ; Monocolpate pollen (occasionally acolpate in nnonaceae) in which the generative cell is cut off towards the distal end; oe - In Magnoliaceae it is n = 19; in Degeneria- ily tiaeee pir in Annonaceae it varies from n — 7 , 8, or 9. The fam- Ee pb resembles Degeneriaceae however, in possessing ” eich ore, quite evident that although these families are closely y cannot be arranged in a linear phylogenetic sequence. They 1971] BHANDARI, MAGNOLIALES 287 have some features common to all, others overlapping with either of the families, and still others unique to each one. It may be concluded that they are closely related but distinct families of the magnolian complex. Winteraceae. Bentham and Hooker (1862-67) recognized one genus Drimys, and along with Jilicium placed it in the tribe Wintereae of the Magnoliaceae. Van Tieghem’s (1900) was the first extensive survey of the family. He proposed the group Homoxylées to include all the vesselless dicotyledons and Drimyacées to comprise the five genera. Pseudowintera, the sixth genus, was added to the Winteraceae by Dandy (1933). The Winteraceae have unique features such as the trends of specializa- tion of the conduplicate carpel; the primitive stamen; permanent tetrads (elsewhere present only in the Lactoridaceae and Annonaceae) with pol- len having the generative cell cut off towards the proximal face; extensive fibrous endothecium, monoporate pollen with conspicuous to minute reticu- lations; phenolic compounds in the outer epidermis of the outer integument; distinctive endosperm, embryo, and seed structure (see also Bailey & Nast, 1945) Smith (1943a,b), however, agreed with Burtt (1938) in transferring Tetrathalamus montana from the Guttiferae but merged this genus with Bubbia as B. montana and therefore, recognized only six genera in the Winteraceae. Hutchinson (1959) and Barkley (1966) consider Tetrathala- mus to be deserving of generic rank and the latter author further favors the recognition of Wintera and Lassonia (= Magnolia, see Willis, 1966) as winteraceous genera. Hutchinson ( 1959) is of the opinion that De- generia, the monotypic genus of the Degeneriaceae (Bailey & Smith, 1942), is closely related to Exospermum and Z ygogynum and should also be in- cluded in the Winteraceae. Embryological information for Tetrathalamus, Lassonia and Wintera is lacking and therefore, any discussion pertaining to the taxonomic place- ment and relationship of these genera must await such data. Bhandari (1963) and Bhandari and Venkataraman (1968) have shown that De- generia differs (Swamy, 1949) from the Winteraceae in many important aspects such as the perianth, stamens, pollen grains, endosperm, embryo, seed coat, and floral and vegetative anatomy, and they support Bailey and Smith (1942) in thinking this genus is rightly placed in a separate family, the Degeneriaceae, and deny any close affinities with the Winteraceae. Bhandari and Venkataraman (1968) considered that J/licium differs from the Winteraceae in having vessels in the xylem; unilacunar node; no dif- erentiation in calyx and corolla; endothecium not extending towards the connective tissue; 2-layered glandular tapetum; pollen grains shedding individually, tricolpate pollen; closed sessile carpel; ephemeral antipodal cells; Asterad type of embryogeny; and seed structure. These dissimilari- ties obviously preclude any possibility of Winteraceae being related to Illicium and justify its separation to Illiciaceae (see also Bailey & Nast, 1945; Erdtman, 1952). Cytologically also the family is distinct. The basic number x = 13 in the species of Drimys section TASMANIA is similar only to that of [/licium 288 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 floridanum which is however of secondary origin by aneuploidy. No such evidence is available in Winteraceae. Secondly, the section WINTERA of Drimys and Pseudowintera have n = 43. Taking into consideration the sum total of evidence from morphology, vegetative and floral anatomy (Bailey, 1944; Bailey & Nast, 1943a,b; 1944a,b; 1945; Nast, 1944), and embryology (Bhandari, 1963; Bhandari & Venkataraman, 1968; Sampson, 1963; Swamy, 1952), the Wintera- ceae form a distinct family of magnolian alliance but not closely related to any other existing family. Myristicaceae and Canellaceae. The Myristicaceae is a homogeneous taxonomic unit. Because of the meager embryological information, not very dependable conclusions can be drawn. Joshi (1946), however, con- siders that the family is related to Annonaceae, and perhaps the ruminate endosperm and arillate seeds add further support to this conclusion. The family Canellaceae has also been placed in either Parietales along with Violaceae, Bixaceae, Flacourtiaceae and Koeberliniaceae or in the woody Ranales near Myristicaceae, Illiciaceae, Schisandraceae, and Eupo- matiaceae (Engler, 1964; Hutchinson, 1959). Wilson (1960) in a compara- tive study of wood anatomy concluded that the family is nearly related to Eupteleaceae, Eupomatiaceae, Illiciaceae, and Schisandraceae. Bessey (1915), Vestal (1935), and Wettstein (1935) regarded the families Myris- ticaceae and Canellaceae as closely related. Although the two families have a number of differences they also possess certain common features such as the simultaneous cytokinesis in the microspore mother cells; the generative cell cut off towards the proximal pole; anatropous, bitegmic, and crassinucellate ovules; Polygonum type embryo sac; ruminate endo- sperm; paratracheal Probably they had a c unidirectionally, Schisandraceae and Illiciace and Hooker (1862-1883), En (1952) included Schisandra an ommon ancestral stock from which they deviated ae. Most taxonomists, such as Bentham gler and Prantl (1889-1897), and Rendle Nea ; d Kadsura in a tribe, Schisandreae or a sub- amily, Schisandroideae, of the Magnoliaceae. McLaughlin (1933), Whit- e (1933), Lemesle (1945, 1955), Ozenda (1946), Smith (1947), and en ey and Nast (1948), on the basis of morphology, wood anatomy, an chromosome number, have concluded that Schisandreae should be raised to family rank as the Schisandraceae, This suggestion has been accepted and incorporated in most taxonom} . mic t 1951) Hutchinson (1959), and reatments as that of Lawrence ( ’ 1971] BHANDARI, MAGNOLIALES 289 1 to 7 traces and embedded microsporangia; bilayered glandular tapetum with binucleate cells; Ubisch granules; monocolpate pollen with smooth exine; unilocular ovary with 2 to 6 ovules; vascularized outer integu- ment; testa differentiated into outer fleshy and inner stony regions; multi- lacunar and multitraced node; stipulate leaves; and m = 19 as the basic chromosome number; and supported the exclusion of Schisandra and Kad- sura to their respective families (see also Kapil & Jalan, 1964). Bhandari and Venkataraman (1968) have shown that embryologically J/licium has no affinities with Drimys (see also Winteraceae) with which it was as- sociated in the tribe Magnolieae of the Magnoliaceae (Bentham & Hooker, 1862-1868). Kapil and Jalan (1964) evaluated the morphological, anatomical and embryological features of the Schisandraceae and Illiciaceae. The Schisan- draceae possess the following characters in contrast to Illiciaceae: (a) clim- bers vs. trees or shrubs, (b) eustelic stem with well developed pericycle vs. pseudosiphonostele with poorly developed pericycle, (c) 3-traced uni- lacunar vs. 1-traced unilacunar node, (d) alternate leaves vs. pseudoverti- cillate, (e) haplocheilic stomata vs. syndetocheilic, (f) non-pitted sclereids with crystals vs. pitted sclereids without crystals, (g) unisexual vs. bisexual flowers, (h) spirally arranged carpels without style vs. whorled carpels with style, (i) stamens monadelphous vs. stamens free, (j) hexacolpate vs. tricolpate pollen, (k) embryo sac Polygonum, Oenothera or modified bisporic (see Swamy, 1964) type vs. Polygonum type, (1) Onagrad type of embryogeny vs. Asterad type, and (m) fruit a berry with succulent peri- carp vs. fruit a follicle with sclerotic pericarp. They (Kapil & Jalan, 1964) concluded that these two families deviate in a large number of characters and there seems to be no close relationship between the Schisandraceae and Illiciaceae as suggested by Whitaker (1933), Smith (1947), and Bailey and Nast (1948). Eames (1961) proposes that Schisandraceae and IIlicia- ceae, though more specialized, are closely related to the Magnoliaceae and it is possible that all three families probably have been derived from a common ancestral stock. On the other hand, Smith (1947) remarked that Jilicium has no close allies other than Schisandra and Kadsura; At the same time he emphasized that the two groups have spe- cialized along different lines and have retained certain primitive features. It is impossible to indicate which is the more primitive. Smith’s remarks made about two decades ago, and those of Eames (1961), and Bailey and Nast (1948) seem unfounded in the light of embryological investigations carried out recently. Further, the recent evidence from the karyotypic analysis for Illiciaceae (Stone & Freeman, 1968) and Schisandraceae (Stone, 1968) clearly indicates that Schisandraceae differs from Illiciaceae in having a nearly symmetrical karyotype and lacking subterminal chromo- Somes. It may very well be that Schisandraceae and Illiciaceae again Tep- resent the relics of the extant magnoliales much like many other families. Such as Winteraceae, Eupteleaceae, and Lactoridaceae. Monimiaceae and allies. The Monimiaceae sensu lato included a large 290 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 number of genera having doubtful affinities such as Amborella, Austro- baileya, Idenburgia, Scyphostegia, Trimenia, Piptocalyx, Calycanthus, and Gomortega (see Money, Bailey, & Swamy, 1950). According to Money et al. (1950) Amborella has characters resembling those of members of the Monimiaceae such as spiral arrangement of leaves, bracteoles, and tepals; the form and vascularization of the carpels; pollen morphology; fruit morphology; absence of ethereal oil cells; pres- ence of multicellular hairs and hippocrepiform sclereids; and absence of pericyclic fibers in the stem. However, it differs in the orientation of anatropous ovules, narrow rays, and a single arc-shaped leaf trace. There- fore, its position in Amborellaceae, closely related to Monimiaceae, is justified (see also Bailey & Swamy, 1948). perhaps with the Lauralian line of Eames (1961), and the theory that they might have had their origin from a common ancestry seems justified. Money et al. (1950) also included these families in their group having monocolpate or its derived forms of pollen grains, ethereal oil cells, and unilacunar node modial tapetum; successive two-traced carpel that therefore, the two families are closely : Similarly Calycanthaceae are closely ré- eriplasmodial tapetum; multinucleate tape- 1971] BHANDARI, MAGNOLIALES 291 tal cells; anatropous, bitegmic, and crassinucellate ovules; multicelled ar- chesporium; and Polygonum type of embryo sac. Pollen morphology (Erdt- man, 1952) suggests close affinities with Hernandiaceae, Gomortegaceae, and Monimiaceae. Sastri (1963) has rejected the possibility of any relationship with the Thymeleaceae since the latter family differs in having glandular tapetum with 4-nucleate tapetal cells; simultaneous cytokinesis; multiporate pol- len; single-celled female archesporium; and both seed coats persistent. The resemblances, according to him, may be a case of parallelism. proper. However, there are also some important differences: (a) the tip of the nucellus is broad, so that the micropyle formed by the two integu- ments is very wide, (b) formation of multiple embryo sacs which are haustorial and start invading the nucellus at the 4-nucleate stage, 5 or 6 mature embryo sacs come out of the nucellus and invade even the over- arching funicular tissue; (c) the antipodal cells are not organized and the three nuclei are ephemeral, (d) the endosperm is cellular in contrast to nuclear in rest of the family, (e) although the basic plan of seed struc- ture is the same there are some differences in details. On the basis of such important dissimilarities as these, the present author considers that it would perhaps not be too unnatural to remove this genus to a separate family, the Cassythaceae, as proposed by Bartl. ex Lindley (1833) and re- cently adopted by A. C. Smith (personal communication). Certainly the two families are most closely related and probably derived from the same ancestral stock. Cercidiphyllaceae and Trochodendraceae. Cercidiphyllum has been variously placed either in the Hamamelidaceae (Baillon, 1871; Croizat, 1947; Hallier, 1903; Lotsy, 1911; McLaughlin, 1933; Solereder, 1900), or in Magnoliaceae (Bessey, 1915; Diels, 1936; Hayata, 1921; Hutchinson, 1959; Lemesle, 1946), in Trochodendraceae (Harms, 1897), or in a separate family, the Cercidiphyllaceae (Cronquist, 1968; Harms, 1918; Takhtajan, 1966; Thorne, 1968). In all these treatments the re- lationships have been sought with other genera such as Trochodendron, Tetracentron, Euptelea, Hamamelidaceae, and members of the Magnolia- ceae. Most recent taxonomists (Cronquist, 1968; Takhtajan, 1966; Thorne, 1968) keep Cercidiphyllaceae in the Hamamelidales. Such em- bryological features as single-layered glandular tapetum; absence of isch granules; tricolpate pollen with conspicuously broad colpi; 1-celled female archesporium; nucellar cap formed by both the parietal cell and the nucellar epidermis; pattern of early growth of the endosperm. Con- 292 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 spicuous basal cell not taking part in the formation of embryo; ovules having vascularized tapering projection, and winged seeds in Cercidiphyl- lum indicate that the Cercidiphyllaceae are not related to the Magnolia- eae. The chief evidences for including Cercidiphylium either in Hamameli- daceae or in Cercidiphyllaceae placed close to Hamamelidaceae are the presence of stipules, palmate venation, marginal glands, tricolpate pol- len, resemblances in wood anatomy to Corylopsis, and trilacunar node. The pollen although tricolpate differs basically. Further the resemblances in the wood anatomy may not indicate close relationships with certainty. The wings in the two families differ markedly in histological details (see Swamy & Bailey, 1949). Swamy and Bailey (1949) have taken into account the (i) idioblasts in the leaves, (ii) trends of specialization in the carpels, (iii) pollen grains, (iv) embryology, (v) seed structure, and (vi) wood characters, and they have concluded that the summation of evidence from all organs and parts clearly indicates that C ercidiphyllum cannot comfortably be included in any of the existing ranalian families. They agreed with van Tieghem (1900) that it should be placed in a separate family, the Cercidiphyllaceae. They remarked “Nothing is to be gained by transferring such genera as Tetracentron, Trochodendron, Eup- telea, and Cercidiphyllum into close relationship with the Hamamelida- ceae or Saxifragaceae, since this would merely serve to expand another order into a less homogeneous assemblage.” To this conclusion I might add that the available evidences from embryology, morphology, anatomy, floral morphology, and cytology suggest that retention of this family with- in the Magnoliales would not make the order more heterogeneous than would its exclusion. Trochodendron alon been placed in the Ma (Parietales). On the b bium in both these pulsion of the two g ceae and inclusion 1971] BHANDARI, MAGNOLIALES 293 on the seed formed by the chalazal projection. Therefore, these families seem to be more closely related to each other than to any others in the Magnolian complex (see also Yoffe, 1965). On the basis of morphology and anatomy the Trochodendraceae are closely related to the Tetracen- traceae (Bailey & Nast, 1945; Nast & Bailey, 1945; Smith, 1945). Chloranthaceae. Various views have been expressed regarding the interrelationship of the Chloranthaceae, which have been considered te- lated to Santalineae, Ceratophyllaceae, Piperales, or Ranales (see Bessey, 1915; Swamy, 1953; Hutchinson, 1959). Recently Takhtajan (1966) and Thorne (1968) have included the family in Laurales or Annonales respectively, whereas Cronquist (1968) retains this family in Piperales along with Saururaceae and Piperaceae. Vijayaraghavan (1964) has worked out the embryology of Sarcandra irvingbaileyi and evaluated its relationships with either Santalaceae or Piperaceae and Saururaceae. Ac- cording to him (Vijayaraghavan, 1964) the summation of evidence fails to support any relationship with the families mentioned previously. The Chloranthaceae differ from Santalaceae in including autotrophic plants, with 2-celled pollen; superior ovary with 1 orthotropous bitegmic and cras- sinucellate ovule; a Polygonum type of embryo sac not developing beyond the ovule; and in lacking endosperm haustoria and having an Onagrad type of embryogeny. The family also differs from Saururaceae and Piperaceae in having unilacunar nodes; pollen with reticulate exine; tetrad of megaspores; and cellular endosperm. No close relationship seems to exist between them. The presence of monocolpate pollen, unilacunar node, and ethereal oil cells point towards ranalian affinities (see also Money et al., 1950). Therefore, this family should be included in the Ranales sensu lato and the Magnoliales sensu stricto. ACKNOWLEDGMENTS I am deeply indebted to my teacher, the late Professor P. Maheshwari who initiated me into the field of embryology of the ranalian complex and without whose constant encouragement and inspiration I could not have undertaken the task. I am also grateful to Professor B. M. Johri for his interest and for making possible frequent access to the Maheshwari Memorial Library to consult the literature. My thanks are also due to Dr. A. C. Smith, Dr. A. L. Bogle, and Dr. James Doyle for their valuable comments; and to my colleague Dr. K. M. M. Dakshini whom I ap- proached many times for suggestions and clarifications. I am particu- larly obliged to my wife for her moral support. LITERATURE CITED Apatta, R. D. The occurrence of entire pollen grains inside the carpels of some Anonaceae. Jour. Univ. Bombay 14: 47-48. 1946. Arora, C. M. New chromosome report.’ Bull. Bot. Surv. India 3: 37. 1961. 294 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Asana, J. J., & R. D. Apatia. Contributions to the embryology of the Anona- ceae. I. Artabotrys odoratissimus R. Br. Jour. Univ. Bombay 16: 7-21. 1947. Bartey, I. W. The development of vessels in angiosperms and its significance in propgaets research. Am. Jour. Bot. 31: 421-428. 1944a. comparative morphology of the Winteraceae. III. Wood. Jour. Arnold jee 25: 97-103. 1944b. Additional notes on the vesselless dicotyledon, Amborella trichopoda Baill. Ibid. 38: 374-378. 1957. ast. The comparative morphology of the Winteraceae. I. Pollen and stamens. Jour. Arnold Arb. 24: 340-346. 1943a The comparative morphology of the Winterscens: II. Car- pels. Ibid. 24; 472-481. 1943b. & . The comparative morphology of the Winteraceae. IV. Anat- omy of the node and vascularization of the leaf. /bid. 25: 215-221. 1944a. The comparative morphology of the Winteraceae. V. Foliar epidermis and sclerenchyma. Jbid. 25: 342-348. 1944b. — & he comparative morphology of the Winteraceae. VII. Sum- mary and conclusions. Ibid. 26: 37-47. 1945a. orphology and relationships of Trochodendron and Tetra- centron, Stem, root and leaf. bid. 26: 143-154. 1945b. . rphology and relationships of Jllicium, Schisandra, and Kadsura. I. Stem and leaf. Ibid. 29: 77-89. 1948. & . SMITH. Degeneriaceae, a new family of flowering plants from me Jour. Arnold Arb, 23: 356-365. 1942 ——— & ———, The family Himantandraceae. Ibid, 24: 190-206. 1943. ; ob. Swamy. A new morphological type of vesselless dicotyledon. Jour. Arnold Arb, 29: 245-254, 1 1948. . The morphology and relationships of Austrobaileya. Ibid. 30: 211-226. 1949, & The conduplicate carpel of eh eri and its initial trends of specialization. Am. Jour. Bot. 38: 373-379. & W. P. THompson. Additional notes upon ie angiosperms Tetracen- tron, Trochodendron, and Drimys. Ann. Bot. 37: 503-512. 1918. Bano, H. Nouvelles notes sur les Hamamelidacées. Adansonia 10: 120-137. BaMBACIONI-MEzzertt, V. Ricerche morfologiche sulle Lauracee. Lo sviluppo dell’ovulo e = sacchi pollinici nel Laurus nobilis L. Ann. Bot. Roma 21: 186-204. 193 BANerjl, I., & A. ae . A contribution to the life history of Artocarpus la- : . dian Acad. Sci. B. 39: 128-132, 1954. err A. A list of the orders and families of ie He Canisoepen — rae examples. Hopkins Press, Providence, R.I. e interpretation of th Sci. Mee of the angiosperm flower. Panis Jour. BenTuaM, G., & J. D. Hooker. Genera Plantarum. Vol, I. London. 1862-67. BESSEY C, E. The ph ylogenetic ta nist Bot. Gard. 2: 109-164. 1915, xonomy of flowering plants. Ann. Miss BATTAGLIA, E. Meiosi anormale Soc. Tose. Sci. Nat. 54: 1-22 Buacavatut Kurt 1 Ama, P, R . Micros Presidency College Bot. om Maka "rap themaggi oe in Drimys. nella microsporogenesis di Laurus nobilis L. Atti 1947, 1971] BHANDARI, MAGNOLIALES 295 Buanparl, N. N. Studies in the family Ranunculaceae — III. Development of the female gametophyte in Adonis annua L. Phytomorphology 12: 70-74. 1962 Studies s in the family Ranunculaceae — V. The female gametophyte of Adonis aestivalis L.—a reinvestigation. bid. 13: 317-320. 1963a. eth om eo - Pseudowintera colorata, a vesselless dicotyledon. Ibid. 13: 303-316. 19 . Studies in ee family Ranunculaceae — VIII. Variations in the devel- opment of the embryo sac of Anemone vitifolia Buch.-Ham. ex DC. bid. 15: 285-291. 1965. . Studies in the family Pep erage Embryology of Adonis Dill. ex Linn. Jbid. 16: 578-587. 1966. Magnoliaceae. (Abstract.) P. 11 Jn: Johri . al. eee Seminar comparative embryology of angiosperms. Delhi. 1967a . Degeneriaceae. Jbid. 12. 1967b. © 2 . Schisandraceae. Jbid. 17. 1967f. Illiciaceae. bid. 18. 1967g. Monimiaceae. Jbid. 19. 1967h. Calycanthaceae. Ibid. 20. 1967i. Lauraceae. Ibid. 21. 1967}. Trochodendraceae. Ibid. 22. 1967k. Cercidiphyllaceae. Ibid. 23. 19671. Studies in the family ee Embryology of Anemone Phytomorphology 18: 487-497. & S. Asnant. Studies in the Hee Ranunculaceae — XI. Morphology and embryology of Ceratocephalus falcatus Pers. Beitr. Biol. Pflanzen 45: 271-290. 1968. & R. N. Kapri. Studies in the family Ranunculaceae — VII. Two types of embryo sacs in Trollius Linn. Ibid. 40: 113-120. 1964. —— & R. Venxataraman. Embryology of Drimys winteri. Jour. Arnold Arb. 49: 509-525. 1968. & M. R. ViyavaraGHAVAN. Studies in the family Ranunculaceae — XII. Embryology of Aquilegia vulgaris. Beitr. Biol. Pflanzen 46: 337-354. 1970. BorcMANN, E. Anteil der Polyploiden in der Flora des Bismarckgebirges von ai uguinea. Zeitschr. Bot. 52: 118-172. 1964. RAN Développement des téguments de la graine. (Magnoliacées.) Re- . Bot. 3: 124-126. 1891. } Burtt, "B. L. The taxonomic position of Tetrathalamus. Bull. Misc. Inf. Kew 1938: 458-460. 1938 Camp, W. H., & M. M. Hupparp, Vascular supply and structure of the ovule and aril of peony and of the aril of nutmeg. Am. Jour. Bot. 50: 174-178. 1963, Canricut, J. E. The comparative morphology and relationships of the Mag- noliaceae, I. Trends of specialization in the stamens. Am. Jour. Bot. 39: 484-497. 1952. ~_——.. The comparative morphology and relationships of the Magnoliaceae — IT. Significance of the pollen. Phytomorphology. 3: 355-365. 1953. ~———. The comparative morphology and siatindeliee of the Magnoliaceae — III. Carpels. Am. Jour, Bot. 47: 145-155. 1960. A II] | | I] oo 296 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 . The comparative morphology and seg of the ren IV. Wood and nodal anatomy. Jour. Arnold Arb. 36: 119-139. Some a eo Sapien ee alas morphology in a Ranales. ___ Abstract.) Am. glh2e2652: 1965. —— & M. P. Pap eee Aion of pollen morphology to the phylogeny of the Annonaceae, Eupomatiaceae, and Myristicaceae. (Abstract.) Am. Jour. Bot. 49: 674. 1962. Cartquist, S. Morphology and relationships of Lactoridaceae. Aliso 5: 421- 435. 1964. Cave, M. S. Embryological characters of taxonomic value. (Abstract.) Proc. Ninth Internat. Bot. Congr. 2: 62. 1959. CuHoupHury, J. K., & J. N. Mirra. Abnormal tricotyledonous embryo and the morphological. stinctur of normal fruit and seed of Cimnamomum cam- phora F. Nees. Sci. & Cult. 19: 159. 1953. Cuuns, T., I. Counc, C. Y. Cao, W. W. L. Hu, & S. C. Kwan. Chromosome numbers of the vascular plants of Taiwan I. Taiwania 1: 51-66. 1963. Corner, E. J. H. The annonaceous seed and its four integuments. New Phytol. 48: 332-364. 1949, Courter, J., & C. J. aceon Morphology of angiosperms. D. Appleton & Company, New York. Coy, G. V. Morphology of Dee in relation to phylogeny of angiosperms. Bot. Gaz. 86: 149-171. 1928. Croizat, L. Trochodendron, Tetracentron, and their meaning in phylogeny. Bull. Torrey Bot. Club 74: 60-76. 1947. Cronguist, A, The evolution and classification of flowering plants. Houghton Mifflin Co., Boston. 1968. Dau, O. A. ep R. RowLey. Pollen of Degeneria vitiensis. Jour. Arnold Arb. 46: 308-323, 1965. me J. E. The genera of Magnolieae. Bull. Misc. Inf. Kew 1927: 257-265. ———.. The Winteraceae of New Zealand. Jour. Bot. 71: 119-122. 1933 DanuiNero, C. D., & A. P. Wytie. Chromosome atlas of flowering ‘plants. Allen & Unwin, London: 1955. etl “ vad Systematic embryology of angiosperms. John Wiley & Sons, New ork, Drts, L, Ueber die Gattung Himantandra ihre me iogcg und ihre syste- matische Stellung. Bot. Jahrb. 55: 126-134. : aarteee Syllabus der Pflanzenfamilien, 7 é Gebrueder Borntraeger, erlin caer ak Morphology of the angios perms. McGraw-Hill, New York. 1961. ARLE, T. T. Origin of the seed coats in Magnolia. Am. Jour. Bot. 25: 221, 274, a. — men de of certain Ranales. Bot. Gaz. 100: 257-275. 1938b. . Embryo and endos pos Jour. Bot. 28: 25. 194 1 TAR adil in Ilicium floridanum Ellis. pera A. Syllabus der Pflanzenfamilien (rev. by Melchior). Vol. 2. Jena. ERpTMAN, G. Pollen morphol : sree = Wal rpho ey and plant taxonomy. Angiosperms. Chronic & 7 len ALFE, “Aint f certain genera Incertae sedis suggested ie en are and hiaetisies anatomy. Kew Bull. 17: 249-256. 1971] BHANDARI, MAGNOLIALES 297 Faun, A., & I. W. Battey. The nodal anatomy and the primary vascular cyl- inder of the Calycanthaceae. Jour. Arnold Arb. 38: 107-117. 1957 Farr, C. ee ‘i division by furrowing in Magnolia. Am. Jour. Bot. 5: 377- 395. 19 FRYXELL, P. mi Modes of reproduction in higher plants. Bot. Rev. 23: 135-233. 957. Garrat, G. A. Systematic anatomy of the woods of the Myristicaceae. Trop. Woods. 35: 6-48. 1933a. ——. Bearing of bre anatomy on the relationships of the Myristicaceae. Ibid. 36: 20-44. 3b. Girrorp, E, M. The structure and development of the shoot apex in certain woody Ranales. Am. Jour. Bot. 37: 595-611. 1950. Gruxtant, A. Contributo alla embriogenia del “Cinnamomum camphora” (L.) Eber. et Nees. Bull. Orto Bot. Univ. Napoli 9: 33-40. 1928. Harr, J. B., & E. J. BruzeNBeRG. Contributions to a chromosome atlas of the New Zealand Flora — 2. New Zealand Jour. Sci. 2: 148-156. 1959. . Contributions to a chromosome atlas of the New Zealand Flora — 4. Ibid. 3: 432-440. 1960. Hatter, H. Ueber den Umfang, die Gliederung und die Verwandtschaft der Familie der Hamamelidaceen. Beih. Bot. Centr. 14: 247-260. 1903. Harms, H. Zur Kenntniss der Gattung Cercidiphyllum. Mitt. Deutsch. Dendrol. Ges. 26: 71-87. 1918. Havasu, Y. On the microsporogenesis and pollen eerarneys in the family Magnoliaceae. Sci. Rep. Téhoku Univ. Biol. 27: 45-52. 1960. embryology of the Magnoliaceae sens. lat. i ae ove gametophyte and embryogeny of Jllicium anisatum. Ibid. 29: 27- 33. 1963. : “The embryology of the family Magnoliaceae sens. lat. III. Mega- sporogenesis, female gametophyte and embryogeny of _e liliflora Desrousseaux and Michelia fuscata Blume. Ibid. 30: 89-9 ae B. The natural classification of plants according to the pao system. Pl. Formosa, 10: 97-234. 1921. Hemons, O. Studies on the taxonomy, geographical distribution and embryol- of the set: Siparuna Aubl. Sv. Bot. Tidskr. 25: 202-228. 1935, cee C. B., Jz. Some observations on pollinations and compatibility in Mag- nolia. Proc. rae Acad. Sci. 72: 259-266. 1962. Hers, W. B. Contributions to the life history of Asimina triloba. Ohio Nat. 8: 211-216. 190 See 1, Mature fruit, X 6; m, semidiagrammatic longitudinal section fruit to show embryo, x 6; n, mature leaf, « 1, 1971] ELIAS, GENERA OF MYRICACEAE 317 usually on older wood, the primary bract usually ovate, pubescent, at least near the base; each flower subtended by 6-8 minute linear to acicu- lar secondary bracts that persist and elongate in fruit (see Ficure 3, g-i, k). Fruit a small ovoid-oblong, smooth, shining nut. Type species: C, asplentifolia (L.) L’Her. ex W. Ait. = C. peregrina (L.) Coult. (Name commemorating H. Compton, 1632-1713, an amateur horticulturist and supporter of botany, who also served as Bishop of London.) — Sweet N. A single extant species, Comptonia peregrina (L.) Coult., 2n = 32, sweet fern, found on poor, often rather dry soils in woodlands, clearings, pastures, and pine barrens from Nova Scotia to Saskatchewan, south to Virginia, western North Carolina (but also in four counties in the lower Piedmont), extreme western South Carolina, and locally in northern Georgia and Tennessee, Ohio, northwestern Indiana, Illinois and Minne- sota. The rather attractive low shrub with its fragrant aroma when crushed has been separated into two varieties. The questionably distinct var. aspleniifolia (L.) Fern., of the Coastal Plain pinelands or pine bar- rens from Long Island, New York, to Virginia, is characterized by the minutely puberulent branchlets, sparsely short-puberulent or glabrous leaves, and fruits 3-4 mm. long. The more northern and inland var. pere- grina has more or less pilose branchlets, pilose leaves, and fruits 4-5 mm. ong. The six to eight secondary bracts associated with a single pistillate flower Suggest that the solitary flower is derived by reduction from a few- flowered cyme. The inner secondary bracts would be the only grossly ob- servable remnants of the cyme. A second possibility is that the inner sec- ondary bracts are actually modified perianth parts that are completely lacking in Myrica. A careful anatomical study of the flowers and aments is needed. Comptonia peregrina is the only survivor of a genus that had perhaps a dozen species in the Eocene and Oligocene. Berry (1906) postulated that Comptonia developed in the Upper Cretaceous as a branch from Myrica stock, but supporting evidence is lacking. Numerous Tertiary fossils of Comptonia clearly demonstrate that several species once oc- cupied regions throughout much of the temperate world. In European Tertiary deposits, C. difformis (Sternh.) Berry is the most common fos- sil species of the genus (Kotlaba, 1967). REFERENCES: Under family references see E. C. Appe; L. B. ABBE; HyJELMQvIST, RENDLE, and YOUNGKEN, Berry, E, W. Living and fossil species of Comptonia, Am. Nat. 40: 485-524. 6 1906, , . Dopce, B. O., & J. F. Avams. Notes relating to the Gymnosporangia on Myrica and Comptonia. Mycologia 9: 23-29. pls. 2, 3. 1917. 318 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 FERNALD, M. L. Noteworthy plants of southeastern Virginia. Rhodora 40: 364- Herb. 123.) [Comptonia, 410-412, pl. 514, including C. peregrina var. aspleniifolia (L.) shila comb. nov.] Fox, W. B., R. K. Goprrey, & H. L. Bromouist. Notes on distribution of North Carolina ei Rhodora 52: 253-271. 1950. [C. peregrina, 261. Hetno, H. E. The excised embryo culture method for controlled seedling growth of the sweet fern, Comptonia peregrina, of the family Myricaceae. Proc. Minn. Acad. Sci. 29: 180-184. map. 1961. Korzasa, F. Taxonomic-nomenclatural notes on the fossil Comptonia diffor- mis (Sternb.) Berry and the recent Comptonia a (L.) Aiton. (In Czech; English summary.) Preslia 33: 130-140. RoprIGuEz- BarRUECO, C. The occurrence of nitrogen- nae root nodules on non- leguminous plants, Bot. Jour. Linn. Soc. London 62: 77-84. 1 pl. 1969. ZiecLer, H. thamnien” bei Comptonia peregrina (L.) Coult. Naturwis- senschaften a 113, 114. 1960. [See also Mitt. Deutsch. Dendrol. Ges. 61: 28-31. 1960.] ARNOLD ARBORETUM Harvarp UNIVER CAMBRIDGE, MAssACHUSETTS 02138 1971] DICKISON, DILLENIACEAE, VII 319 COMPARATIVE MORPHOLOGICAL STUDIES IN DILLENIACEAE, VII. ADDITIONAL NOTES ON ACROTREMA WILLIAM C. DickIson I HAVE POINTED ouT during this series of papers that the Dilleniaceae is comprised of genera possessing considerable morphological diversity, a result in part of the wide range of growth habits encountered. One of the most interesting, yet least studied genera in the family, is Acrotrema Jack, a rhizomatous, semi-herbaceous plant. Acrotrema is of interest not only because it is a small, nearly herbaceous plant in an otherwise woody family (the Australian genus Pachynema is also for the most part non-woody), but its constituent species show a remarkable degree of morphological variation (see FicurEes 35—43). From an evolutionary viewpoint, the genus is noteworthy because it is one of the few dicotyledonous herbs in which the vessel elements possess exclusively scalariform perforations (Takhtajan, 1954, p. 119). Further- more, Meeuse (1966, p. 204) suggests that, “the pinnate or bipinnate leaves of the acaulescent Acrotremeae are also suggestive of a fern-frond, lea pteridospermous-cycadopsid phyllome, and one of the possible habit forms of the ‘reconstructed’ group of Protodicots.” The genus has been generally considered to consist of about ten species and a number of varieties, all of which are confined to Ceylon with the exception of A. costatum Jack from Thailand and the Malay peninsula, and A. arnottianum Wight from southern India. Indications are that some of the species are very narrowly endemic and apparently quite rare. The only comprehensive taxonomic treatments available are those pre- pared by G. H. K. Thwaites (1864) in his Enumeratio plantarum Zey- laniae: An Enumeration of Ceylon Plants . . ., by Hooker and Thomson (1872) for J. D. Hooker’s The Flora of British India, in which they com- ment that Acrotrema is, “A remarkable genus of very variable species,” and by H. Trimen (1893) for his A Hand-Book to the Flora of Ceylon. Trimen commented, “Most of the species are ill-defined and extremely variable, and I have failed to obtain any clear idea of their limits. They doubtless hybridise freely.”’ Anatomical information is meager and accu- rate illustrations do not exist. I recently had the good fortune to secure an excellent assemblage of well-preserved specimens of Acrotrema resulting from the collecting ef- forts of Dr. R. D. Hoogland of the Australian National University. The assistance Dr. Hoogland has given me now and in the past is acknowl- edged with much appreciation. Although many of these collections are Currently without names, they are all vouchered and are being studied in 320 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 connection with a new generic treatment for a proposed flora of Ceylon, The present paper draws attention to this uncommon and little-known dilleniaceous genus and provides additional morphological information, including the report of new anatomical features for the family. MATERIALS AND METHODS Liquid-preserved vegetative and floral material was received from Dr. R. D. Hoogland. Hoogland collection numbers 11423, 11460, 11469, 11572, 11575, 11580, and 11583 were collected in Ceylon and have voucher specimens housed at the Commonwealth Scientific and Industrial Research Organization, Division of Plant Industry, Canberra, Australia (cANB); Department of Agriculture, Peradeniya, Ceylon (ppa); and the United States National Museum, Washington, D.C. (us). The Hoog- land collection of Acrotrema costatum (11635) was obtained in Penang, Malaya, and has supporting material in cans. Additional preserved plants of A. costatum were received through the efforts of Mr. K. Chiang at the Botanic Gardens, Penang. The following specimens of Acrotrema at the United States National Museum were also examined; these for the most part are very old sheets with Thwaites collecting numbers: Acrotrema sp. Ceylon: Thwaites CP. 3898; A. appendiculatum Thw. Cey- ites C.P. 239; 4. T. hwaitesii Hook. Ceylon: Thwaites C.P. 3364; Thwaites C.P. 3969; A. uniflorum Hook. Ceylon: Thwaites C.P. 3486; A. Walkeri Wight. Ceylon: Thwaites C.P. 604. Material of A. arnottianum was not available for study. bs NaOH to study vasculation. Epidermal features were examine by making peels from preserved leaves. Pollen was acetylized and mounted in glycerine jelly. The assistance of Miss Marion Seiler, staff artist in the Department of Botany, The Universit of North » . pari ; preciated, : orth Carolina, Chapel Hill, is sincerely ap GENERAL MORPHOLOGY Pee co Acrotrema is composed of perennial herbs with simple leaves. rprisingly, Hoogland (personal communication) mentions collecting 1971] DICKISON, DILLENIACEAE, VII 321 leaves of Acrotrema gardneri (FicurE 39) are narrowly oblong to linear- spathulate whereas those of A. lanceolatum (FIGURE 40) are characteris- tically linear-lanceolate in outline. The margins are entire, serrate, or dentate. Ficures 41 and 42 show specimens of A. Thwaitesii with linear- lanceolate, pinnatifid leaves. The most dissected situation occurs in A. dissectum (Ficurr 43) in which the leaves are bipinnate with the largest segments deeply cleft and one or more pairs of small segments situated between the large ones. Acrotrema lyratum (not pictured) is described as having lyrately lobed leaves with four to five pairs of basal lobes, and A, arnottianum has been pictured as possessing large, entire leaves (Swamy & Periasamy, 1955). A study of the stability of leaf morphology in this genus in addition to chromosome data would be of interest. Thwaites’s collection of A. appendiculatum Thw. (C.P. 3880) is distin- guished by the presence of rows of large, circular cushions between the secondary veins of the leaf (Ficurr 37). The foliage of A. costatum and many of the Ceylonese plants is variegated along the midrib as shown in FicurE 20. The petioles sheath the stem and possess membranous, caducous wings. Young leaves exhibit conduplicate vernation (FIGURE 29) as is characteristic of other Dilleniaceae. The stem is typically very short and in some cases the plants appear acaulescent. Acrotrema costatum (FicuRE 20) has a comparatively elongated axis which is readily divided into a pubescent aérial segment bearing crowded leaves and leaf scars, and an underground rhizomatous portion, With the exception of the primary root, all roots are adven- titious. The endogenous origin of the roots from the stem has been illus- trated in Plate I, Figure 4 of Dickison (1969). The inflorescence is a terminal or axillary raceme, occasionally reduced to a single flower. Each flower is subtended by a membranous bract. The length of the peduncle and degree of pubescence has been used to separate species. On the basis of general floral morphology, A. costatum from Ma- laya (Ficures 9-15) is clearly distinguishable from most species in- habiting Ceylon (Ficures 1-8). In the latter, the androecium is ex- ternally fasciculate with three stamen fascicles alternating with three Carpels (A. Thwaitesii is an exception). Individual stamens possess long filaments and short anthers which open by longitudinal slits (FicurEs 6, 8). The androecium of A. costatum does not show external fasciculation but rather the fifteen stamens arise from the receptacle uniformly around the carpels. Stamens have elongated microsporangia which dehisce by apical pores (FicurE 12). The gynoecium is usually described as con- sisting of three basally connate carpels; however, I have examined ra- cemes in which some or all of the flowers were bicarpellate (FicuRE 11). The morphology of the carpels agrees essentially with previous descrip- tions (Dickison, 1968). Carpels are glabrous and differentiated into a swollen ovary and slender, sometimes recurved, style terminated by an in- distinct stigmatic surface. The two or three conduplicate carpels are lat- erally concrescent (FIGURES 25, 26). At the level of placentation the ven- tral suture opens to produce a central cavity. Progressing distally, the 322 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 carpels close and become totally apocarpous at the initiation of the stylar region. The bicarpellate gynoecium of A. costatum has an oppo- site orientation as interpreted by Wilson (1965). The ovules are arranged in two rows along the placenta and range in number from two to many. Raphides are a characteristic feature in the carpel wall. The fruit is a follicle and the seeds are arillate. The flowers have five pubescent sepals and five glabrous petals which contain numerous raphides. VEGETATIVE ANATOMY Trichomes. The trichome complement of Acrotrema is unusually di- verse and includes types previously unknown for the family. A dense pubescence may cover the aérial stem (F1cuRE 20) and also young floral buds, pedicels and peduncles (Ficures 1, 9, 10). On the leaf, hairs are conspicuously abundant along the major veins of the abaxial surface and margins of the lamina. Foliar trichomes are not uniform in length but are noticeably longer along the veins than in intercostal regions. Compared to other Dilleniaceae, the prevalent trichome type is a uni- cellular, unbranched form. Frequently, however, these hairs are raised on pedestals of enlarged cushion cells (Ficure 24), a feature not typical among other members of the family. A few septate hairs with varying numbers of cells and their outer walls divided into two layers were ob- served (FicurE 23). ee - ha short stalk and spreading, multicellular arms that ter- is on ne a much enlarged, clavate, thin-walled cell. This trichome “cep tonal for the family and has surprisingly been overlooked in previous descriptions of the genus. : Leaf. Dorsiventral. composed of uniform] epidermis tend to be the adaxial and tricho Hairs of the type described above. Epidermis is y thickened, thin-walled cells. Cells of the abaxial is is the first report of the anisocytic stomatal + hd ; . . pattern in the Dilleniaceae. Guard cells have an average length of 26 1971] DICKISON, DILLENIACEAE, VII 323 Ficures 1-15. Floral morphology of Acrotrema. 1-8, Acrotrema sp. (from Hoogland 11469, Ceylon). 1, Lateral view of open flower, X 2; 2, open flower e seen obliquely from above, ; 3, flower seen obliquely from beneath, x 2; 4, gynoecium of 3 : 5, , X 3; 6, stamen, X 6; al, X 3; 8, anther, showing longitudinal dehiscence, -15, A. costatum (from Hoog- 6. land 11635). 9, Oblique view of flower from above, x 0, same, from be- neath, X 2; 11, lateral view of bicarpellate gynoecium, < 3.5; 12, front view of stamen, X 4; 13, same, from side, X 4; 14, sepal, X 3; 15, petal, X 3. and contain numerous chloroplasts. The mesophyll is differentiated into Palisade and spongy layers although they are not sharply defined (Fic- URE 31). A weakly differentiated hypodermis is evident in A. costatum. Raphides are common. Subepidermal layers of collenchyma occur around the midrib. I have drawn attention elsewhere (Dickison, 1969) to the marked differences in petiole vasculation between A. costatum and the 324 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Ceylonese species, where initially three bundles give rise to an abaxial from Ceylon have secondary veins which fork noticeably at the margins . oles Qa W 6 / {WE al Pata Wa id (\\ Wr ! \ Al ‘ 3 - yf ae ae de " _—-~ F * : % ' 4 = oS Fine” j (i 1 of / if ) \ " 8 eS a3 ‘ \ > e Deh re 7 ie sss 8 ony Lad pk (eX judas ala steel wl fotarcee Laeee S ; \ Ne, \ onl, dN Nn! —_ oo n> S44 o = Ss ose : Se ‘ eels \ WLD ond MF Lad tn FIGURES 16- : glandular t be Bes Veeetative morphology of Acrotrema, 16, Lateral view of same, seen from shove So of A. costatum (Hoogland 11635), X 62; 17 ri ——— 11423), 62° ‘ glandular trichome from sepal of Aqua . costatu , ) bundles Mee bhai Malaya, s.n.) showing widely separated vascular section ‘through Hee " iy A. costatum (from same coll.), X 1/2; 21, as tion of periderm (p) . — level showing departing leaf traces (It), imita- (Tn So ogre composed of collateral bundles, x 4; 22, lea ~ A asa: ge : 4), showing pubescence and pinnat (Hoogland’ Peppy bation trichome from the sepal Of Acrotrema Sp- sepal of A. costatum (Hoogling 11635) ye ae ee ae 1971] DICKISON, DILLENIACEAE, VII 325 Stem. The stem is circular in outline in transverse section. The aérial chymatous interfascicular areas (FIGURE 21). In this case, the stem possesses fifteen to twenty collateral bundles each with its own vascular cambium. In the same species, a periderm arises in the inner cortex (Ficure 21). Current observations on the secondary xylem agree with those already published (Dickison, 1967a), On the transverse section it is often difficult to distinguish vessel elements from fiber-tracheids owing to similar mean diameters and the radial alignment of xylem elements (Ficure 27). Average vessel element length is 700 » While the mean length of imperforate tracheary elements is 820 up. Vessel elements pos- sess scalariform perforation plates with an average of 24 fully-bordered bars. The secondary phloem is composed of sieve tube elements, com- panion cells and crystal-containing parenchyma. Sieve tube elements correspond to Type III of Zahur (1959) in possessing simple, transverse sieve plates (FIcuRE 28). Sieve tube slime is abundant. The nodal anatomy of all Ceylonese species examined is trilacunar (3:3) whereas A. costatum is distinguished by a multilacunar condition (ca. 9 traces and gaps). Root. The epidermal cells are thin-walled and distinguished from the cortex by dark-staining contents. The wide cortex is composed of uni- form parenchyma containing starch and raphides. The stele (FicurE 30) is sharply delimited by an endodermis and 1-layered pericycle. Primary xylem contains a variable number of protoxylem points ranging from triarch to pentarch. Patches of primary phloem are located between the protoxylem poles. FLORAL ANATOMY The vascular anatomy of flowers of Acrotrema (Ficures 25, 26) is identical with that described by Wilson (1965) for species of Hibbertia. The pedicel of A. costatum contains four or five distinct bundles whereas a siphonostele is typical in other species. The sepals are vascularized by three traces which either depart from separate gaps, or result from the: trifurcation of a single large bundle which diverges from a single gap. Lateral traces to adjacent sepals are commonly fused along part of their length. The petals are each supplied by a single bundle; which, however, 326 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 52 FIGURE 25, Acrotrema sp. transverse sections through th (i) showing details of vasculation. The peria (Hoogland 11469, Ceylon), flower. Drawings of € mature flower from pedicel (a) to stylar i nth parts have not been included. 1971] DICKISON, DILLENIACEAE, VII 327 branches soon after entering the petal base. Owing to the flattened re- ceptacle and frequent fusion among perianth traces, it is frequently dif- ficult to follow the course of floral traces as they progress distally in the flower. Carpels are uniformly vascularized by three bundles, one dorsal and two ventrals. In the Ceylon taxa the receptacle and stele assume a decidedly triangular appearance in transverse section after the departure of perianth and androecial vascular bundles. The dorsal carpel bundles arise from each of the three angles, and terminate at the style apex. The remaining stelar vascular tissue organizes itself into the ventral carpel bundles with the number dependent upon the number of carpels, two or three, in the gynoecium. The ovular supply is derived from the ventral bundles or the ventral bundles terminate within the ovules directly as in A. costatum. No evidence of vestigial stelar vascular tissue was observed. The greatest variation in floral anatomy is found in the vasculation of the androecium, which diverges from the stele in the form of varying numbers of stamen trunk bundles which subsequently divide to produce individual stamen traces (see Wilson, 1965). In A. costatum, where fif- teen stamens encircle the carpels, six to eight concentric stamen trunk bundles arise from the stele. A few of these bundles remain unbranched and directly vascularize individual stamens; more commonly, however, they divide into three or four stamen traces. In the Ceylonese species studied, the stamens are externally clustered into three fascicles which alternate with the carpels. The androecial vas- cular supply originates as three strong stamen trunk bundles from the angles of a somewhat triangular stele. Occasionally trunks are double but arise from a single gap. The degree of branching of the main stalk varies from a single bifurcation to five or more divisions depending upon the number of stamens in a fascicle. POLLEN The pollen of A. costatum (FicurE 32) has an average size of 22 p (polar) x 19 pw (equatorial). Pollen shape is oblate spheroidal and semiangular in polar view. Grains are tectate with the exine thicker at the furrows, Sculpturing is reticulate. — type is tricolpate with furrows that are distinctly granular and narro DISCUSSION The genus Acrotrema exhibits a range of variation in vegetative and floral anatomy of potential phylogenetic and taxonomic significance. Its origin from a woody ancestor seems certain as a result of reduction in cambial activity. However, the limited secondary xylem produced is un- . dorsal carpel bundle; ovb, Bites bundle; st, stamen trace; str, stamen trunk; v, ventral carpel bun J , . 2 3 i lm, fc 27% 2S a FicurE 26, Acrotrema transverse sections through sah costatum (Hoogland 11635), flower. Drawings | (i) showing details of usually primitive for bers of the subfamil form perforation pla mem- a plant of this stature and resembles ay y Dillenioideae. The vessel elements hav rly fifty tes which may be composed of up to nea 1971] DICKISON, DILLENIACEAE, VII 329 fully bordered bars. Vessel element side wall pitting is scalariform, Im- perforate tracheary elements are fiber-tracheids with conspicuous bor- dered pits. The decrease in length of xylem elements is presumably correlated with a concomitant decrease in length of the cambial fusiform initials. known to contain anomocytic, paracytic, and anisocytic stomata, as well as a specialized pattern called “gordoniaceous.” For the most part these patterns are restricted to particular genera. Unfortunately, sufficient in- formation has not been obtained to determine the nature of the primitive trema is also especially noteworthy. The only other dilleniaceous genus to have glandular hairs is Dillenia (Dickison, 1970a). The structure of the stem shows no outstanding differences from other Dilleniaceae. The phloem is of a comparatively advanced type with crystal-containing parenchyma present. The range of intrageneric varia- tion, however, is of importance. Acrotrema costatum is characterized by a eustele of dissected collateral bundles, multilacunar nodes, and petioles vascularized by a ring of widely spaced bundles, The Ceylon plants ex- amined had a complete vascular cylinder in the stem, trilacunar nodes, and a more complex petiolar anatomy. The course of evolution within the genus remains unclear at this time. No one species or group of species, contains totally unspecialized features. Wilson (1965) has pointed out that the nature of the stamen trunk androecial vascular supply in the large dilleniaceous genus Hibbertia is of possible taxonomic and phylogenetic significance. In his opinion, the following major trends of stamen trunk specialization may be recognized: Primitive than those with reduced branching. With the above considera- tions in mind, the following observations may be noted regarding Acro- trema. The androecial vasculation of A. costatum is significantly distinct from the Ceylonese species studied, as a result of the high number (6-8) of stamen trunks which depart the stele. The degree of branching of the trunks, however, is reduced to usually three and a few of the stalks do 330 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 52 Thus, the evidence from both vegetative and floral organs indicates that while Acrotrema is often very divergent in growth habit and gross morphology, the level of anatomical specialization has surprisingly re- mained comparable to other members of the family. The genus appears only remotely related to extant genera of the Dilleniaceae although anatomically it is most similar to Dillenia and Hibbertia. LITERATURE CITED Dickson, W. C. 1967. Comparative oo studies in Dilleniaceae, I. Wood anatomy. Jour. Arnold Arb. 48: 1-29. . 1968. Comparative gee Ja in Dilleniaceae, III. The carpels. Jbid. 49: 317-332. . 1969. Comparative morphological studies in Dilleniaceae, IV. Anatomy of the node and vascularization of the leaf. Jbid. 50: 384-410 ; Comparative = rt studies in Dilleniaceae, V. Leaf anatomy. Jbid. 51: 89-1 HALuier, H. 1912. L’origine a le systeme phylétique des ee exposés 4 Vaide de leur arbre généalogique. Arch. Néerl. Sci. Nat. sér. B. 1: 146-234. Hooker, J. D., & T. Tomson. 1872. Dilleniaceae. In :; D. hed: The Flora of British India. Vol. 1. L. Reeve & Co., Ken Meeuse, A. D. J. 1966. Fundamentals of “at ones The Ronald Press, New York, Metcatre, C. R., & C. CHALK. 1950. Anatomy of the Dicotyledons. 2 Volsfl The ae Press, Oxford. Swamy, B. G. L., & K. Pertasamy. 1955. stl dagi 2 the embryology of Acrotrema arnottianum. Phytomorphology 5: 30 TakHtTayan, A. 1954. Essays on the og aes “hacibetes of Plants. (Transl. O. H. Gankin.) Am. Inst. Biol. Sci., Washington 1969. Flowering Plants, Origin and Dispersal. (Transl. C. Jeffrey.) Smithsonian Inst, im Washington. Tuwartes, G. H. K. 186 seats plantarum Zeylaniae: an enumeration of cau plants, wi descriptions of the new and little known genera and nae a ae on their habitats, uses, native names, etc. Dulau 0., Lon x 1893. A Hand-Book to the Flora of Ceylon. Pt. 1. Dulau & Co., Witéen, C. re 1965. The & TRIMEN, Lon floral gt of the Dilleniaceae. I. Hibbertia Andr. 274. mparative study of secondary phloem of 423 species of ge cotyledon belonging to 85 families. Cornell Agr. Exp. Sta. ree DEPARTMENT OF Botany THE UNrversity OF NortH CAROoLIN CHAPEL Hit, Norru CAROLINA ay 1971] DICKISON, DILLENIACEAE, VII 331 EXPLANATION OF PLATES PLATE I Ficures 27-34. Morphology of Acrotrema. 27, Transverse section through the rhizome of Hoogland 11423 showing cambial zone (c), secondary xylem an phloem, X 156; 28, same, longitudinal section of phloem showing sieve tube i 781 29, transverse section through the stem apex of Hoogland 11575 showing condupli- cate vernation of youngest leaf, X 156; 30, transverse section through the stele of i toxylem (px) and metaxylem (mx), X 312: 31, transverse section of the leaf PLATE II Ficures 35-43. Herbarium specimens of Acrotrema illustrating variation in gross morphology. 35, A. costatum (Smitinand 2999, US 2211867); 36, A, uni- s 94, US 1576876); 39, A. gardneri (Thwaites C.P. 253, US 597060); 40, A. lanceolatum (Thwaites C.P. 2660, US 1576879; 41, A. thwaitesii (Thwaites C.P. 3969, 597067); 42, A. thwaitesii (Thwaites C.P. 3364, US 1576882); 43, A. dissectum (Thwaites C.P. 3393, US 1576883). Jour. ARNOLD Ars. VoL. 52 PLatTE I Dickison, DILLENIACEAE VII ouR. ARN f fou. 5 Jo RNOLD ARB. VOL, 52 PLaTeE II DICKISON, DILLENIACEAE VII 334 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 A REVISION OF THE BORAGINACEAE OF WEST PAKISTAN AND KASHMIR * S. M.A. Kazi 24. Lindelofia Lehm. Hamburg. Gart.-Blumenzeit. 6: 351. 1850. Type species: L. longiflora (Benth.) Baill. Perennial pubescent herb. Basal leaves usually petiolate, cauline leaves alternate, ovate to linear-lanceolate. Inflorescence ebracteate. Calyx usual- ly divided to the base, slightly enlarged and spreading in fruit. Corolla blue, purple or dark purple, cylindrical or cylindrical-campanulate to in- fundibuliform, lobes usually spreading; faucal appendages usually large, rarely small, trapeziform to oblong, developed at the level of the fila- ments, or above, on the corolla throat. Stamens 5, anthers large, linear to oblong, usually three times as long as broad, often sagittate; filaments short. Ovary 4-lobed; style long, slender, stigma small. Nutlets de- pressed-ovoid to depressed-globose, dorsally deeply concave to plane, ap- pendiculate at the margins, appendages subulate and glochidiate at apex, dilated and confluent at base, ventrally attached to a short gynobase. Species about 12, distributed in northern Iran, Afghanistan, West Pak- istan, Kashmir, northwest India, and adjoining areas of URSS and China. KEY TO THE SPECIES a. Corolla tube much shorter than the limb, limb divided for 1/3 to 1/2 its ength; filaments shorter than the anthers, inserted a little to much below the bases of the faucal appendages 1. L. anchusoides. a. Corolla tube equalling to much longer than the limb, limb divided nearly to the base; filament d S + equalling the anthers, inserted adjacent to the bases faucal appendages very short, much shorter than a 2. L. stylos b. Corolla lobes ovate, faucal appendages large, equalling or longer than the ee 3. L. longifiora. 1. L. anchusoides (Lindl.) Lehm, Hamburg. Gart.-Blumenzeit. 6: 352. 1850; Linnaea 24: 216, 1851; Riedl in Rechinger, Fl. Tranica 48: 138. 1967 Cynoglossum anchusoides Lindl . DC - Bot. Reg. 28: - A. DC. in DC. Prodr. 10: 151. 1846, ot. Reg. 28: ¢. 14. 1842; A C. emodi Schouw. Ind. Sem. Hafn. Coll. 4. 1846, * Continued from volume 52, page 136. 1971] KAZMI, BORAGINACEAE 335 C. macrostylum Bunge, Lehm. Relig. Bot. 412. 1847; Mém. Acad. St. Pé- tersb. Sav. Etr. 7: 412. 1851; Boiss. Fl. Orient, 4: 266. 1875. Paracaryum heliocarpum Kerner, Ber. Naturw.-Med. Innsbruck 1: [suppl.] 105. 1870; C. B. Clarke in Hook. f. Fl. Brit. India 4: 161. 1883; Boiss. FI. Orient. suppl. 354. 1883; Hook. in Curtis’s Bot. Mag. 113: t. 7520. 1897. Adelocaryum anchusoides (Lindl.) Brand, Repert. Sp. Nov. 13: 548. 1915: Pflanzenr. IV. 252(Heft 78): 77. 1921. Lindelofia cynoglossoides Brand, Pflanzenr. IV. 252(Heft 78): 88. 1921. Lindelofia aspera Rech. f. Ann. Naturh. Mus. Wien 58: 48. 1951. L. macrostyla (Bunge) M. Pop. Fl. URSS. 19: 627. 1953. TYPE: Cultivated from the seeds obtained from Kashmir and Tibet, Herb. East India Company without citation of collector’s name, 5.”. (K). Icon.: Lindl. 1. c. ¢. 14. 1842, under Cynoglossum anchusoides; M. Pop. 1. c. t. 35. fig. 1. 1953, under Lindelofia macrostyla; Hook. in Curtis’s Bot. Mag. 123: t¢. 7520. 1897, under Paracaryum heliocarpum. Perennial, erect or rarely decumbent herb, to 100 cm. tall. Stems fis- tulose, striate, solitary or few with ligneous bases, simple or branched above, covered sparsely or densely with long, soft, usually retrorsely sub- appressed to appressed trichomes, nontuberculate at base, or sometimes with spreading trichomes arising from tuberculate bases. Basal leaves petiolate, petioles to 18 cm. long, lamina lanceolate, entire, acute, gradu- ally narrowed towards the base, to 25 cm. long, 7 cm. broad, upper sur- face covered sparsely with soft, long, + retrorsely subappressed trichomes, lower surface sparsely to densely pubescent in the young leaves, rarely the old leaves covered with stiff trichomes arising from tuberculate bases; cauline leaves short petiolate, lanceolate, to 13 cm. long, 2 cm. broad; up- per cauline leaves sessile, lanceolate to linear-lanceolate, to 8 cm. long, 1 cm. broad. Inflorescence terminal and axillary, long pedunculate, gem- inate or ternate, few flowered. Pedicel erect, to 4(-5) mm. long, in flower, horizontally recurved or slightly reflexed, densely pubescent and 10(-20) mm. long in fruit. Calyx divided nearly to the base, lobes ob- long to oblong-ovate, obtuse, densely and uniformly pubescent, 2.5—5 mm. long, to 1 mm. broad in flower, enlarged to 6(—7) mm. long, 1.5(—2) mm. broad in fruit. Corolla pink, blue or purple, campanulate, to 10 mm. long, tube equalling or longer than the calyx, limbs about 1.5 times longer than the tube, divided 1/3 to 1/2 its length, lobes suborbicular; faucal appendages usually puberulous with a prominent medial line, to 2 mm. long, 1 mm. broad at the base, slightly tapering towards the tri- lobed apex, margins usually ciliate. Anthers to 2 mm. long, 0.5 mm. broad, Sagittate; filaments fleshy, shorter than the anthers, inserted at the mid- dle or below the bases of the faucal appendages on the corolla tube. Style + 9 mm. long, filiform; stigma capitate. Nutlets orbicular to ovate, 4-5(-6) mm. long, 3-4 mm. broad, dorsal surface, usually appendiculate along the medial line and the margins, dorsal appendages generally shorter than the marginal ones, marginal appendages puberulous, triangular, to 2 mm. long, confluent at base and forming a broad marginal wing, usual- 336 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 ly more or less inflexed on the dorsal surface; appendages on the lateral and latero-ventral surfaces slightly shorter than the marginal ones. DistripuTIon: Afghanistan, West Pakistan, Kashmir, northwest India, Pamir Alaj, Tien Shan. West Pakistan: CHITRAL STATE: Brumboret, 2000 m., on Hindukush south, Bowes Lyon 644 (pm); Ziarat, Lowarai Pass, 2100 m., Stainton 2552 (BM, E). Gitcir AGENCY: Rama Valley, s.-w. of Astor, 3000 m., Lankester & Pearson 1453 (pM); Naltar lake, 3500-4000 m., 21 July, 1954, R. R. Stewart s.n. (micH); Gurez Valley, Badwan, 2500 m., R. R. Stewart 19617 (cH); Chillam Chowki, 3100 m., Siddiqui, Nasir & Zaffar 4229 (pm); Chhacher pass, 3500 m., Webster & Nasir 6392 (GH); Astor valley, 2300-2700 m., 23 July, 1892, Duthie s.n. (pm); Gilgit, Toppin 1005 (K); Gubbar, south of Hindukush, 3000 m., Giles 745 (K); South of Hindukush, Giles s.n. (E); South of Astor, 2500 m., Giles sn. (K); Kunslwan (south of Gurez), 2500 m., 21 July, 1876, C. B. Clarke 29366a, 39388 (K), 29366b (BM); Kero Lugma Glacier, left bank, 3500 m., Russell 1319 (BM, GH). Hazara Dist.: Kaghan Valley: Burawai-Basal road, Siddiqui 24161 (Bm); Naran, 2700 m., Kazmi 573b (pes), Jafri & Ali 3416 (E); between Naran and Saiful Maluk, ca. 2800 m., Kazmi 802b (PES). Kurram AGENcy: Kurram Valley, Hariab, Aitchison 576 (GH); between Kurram and Kaiwas, Aitchison 579 (Bm). Swat State: Batani, above Ushu, 1971] KAZMI, BORAGINACEAE 337 times is correlated with smaller flowers and shorter pedicels and some- times not. On examination of more material the plants with stiff, tuber- culate trichomes and short pedicels may prove to be a variation worth recognition. From the other two species of our area, Lindelofia longiflora and L. stylosa, L. anchusoides differs in having a much longer corolla limb com- pared to the length of the tube, and filaments shorter than the anthers, which are obviously inserted below the bases of faucal appendages. 2. L. stylosa (Kar. & Kir.) Brand, Pflanzenr. IV. 252(Heft 78): 85. 1921; M. Pop. in Fl. URSS 19: 633. 1953; Riedl in Rechinger, FI. Tranica 48: 141. 1967. Cynoglossum stylosum Kar. & Kir. Bull. Soc. Nat. Moscou 15: 409. 1842. Solenanthus nigricans Schrenk ex Fisch. & C. A. Mey. Enum. PI. Nov. 2: 29. 1842; DC. Prodr. 10: 165. 1846; Ledeb. Fl. Rossica 3: 171. 1847. Lindelofia benthamii Hook. f. Fl. Brit. India 4: 159. 1883. Type: In pratensibus subalpinus Alatau ad fi. Sarchan, Karelin & Kirilow s.n. (LE-holotype, us-isotype). Icon.: M. Pop. 1. c. ¢. 36. fig. 1. 1953. Perennial herb, to 60 cm. tall. Stems single to many, striate, simple or branched above, sparsely covered with usually spreading, or sometimes in young plants, retrorsely subappressed trichomes, hardly tuberculate at their bases; old stems sometimes quite glabrous below. Basal leaves lan- ceolate, entire, acute, narrowed towards the bases into long petioles, in- cluding petioles to 35 cm. long, 6 cm. broad; covered on both surfaces with white, subappressed trichomes, to 0.5 mm. long, arising from tubercu- late bases; lower cauline leaves subsessile to short petiolate, to 20 cm. long, 1.5 cm. broad; upper cauline leaves few, sessile, linear. Inflorescence terminal, short in flower, to 15 cm. long and paniculately branched in fruit. Pedicels to 7 mm. long in flower, to 22 mm. long and arched out in fruit. Calyx divided to the base, lobes linear-lanceolate, acute, dense- ly covered with suba ressed, white trichomes, 6-8 mm. long, 1 mm. broad, slightly enlarged in fruit. Corolla dark purple, tubular to tubular- campanulate, 11-13 mm. long, tube ca. twice as long as the limb, limb divided to 3/5 its length, lobes narrow-oblong to oblong, obtuse to round- ish, suberect, 3-5 mm. long, ca. 1.5 mm. broad; faucal appendages, to 1 mm. long and broad, semilunar. Anthers exserted from the corolla tube, 3-3.5 mm. long, 0.5 mm. broad, apiculate, sagittate; filaments fleshy, 1.5 mm. long, inserted on the corolla tube adjacent to the bases of faucal appendages. Style filiform, 13-15 mm. long; stigma capitate. Nutlets 4, broad ovate, to 7 mm. long, 6 mm. broad, dorsal surface glabrous or tuberculate, short appendiculate at the margins and all around, appen- dages not exceeding 0.7 mm. in length, apices glochidiate, sometimes mar- ginal appendages confluent at their bases forming a very narrow marginal wing, 338 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 52 DIsTRIBUTION: Pamir Alaj, Tien Shan, Dsungaro-Tarbagatai, Tibet, Alatau, Afghanistan, West Pakistan, Kashmir, northwest India. riff 8368 (GH); Zanskar: Kangi La to Rangdum, 3500 m., Koelz 2874 (cn); Sulle to Padum, Schlagintweit 6708 (cH); top of the Pentse La Pass (pass be- tween Zanskar and Dras), Schlagintweit 7476 (GH). Brand [Pflanzenr. IV. 252(Heft 78) : 87. 1921] considering Lindelofia benthamii Hook. f. (1. c.) identical to L. angustifolia (Schrenk) Brand, 1. c. IV. 252(Heft 78): 87. 1921, cited a number of specimens from our area. In fact L. benthamii is a synonym of L. stylosa and agrees in all its details with the latter. L. stylosa, is distinguished from L. angustifolia by having distinct faucal appendages in the corolla throat, whereas such ap- pendages are totally absent in L. angustifolia. I did not see any specimen of L. angustifolia from West Pakistan, Kashmir or northwest India and its occurrence in these areas appears to be quite doubtful. The specimen (Schlagintweit 6708) cited by Brand belongs to L. stylosa. Lindelofia stylosa is closely related to L. longiflora from which it is distinguished by having narrower corolla lobes and very short faucal appendages. 3. L. longiflora (Benth.) Baill. Hist. Pl. 10: 379. 1890; Giirke in Engl. & Prantl, Nat. Pflanzenfam. IV. 3a: 103. #. 42, fig. AC. 1893; Brand, Pflanzenr. IV. 252(Heft 78): 85. 1921. Lindelofia spectabilis Lehm. in Hamb. Gart.-Blumenzeit. 6: 352. 1850; Lin- naea 24: 216. 1851; C. B. Clarke in Hook. f. Fl. Brit. India 4: 159. 1883. Type: Cashmere, Royle sn, (x?), Icon.: Lindl. Bot. Reg. 26: ¢. 50. 1840. Perennial herb, to 60 cm, tall. Stems usually solitary, sometimes few, simple, covered sparsely to densely with white, crisped, spreading to sub- appressed trichomes, to 1.5 mm. long, some from tuberculate bases others not. Basal leaves lanceolate, entire, acute, attenuate towards the base into petioles, usually shorter, sometimes equal or rarely longer than the lamina, including petioles to 18 cm. long, 8 cm. broad, sparsely to densely covered 1971] KAZMI, BORAGINACEAE 339 sile, oblong-lanceolate to ovate, acute, at the bases slightly dilated, more or less rounded or subcordate; upper cauline leaves semiamplexicaul, shorter and narrower than the lower ones. Inflorescence terminal or subterminal, short in flower, later elongated to 15 cm. Pedicel erect, to 5 mm. long in flower, pubescent, elongated up to 12 mm., horizontal to sub- reflexed in fruit. Calyx divided to the base or connate for its lower 1/4, 7 mm. long in flower, lobes oblong to ovate, obtuse to subacute, ca. 3.5 mm. broad, variously hairy, to 10 mm. long, lobes 5.5 mm. broad in fruit. Corolla bright to deep blue to purple, campanulate-infundibuli- form, tube to 13 mm. long, limbs to 6 mm. long, divided nearly to the base, lobes ovate, spreading; faucal appendages to 4 mm. long, 2 mm. broad, sometimes slightly tapering toward the more or less emarginate apex, margins ciliate. Anthers ca. 2 mm. long, 0.5 mm. broad, more or less rounded at both ends; filament ca. 2 mm. long, fleshy, inserted ad- jacent to the bases of the faucal appendages on the corolla tube. Nutlets ovate, excluding appendages to 4 mm. long, 2.5 mm. broad, margins ap- pendiculate, appendages to 2 mm. long, glochidiate at apex, dilated at base, free or confluent forming a narrow marginal wing, spreading, rarely inflexed on the concave dorsal surface, which has a more or less pro- minent medial ridge, bearing few to many vertical appendages similar to the marginal ones. Style to 15 mm. long, filiform; stigma capitate. DISTRIBUTION OF SPECIES: West Pakistan, Kashmir, northwest India. KEY TO THE VARIETIES e Upper cauline leaves broad-lanceolate to ovate; calyx lobes with prominent midrib, bearing long crisped trichomes usually only on the midrib and stele ihn Goh 6 ae be S ire a anata Si ide te gear 3b. var. levingit. . Upper cauline leaves lanceolate; calyx lobes with inconspicuous midrib, bearing short trichomes usually all over the surface and margins. b. Calyx lobes subacute to acute at apex, lobes uniformly pubescent on the Septal ETO ng. hos oy ee once cs PR eee ee Ee 3c. var. falconeri. b. Calyx lobes obtuse to roundish at apex, densely pubescent on the mar- gins with short trichomes, and uniformly pubescent on the dorsal POE oS bin is Sas Se rt Pont ater Larne POG TCA 3a. var. longiflora. & 3a. Var. longiflora. Leaves on the upper surface sparsely covered with subpatent, long trichomes, arising from tuberculate bases, densely covered on the lower surface with subappressed, crisped, soft trichomes, not tuberculate at their bases, to 0.3 mm. long. Calyx usually divided for 3/4 of its length or sometimes to the base, lobes to 7 mm. long, 3.5 mm. broad in flower, enlarged to 10 mm. long, 5.5 mm. broad in fruit, with inconspicuous mid- rib, covered densely over the dorsal surface with long, white trichomes,. usually longer on the middle and towards the base, margins very densely pubescent with trichomes to 0.2 mm. long, ventral surface more or less: pubescent, usually on the upper parts and margins. Corolla dark blue, in- 340 JOURNAL OF THE ARNOLD ARBORETUM [von. 52 fundibuliform, tube to 13 mm. long, limb to 7 mm. long; faucal appen- dages comparatively narrower and shorter than in var. levingii. West Pakistan: CHITRAL STATE: Brumboret, M. A. Siddiqui & A. Rehman 26853 (gM). GiLcir AceNncy: Kamri Pass, 3000-3500 m., Duthie 12581 (BM); Baltistan, 2700 m., Hunter-Weston 10251 (micH). Hazara Dist.: Kaghan Valley, Saiful Maluk, Shaukat Ali 129 (sm). ir: Tragbal, 19 July 1876, C. B. Clarke 29226 (x), 3200 m., Sept. 1913, Evershed s.n. (BM), Duthie 13524 (k); Rajdhiangan Pass, R. R. & I. D. Stewart 18005 (GH), 3000-3500 m., R. R. Stewart 19517 (cH), 3000 m., R. R. Stewart 19307 (GH); Sonamarg, 3300 m., R. R. Stewart 6560 (K), 9801A (cH); Sind Valley, 3000 m., Stainton 5026 (BM); vicinity of Sonamarg, on the Sind river, 50 road-miles east-northeast of Srinagar, 4000 m., Dickason 91 (micH); Gangangir, Sind Valley, 3200 m., Ludlow & Sherriff 7971 (pm, cH), 7979 (cH); Gumber valley in Sind Valley, 3300 m., Ludlow & Sherriff 7636 (GH); Sekiwas, Yamkar Pass, 3500 m., R. R. Stewart 12552 (GH); Mengandob, 3300 m., Polunin 56/724 (pm); Butin Pandsab, Wangat Valley, 3300 m., Pinfold 269 (sm); Dras Valley, Zoji La, 2 June 1870, Henderson 324 (kK); Gangabal Lakes, 3000 m., R. R. & I. D. Stewart 4552 (micu); Zaiwan, 3000 m., Polunin 56/603 (BM); near Karagbal, 2700 m., R. R. Stewart 22559 (cH); Allyabad, Pir Pan- jal, /nayat 25720 (xk); Surgun, 3200-4000 m., Schmid 549 (pm); Budnambal, 000 m., Fuller 226 (xk); Mantar Valley, near Desu, 3000 m., Ludlow 95 (BM). Miscellaneous: Kashmir, Price 72A (x), Coventry 577 (kK). The typical variety is very variable in the form of its leaves. Gen- erally the basal leaves have long petioles; the cauline leaves are oblong- lanceolate, acute, usually narrowed at both ends and truncate, roundish or subamplexicaul at base. The lower surfaces are densely covered with short trichomes to sometimes slightly pubescent. All the trichomes lack tuberculate bases. These characters are well correlated with the ovate, obtuse calyx lobes densely ciliate at the margins. The corolla tube is narrower than that of var. Jevingii. Corollas in almost all the collections are deep blue rather than purple. i a gS a (C. B. Clarke) Brand, Pflanzenr. IV. 252(Heft 78): fee Berg iene on var. levingit C. B. Clarke in Hook. f. Fl. Brit. India 4: . 3 L. platycalyx Riedl in Kgie & Rechinger, Biol. Skr. 13(4): 202. 1963, in Rechinger, Fl. Iranica 48: 140. 1967, syn. nov. Type: Kashmir: Pir Pingul, alt. 11500 ft., Levinge s.n. (x). Icon.: Riedl, 1. c. fig. 153. 1963, under Lindelofia platycalyx. Leaves sparsely covered on both surfaces with trichomes arising from tuberculate bases; calyx shape more or less like that of the typical varie- ty; calyx lobes with prominent midrib, covered sparsely or densely, usually on the midrib and margins, with crisped trichomes, to 0.5 mm. long, internally usually glabrous. Corolla mostly purple to sometimes dark 1971] KAZMI, BORAGINACEAE 341 blue, campanulate-infundibuliform, tube to 10 mm. long, limbs 5-6 mm. long, divided nearly to the base; faucal appendages to 4 mm. long and 2 mm. broad West Pakistan: Gitcir AcENcy: Upper Astor Valley, Kalapani to Shan- kargarh, 3300 m., R. R. Stewart 22745a (cH). Hazara Dist.: Thadal (Shadal), 16 Aug. 1899, Inayat s.n. (w-type of L. ergot Riedl); Kaghan Valley, be- tween Balakot and Babusar Pass, Abel 134 (pm); between Shogran and Sari Rest House, 2500 m., Kazmi 2158 (PEs); pire: of Saiful Maluk Sar, 3570 m., Schmid 331 (BM, "g); Saiful Maluk, Shaukat Ali 129 (pm, mixed with var. longiflora : Kashmir: Pir Panjal, 3300 m., Levinge, s.n. (K-type); Pir Panjal Pass, 3200 m., Ludlow 31 (Bm); Banihal Pass, Pir Panjal Range, 2575 m., R. R. Stewart 12203, 14709 (GH); Gulmarg, 2750 m., Duthie 13045 (Bm), 3000 m., R. R. Stewart 10395 (GH), 2700 m., Tamin 158 (pm); Gulmarg, Nagini, 2700 m., Stewart 14843a (GH); Killanmarg, above Gulmarg, 3000 m., R. R. Stewart 10420A, 8665 (GH); Pahlgam, R. R. & J. D. Stewart 5904 ‘(wicH): Vicinity of Kolahoi mountains and glacier, about 50 miles north of Islamabad, via Pahlgam and Arau, Dickason 94 (micw); Tuleon, above Pahlgam, 3200 m., R. R. Stewart 21844a (cH); Tuleon Way, east of Pahlgam, on east Lidder river, 27 road-miles north of Islamabad, 2700 m., Dickason 93 (micH); Shisha than Pass, 3500 m., Ludlow & Sheriff 9291 (Bm, cH); Ningle Nullah, Pinfold 217 (BM). Miscellaneous: Kashmir: R. R. Stewart 12957 (BM), 2700 m., 1957, Mrs. Prescott-Decie s.n. (BM) The variety Jevingii of Lindelofia longiflora is very variable in the size and form of leaves, but generally the cauline leaves are much broader and shorter compared to those of the other two varieties. The leaves are always sparsely pubescent on both surfaces. These characters are cor- related with the pubescence of the calyx lobes, which are covered only on the midrib and the margins with long, crisped white trichomes. The corolla in this variety is comparatively broader and more or less cam- panulate in form, sana) eid faucal appendages compared to the corolla in the other two varietie Lindelofia platycalyx oe in n its details to L. longiflora. 3c. Var. falconeri (C. B. Clarke) Brand, Pflanzenr. IV. 252(Heft 78): 85. 1921. Lindelo fia ati tighend Lehm. var. falconeri C. B. Clarke in Hook. f. Fl. Brit. ndia 4: Type: Kasumir: Falconer s.n. (K-Herb. Late East India Co., holotype, GH-isotype). Leaves very similar to those of var. longiflora; calyx obviously divided to the base, lobes oblong-lanceolate, acute to subacute, to 6 mm - long, 1.5 mm. broad in flower, enlarged to 10 mm. long and 3 mm. broad - fruit; corolla as that of var. longiflora. West Pakistan: Gitcrr AGENCY: Burzil Pass, north slope, 3300 m., R. R. 342 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Stewart 22004 (cH), south slope, R. R. Stewart 19126 (GH); Kamri Pass, 4000 m., Giles 700 (K); Mir Panzil Pass, ascent to the Deosai, 3300 m., R. R. Stewart 19985 (GH). Miscellaneous: Gilgit-Chitral, Leichtlin s.n. (x, received in herb. on July 1889). Kashmir: Koragbal, north slope, Rajdhiangan Pass, ca. 2750 m., : Stewart 22586c (GH); Sonamarg, 3000 m., R. R. Stewart 9858A (GH); vicinity of Sonamarg, on Sind river, 50 road-miles east-northeast of Srinagar, 3500 m., Dickason 92 (MicH); above Gulmarg, 3000 m., R. R. Stewart 10329 (cH); Koslorkut, 8 miles n.w. of Vishensar, 3300 m., Polunin 56/701 (BM). Mis- cellaneous: Kashmir, Falconer s.n. (K, GH, types). Lindelofia longiflora var. falconeri can very easily be distinguished from the other two varieties of the species by its lanceolate, acute to subacute calyx lobes. In the type specimen the upper cauline leaves are narrower than in any of the other specimens cited. In the other collections the cauline leaves are usually oblong-lanceolate and broader than those in the type specimen. 25. Cynoglossum L. Sp. Pl. 134. 1753; Gen. Pl. ed. 5. 65. 1754. TYPE SPECIES: C. officinale L. (Lectotype). _ Species about 60, distributed all over the world, except in arctic re- ions. KEY TO THE SPECIES and forming narrow to broad RM ah dh swans icra si short, = 4-angled; nutlets roundish, dorsally Cony Y appendiculate, appendages + lencth. usually 10 confluent at their bases, wae Of equal length, c. Leaves uniformly covered dorsally with trichomes, to 0.2 mm. long, 2. C. zeylanicum. ao 1971] KAZMI, BORAGINACEAE 343 c. Leaves covered dorsally with trichomes of unequal lengths, usually longer than 0.2 mm., obviously arising from tuberculate bases. ...... By ee eae Tey ey ee MO Mey eI aD IF 1S ON 1. C. lanceolatum. a. Nutlets 3.5-5 mm. long. . Basal leaves petiolate, lamina ovate, abruptly narrowed at the base, truncate or rarely subcordate. ...................... . C. stewartii, Basal leaves petiolate, lamina oblong, gradually attenuate at the base towards the petioles. e. Cauline leaves oblong, obtuse to subacute, gradually narrowed to- bases Rs 1, C. lanceolatum Forssk. Fl. Aegypt.-Arab. 41. 1775; DC. Prodr. 10: 149. 1846; C. B. Clarke in Hook. f. Fl. Brit. India 4: 156. 1883; Brand, Pflanzenr. IV. 252(Heft 78): 137. 1921, excl. vars.; Riedl in Rechinger, Fl. Iranica 48: 145. 1967. micranthum Desf. Tab. Ecole Bot. ed. 1. 220. 1804. hirsutum Jacq. Hort. Schénbr. 4: tab. 489. 1804. canescens Willd. Enum. 180. 1809. racemosum Roxb. Hort. Beng. 13. 1814; Fl. Indica, ed. Carey, 2: 6. 1824. platyphyllum Klotzsch in Peter, Reise Mozambique Bot. 254. 1862. abyssinicum Hochst. & Vatke ex Engler, Abh. Preuss. Acad. Wiss. 1891. 2: AAANAN Ww on - 1892. C. johnstoni Baker, Kew Bull. 1894: 29. 1894. Type: Hadie, Forsskdl s.n. (c). Icon.: Brand, |. c. fig. 18, A-G. 1921. Erect biennial or perennial herb, 60-150 cm. tall. Stems woody at base, branched, branches long, spreading; densely covered with stout, white trichomes arising from tuberculate bases up to 2 mm. long, tri- chomes on the upper part of the stem comparatively shorter and + sub- appressed. Basal leaves petiolate (petioles to 8 cm. long), lanceolate, en- tire, acute, attenuate at the base, including the petioles to 30 cm. long and 4 cm. broad, nerves on the upper surface inconspicuous, raised and prominent below, upper surface covered uniformly with thin, subap- pressed trichomes, intermixed in different densities with longer, thicker trichomes to 1 mm. long, arising from large tuberculate bases, sometimes surrounded with tiny white tubercles, lower surface covered more dense- ly, especially on the nerves, with similar trichomes; cauline leaves sub- sessile to short petiolate, to 12 cm. long and 2 cm. broad; upper cauline leaves sessile, much reduced in size. Inflorescence axillary or terminal, bifurcate to dichotomously branched, branches spreading, ebrac- teate, short, scorpioid in flower, elongated to 25 cm. when in fruit. Pedi- cels to 0.1 mm. long in flower, to 2 mm. long and usually reflexed in fruit. Calyx divided to the base, lobes ovate, obtuse, densely hairy externally, glabrous within, + 1 mm. long, enlarged and spreading in fruit, to 2 mm. and 1 mm. broad. Corolla light blue, 2-3 mm. long, limbs divided to the 344 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 base, lobes ovate; faucal appendages broader than long. Stamens included, anthers ovate, minute, filaments short. Nutlets 4, ovate, 1-2 mm. long, ++ equally broad, dorsally convex, appendiculate all around, appendages of equal length, vertical, not confluent at base, apex glochidiate. Style short, fleshy, 0.5-0.7(-1) mm. long in fruit. DistripuTion: Africa excluding the northwestern parts, Arabia, Pak- istan, Kashmir, India, Nepal, Burma, Thailand, Philippines, Formosa, Riu Kiu, and China. West Pakistan: Dir State: Dir village, 1500 m., Salim 105 (x). HaAzaRA Dist.: Tribal area, Nilishang, ca. 1200 m., Burtt 1426 (ce); Kaghan valley, be- tween Balakot and Babusar Pass, Abel 13 (Bm); Siran Valley, Jabori, Burtt mir: Bandipur, 1500 m., R. R. Stewart 22521 (cH); Ganderbal, R. R. Stewart 7497 (kK); Gund, 2000 m., Rich 278 (x); Chokoti to Uri, Aug. 1880, Young s.n. (sm). Miscellaneous: Himal. bor. occ. 3—7000 ped., Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (GH, MICH), 2. C. zeylanicum (Vahl) Thunb. ex Lehm. Neue Schriften Naturf. Ges. Halle 3(2): 20, 1817; Asperif. 116. 1818; Brand, Pflanzenr. IV. 252(Heft 78): 134, 1921, excl. var. lanceolatum C. B. Clarke; Riedl in Rechinger, Fl. Iranica 48: 145. 1967. Anchusa zeylanica Vahl ex Hornem. Enum. Hafn. 3. 1807, Hort. Hafn. 1: 176. 1813, nomen nudum. Echinospermum zeylanicum Lehm. Asperif. 116, 1818: A. DC. in DC. Prodr. 10: 142. 1846. epnoelocmen coeruleum Buch.-Ham. ex D. Don, Prodr. Fl. Nepal. 100. 1825. - micranthum Desf. var. decurrens Trimen, Handb. Fl. Ceylon 3: 203. 1895. TYPE: not indicated, pois ++ Wight, Icon. Pl. Indiae Orient. 4: £. 1395. 1850, under C. furca- Biennial erect herb, to 85 branched, covered densel 1971] KAZMI, BORAGINACEAE 345 trichomes to 0.2 mm. long, not arising from conspicuous tuberculate bases; cauline leaves short petiolate, upper cauline leaves sessile, oblong- lanceolate, gradually reduced in size upwards. Inflorescence terminal or subterminal, short in flower, ebracteate and elongated in fruit. Pedicel shorter than the calyx in fruit. Calyx divided for 3/4 of its length, lobes ovate, 3-4 mm. long, obtuse to roundish at apex, covered with trichomes like those on the leaves. Corolla blue, infundibuliform, 5 mm. long, limbs 8-9 mm. in diameter, divided nearly to the base; faucal appendages blue, subquadrate, larger than the anthers. Style fleshy, much shorter than the calyx. Nutlets 4, ovoid, 3-4 mm. long, uniformly appendiculate all around, emarginate, appendages glochidiate at apex, free at base, dor- sal surface convex. DistrisuTIon: Afghanistan, Pakistan, India, Ceylon, Philippines, Japan, ina t Pakistan: Reported from Swat: inter Madyan et Kalam, 1700 m., Rechinger 19377 (w). I did not see a specimen of this species from Afghanistan or West Pakistan; its distribution in these areas needs thorough investigation. 3. C. wallichii G. Don, Gen. Syst. 4: 354. 1838; A. DC. in DC. Prodr. 10: 150. 1846; C. B. Clarke in Hook. f. Fl, Brit. India 4: 157. 1883. stellatum Wall. Cat. No. 924. 1828, nomen nudum. canescens Wall. Cat. No. 918. 1828, nomen nudum. calycinum Wall. Cat. No. 923. 1828, nomen nudum, non C. A. Meyer. vesiculosum Wall. Cat. No. 920. 1828, nomen nudum. edgeworthii A. DC. in DC. Prodr. 10: 150. 1846. denticulatum A, DC. in DC. Prodr. 10: 150. 1846. micranthum Hook. f. & Thoms. ex C. B. Clarke in Hook. f. Fl. Brit. India 4: 157. 1885. Type: Native of Nepaul, in Gosainsthan, Wallich 923 (x). AANAAAN Erect biennial herb, to 75(-100) cm. tall. Stems solitary to many, Striate, usually branched above, branches long, divaricate; stem and branches in the lower parts covered densely with spreading, white tri- chomes to 3 mm. long, arising from tuberculate bases, in the upper part with shorter, sometimes subappressed ones. Basal leaves petiolate, in- cluding petioles, to 17 cm. long and 32 mm. broad, lamina usually shorter, sometimes equalling or longer than the petiole, lanceolate to obovate, obtuse to acutish, gradually narrowed at base, nerves on the upper sur- face sunken, on the lower surface raised and prominent, both surfaces densely covered with long, spreading, white trichomes, usually arising from large, prominent tuberculate bases, trichomes on the margins of petioles are sometimes even longer; lower cauline leaves acute, usually longer and broader than the basal ones and with much shorter petioles; middle and upper cauline leaves subsessile to sessile, gradually reduced 346 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 in size upwards. Inflorescence terminal or axillary, usually geminate, branches divaricate, ebracteate, short scorpioid, bearing closely set flow- ers, much elongated after flowering, with fruits to 20 mm. apart. Pedicels none to very short in flower, hairy, reflexed to 3 mm. long in fruit. Calyx divided to the base, lobes oblong, erect, ca. 2 mm, long and 0.5 mm. broad, slightly enlarged in fruit, spreading, densely pubescent to sometimes ciliate at the margins, rarely with the trichomes arising from minute tubercles. Corolla pale to deep blue, campanulate, 3.5-4 mm. long, tube + equalling the calyx, limbs equalling the tube, divided nearly to the base, lobes rounded; faucal appendages subquadrate, slightly emarginate, puberulent. Anthers ovate, 0.8 mm. long, 0.4 mm. broad; filament short, inserted nearly at the middle of the corolla tube. Nutlets ovate, apex roundish to acute, excluding the appendages, 2-2.5 mm. long, 1.5—2 mm. broad, dorsally concave, bearing few to many vertical glochidiate ap- pendages, margins appendiculate with the appendages more or less con- fluent at their bases and usually forming a marginal wing, appendages on the ventral side like those on the margins, not confluent at bases and usually smaller. Distripution: Afghanistan, Pakistan, Kashmir, Pamir Alaj, India. 3a. Var. wallichii. Dorsal surface of the nutlets broad-ovate, sunken, sparsely to densely covered with vertical appendages, dorsal longitudinal keel not conspicu- Ous, apex roundish, marginal appendages many, long, dilated and con- fluent at base forming broad elevated wings. West Pakistan: Currra STATE: Brir, 1200 m., Siddiqui & Rahman 26851 (BM). Grucrr AcENcy: Naltar Valley, 2000 m., Duthie 12437 (BM); apanegg : R a pinkoley, 3000 m., Clarke 30408 (x); Gilgit, Toppin 1038 (x); KA np ee ero Lugma Glacier, 3000-3500 m., Russell 1337 (pm); Kushuchun a ie neygian m., Russell 1776 (gm). Hazara Dist.: Kaghan Val- Balakot eon ee m., Kazmi 799A (prs), Shaukat Ali 133 (BM); m Kazmi pee le & Al 5267 (E); Shogran to Sari Rest House, ca. 2500 Tribal territory, Conic arth Inayat 22034 (x); Shinu, Inayat 22035 (B); 1259 (z) age anja Kandoo to Hill Durra Nullah, 3000 m., Burtt & Kazmi Afandi 671 eae Rina; Kurram Valley, Parachinar, Barbour s.n. (BM); Mourree Hits: ¢, AWALPINDI Dist.: Rawalpindi, 1960, McVean s.m. (); rotates! S: near Rosenheim, 2000 m., Rodin 5360 (us). QueTTA DIST.: at, Lace 3401 (E). Swar State: Sho Nala, Rahman 26 (BM). — hh ~~ 1971] KAZMI, BORAGINACEAE 347 Kashmir: Gulmarg, 2700 m., Sept. 1922, Barbour s.n. (Bm), 3500 m., Aug. 1922, Barbour s.n. (pm); Purig, Tangola, Koelz 6075 (us); Parkachen, Koelz 5977 (MICH, GH, us); Pahlgam, Lidder Valley, 2000 m., R. R. Stewart 21623 (cH); Pahlgam, 2100 m., R. R. Stewart 21628, 21762 (cH); Islamabad, 17 Nov. 1896, H. H. Johnston 2 (£); Kootihar Valley, 3 Dec. 1896, H. H. Johns- ton s.n. (E); Banks of Jhelum below Islamabad, ca. 2000 m., Duthie 13059 (x); Kajei-utholi, Padar, upper Chenab, 2000 m., R. R. Stewart 2925 (x), 2919 (x, us); Upper Chenab Valley, Ellis in Gamble 189 (x); Goochi, 2000 m., 172 (kK); Kui Mi m., R. R. Stewart 12553 (GH); Dah, right side of Indus Valley, Schlagintweit 1212 (cH); Marsho, Koelz 6393a (GH); Dras, 3000 m., 31 - 1922, R. R. Stewart s.n. (K); Chi ani Valley, Poosiana, Levinge in Clarke 3b. Var. glochidiatum (Wall. ex Benth.) Kazmi, comb. nov. Cynoglossum glochidiatum Wall. ex Benth. in Royle, Illustr. Bot. Himal. Mount. 1: 306. 1839. Type: Nepaulia, Wallich 922 (x). Icon.: Lindl. in Bot. Reg. t. 15. 1841. ith Dorsal surface of the nutlets ovate, less sunken, covered seer few vertical appendages along the more or less prominent ‘ong o0pe keel, apex obtuse to acutish, marginal appendages at bei : all, to slightly dilated, sometimes confluent, forming 4 nar wing, 00-3500 m., West Pakistan: Gircrr AceNncy: Baltistan, Lencepiaes? bere Duthie 13816 (x); Chutrun, 2000 m., Ludlow 262 paper t and Babusar Pass, reas Silan, 12 July 1899, Znayat s.n. (K); between . 128 (pm), ; miles east-north- ir: Vicinity of Sonamarg, on the Sind river, pba Dickason 106 east of Srinagar, 2000 m., Dickason 107 (MICH, US), | z MICH ; on (at ); Purig, Tangola Ko 6075 (MICH us); vicinity of Pi ; 3 ’ rig, gola, elz 6075 ( ’ ‘ : Pahlgam dder River, 27 road-miles north of Islamabad, Dickason 111 (micH) to Lake Tuliori, 2300 m., R. R. Stewart 21848 (Us). 348 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 4. C. microglochin Benth. in Royle, Illustr. Bot. Himal. Mount. 1: 305. 1839; A. DC. in DC. Prodr. 10: 151. 1846; C. B. Clarke in Hook. f. Fl. Brit. India 4: 158. 1883; Brand, Pflanzenr. IV. 252 (Heft 78): 133. 1921. Type: Kashmir, Royle s.n. (K?). Perennial herb, erect, 35-90 cm. tall. Stems simple or branched, dense- ly covered with white trichomes. Basal leaves petiolate, lanceolate to oblong-lanceolate, on both surfaces densely covered with soft white tri- chomes, attenuated at both ends, 6-20 cm. long including petioles, 12- 30 mm. broad; cauline leaves sessile ovate to elliptic, acuminate, round towards the base, to 6 cm. long, 30 mm. broad, nerves sunken above, raised below, upper surface densely covered with thin, antrorsely sub- appressed to appressed, white trichomes 0.5-1 mm. long, arising from prominent white tuberculate bases, lower surfaces densely albido-tomen- tose; upper cauline leaves smaller in size. Inflorescence terminal or ax- dinal keel, with few scattered vertical glochidiate appendages, appendages on the lateral and ventral sides smaller than those of the margins. Style DIstTRIBUTION: West Pakistan, Kashmir, northwest India. West Pakistan: Hazara Dist.: KacHAN Vazuey: between Balakot and Babusar Pass, Abel 135 (BM). hmir: Pooniana (?), 3000 m., Clarke 28819 (BM). Very little material belonging to Cynoglossum microglochin is available in herbaria. The type collection b ir i t y Royle from Kashmir is also no traceable. Bentham (/. c.), C. B. Clarke (J. c.) and Brand (J. c.) describe 1971] KAZMI, BORAGINACEAE 349 ovate to elliptic, and acute, densely pubescent above, pubescent below, whereas C. nervosum has oblong to elliptic cauline leaves, sparsely hir- sute on both surfaces. From the specimens of both the species available to me, the differentiating characters seem to be quite stable. The two specimens cited above are immature and lack fruits, but they agree in the form of leaves, their indumentum, and in the floral parts with a very good and complete specimen collected from Dalhousie by C. B. Clarke (23 Sept. 1874, Clarke 2298 at Kew) and determined by him, which agrees in all its details with Bentham’s original description. 5. CC, nervosum Benth. ex C. B. Clarke in Hook. f. Fl. Brit. India 4: 158. 1883; Hook. Bot. Mag. 123: ¢. 1713. 1897; Brand, Pflanzenr. IV. 252(Heft 78): 142. 1921; Riedl in Rechinger, FI. Iranica 48: 146. 1967. Omphalodes nervosa Edgeworth ex C. B. Clarke in Hook. f. Fl. Brit. India 4: 158. 1883, nomen nudum. Type: WESTERN Hrmatraya: Kulu, Jalauri Pass, Edgeworth s.n. (K). *c0n,: Hook. Lc. £1713,.1897. Perennial erect herb, to 90 cm. tall. Stems solitary or few, branched above the base; stem and branches covered with crisped white, spreading trichomes, to 1.5 mm. long, arising from tuberculate bases on the lower part; with shorter, subappressed trichomes on the upper part. Basal leaves petiolate, petioles to 4 cm. long, lamina ovate to oblong-ovate, en- tire, rounded at apex, attenuated towards the base, to 4 cm. long and 1.5 cm. broad, covered on both surfaces with soft, white, crisped, spread- ing trichomes, usually arising from minute tuberculate bases; lower cau- line leaves petiolate, petioles to 10 cm. long, lamina lanceolate, acute, gradually narrowed at base, to 12 cm. long and 2 cm. broad, indumen- tum like that of the basal leaves; middle and upper cauline leaves sessile, natrow to broadly lanceolate, to 7 cm. long, gradually reduced in size upwards. Inflorescence terminal or axillary, bifurcate, branches 2-3 cm. long, scorpioid in flower, later elongated to 10 cm., bearing 10-20 mm. distant fruits. Pedicels to 1 mm. long, in flower, sometimes elongated to 3 mm. and reflexed in fruit. Calyx divided to the base, lobes oblong- ovate, obtuse, sparsely to densely covered with long, soft, spreading tri- chomes, dorsally sometimes ciliate at the margins, more or less glabrous within, up to 3.5 mm. long and 1 mm. broad in flower. Spreading and enlarged to 4.5(—5) mm. long, 1.5 mm. broad in fruit. Corolla blue, in- fundibuliform, tube equalling or longer than the calyx, limbs 7-8 mm. in diameter, divided for 4/5 of its length, lobes transversely ovate, ca. 3 mm. long, 4 mm. broad, patent; faucal appendages emarginate, ciliate at the margins, puberulous above, 1 mm. long and broad. Anthers oblong- ovate, 1-1.5 mm. long; filaments 0.3 mm. long, inserted adjacent to the bases of the faucal appendages. Style to 3.5 mm. long, filiform, stigma capitate. Nutlets ovate, to 4.5 mm. long, 3 mm. broad, appendages 350 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 glochidiate, marginal ones dilated at their bases, confluent, forming a narrow or broad marginal wing, dorsal appendages few to many. DistripuTIon: Afghanistan, West Pakistan, Kashmir, northwest India. 5a. Var. nervosum. Petioles of the lower cauline leaves 3-4 cm. long; leaves covered on both surfaces with trichomes usually arising from prominent tuberculate bases; calyx lobes to 6 mm. long in fruit; fruiting pedicels usually shorter than the calyx; corolla 9-11 mm. in diameter. Kashmir: above Tragbal, 3500 m., Osmaston 13 (K); Hunan Nullah, Chamba Pangi, ca. 3000 m., R. R. Stewart 2827 (x); Chamba, Aug. 1880, Ellis 187, 405B (xk). Reported from: Gorai, Meebold 4183; Margan Pass, Meebold 4182. Sb. Var. petiolatum (Hook.) Kazmi, comb. nov. Anchusa petiolata Hook. Bot. Mag. 67: t, 3858. 1841. Cynoglossum petiolatum (Hook.) A. DC. in DC. Prodr. 10: 149. 1846; C. B. Clarke in Hook. f. Fl. Brit, India 4: 158. 1883; Brand, Pflanzenr. IV. 252 (Heft 78): 141. 1921, Type: e Nepalia in hort. bot. Glascovium introduc. Icon.: Hook. 1. c. ¢. 3856. 1841. Petioles of the lower cauline leaves 5—8 cm. long; leaves covered on both surfaces with soft trichomes usually not arising from prominent tu- berculate bases; calyx lobes to 8 mm. long in fruit, fruiting pedicel usu- ally longer than the calyx; corolla 16-18 mm. in diameter. Kashmir: Bhujaz, foot of Umasi La, Zanskar Frontier, 3300 m., Ludlow & Sherriff 9197 (BM). Report f ed from: alpine western Himalaya, Zanskar, neat Umasi Pass, Thomson s.n. (xk), 6. C. stewartii Kazmi, sp. nov. Type: Kasumi: Pahlgam to Lake Sorus, 10,000—11,000 ft., flowers dark blue purple, basal leaves large, 8 Aug. 1945, R. R. Stewart 21584 (cH-holotype). » petioli ad 18 cm. longi, lamina ovata vel ovato- » ad basim plerumque truncata, interdum Taro subcordata, manifeste nervata, ad 26 cm. longa, 12 cm. lata, supra uniformiter sparseque trichomatibus albis, tenu!- bus, patentibus vestita, trichomata ad 0.5 mm. longa, ad basim magni-tu- 1971] KAZMI, BORAGINACEAE 351 berculata, infra trichomatibus densioribus, patentibus vel appressis, etuber- culatis; foliis caulina petiolata, petioli ad 6 cm. longi, plerumque breviores, lamina ovato-lanceolata, acuminata, gradatim attenuata ad basim; folia caulina superiora sessilia vel subsessilia, in magnitudine sursum gradatim reducta. Inflorescentia terminalis vel axillaris, plerumque geminata, in- terdum dichotome ramosa, brevis, scorpioidea, in juventute flores ap- proximati, deinde ad 15 cm. longa et fructibus distantibus ad 15 mm. Pedicelli ad anthesin perbreves; in fructu calycem aequantes, vel brev- iores vel interdum longiores. Calyx paene 4/5 divisus, lobi oblongo-ovati, obtusi, 3.5-4.5 mm. longi, 1.5—2 mm. lati ad anthesin, extus trichomati- bus mollibus albis antrorse appressis vestiti, trichomata ad 0.5 mm. longa, intus densiore vestiti, trichomata breviora; lobi ad margines albido-to- mentosi, in fructu ad 5 mm. longi et 3 mm, lati, patentes. Corolla pur- pureo-caerulea, infundibuliformi-campanulata vel rotata, tubus calycem aequans, 4-5 mm. longus, limbus ad basim divisus, lobi late ovati, pa- tentes, 3-4 mm. longi, 3.5-4.5 mm. lati; appendices faucales trapezi- formes, subemarginatae, puberulae. Antherae ovatae, 1.6 mm. longae, 0.7 mm, latae; filamenta perbrevia, supra basim ca. 3.5 mm. inserta. Nuculae late ovatae, ca. 4 mm. longae, 2.5 mm. latae, plus minusve acu- tae, bases rotundatae, margines appendiculatae, appendices ad 1.5 mm. longae, 0.5 mm. latae ad basim, apex glochidiatus, bases liberae vel sub- confluentes cum ala angusta marginale, dorsum planum, glabrae vel in- terdum puberulae, longitudinaliter subcarinatae, 1-3 appendicibus verti- calibus, appendices laterales et ventrales erectae, glochidiatae. Stylus 2.5 mm. longus, filiformis, stigma capitatum. DIsTRIBUTION: West Pakistan, Kashmir, northwest India. West Pakistan: Hazara Dist.: KAGHAN VALLEY: Kund Bangala, 2500 m., Kazmi 2572 (pres); Bala Bhurj, 3000 m., Inayat 19520 (kK); Kamalban, 17 June 1899, Inayat s.n. (K), Kazmi 2528a, 2532a (PES); Silan, 12 July 1899, Jnayat s.n. (Kk). Rawatpinpr Dist.: Murree Hills, along the pipeline from Murree north to Nathia gali, 2500-3000 m., Dickason 109 (micH). Swat STATE: Bishigram, 3000 m., R. R. Stewart & Rahman 24957 (pM); Bahrain, 2500 m., Kazmi 2510 (PES); above Utror, 2500-3000 m., R. R. Stewart & Rahman 25228 (BM). Kashmir: Vicinity of Pahlgam, on east Lidder River, 27 road-miles north of Islamabad, Dickason 108 (MicH); Pahlgam to Lake Sorus, 3000-3500 m., R. R. Stewart 21584 (cH-type); Pahlgam, R. R. Stewart 8319, 21723a (GH), 2500 m., R. R. Stewart 21504 (cH, K); Ferozpur Nullah near Gulmarg, 3000 m., R. R. Stewart 12963 (cH); Sonamarg, R. R. Stewart 6354, 7198 1/2 (K), 3000 m., R. R. Stewart 12960 (GH); Chinari, Kazmi 241 (PEs); Gulbarg, 2700 m., R. R. Stewart 12961 (cH); Jhelum Valley, Tilpatra forests, near Rampur, Ludlow & Sherriff 7714 (cH); Sinp VALLEY: Gangangir, hill face, Ludlow & Sherriff 7970 (pm); Gund, 2000 m., Ludlow & Sherriff 7695 (pm); Kishanganga Valley, Sharda, R. R. & I. D. Stewart 17769 (mtcH). Miscellaneous: Himal. bor. occ., 6000 ped., Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (BM). The specimens cited above are mostly named as C. wisliagscirs ae: microglochin in herbaria, but they are very different iron ee 352 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 species and deserve specific recognition based on the following dis- tinguishing characters. Cynoglossum stewartii is a tall perennial, taller than either of the species cited above; its basal leaves are very large, petiolate with petioles to 18 cm. long; lamina ovate to ovate lanceolate, truncate at base, abrupt- ly narrowed or rarely subcordate, to 20 cm. long and 12 cm. broad. In the other two species petioles and lamina of the basal leaves never ex- ceed 10 cm. and 12 cm. in length, respectively. In shape the lamina are ovate to oblong-ovate and always gradually narrowed towards the base. The middle and upper cauline leaves in C. stewartii are distinctly pe- tiolate, with petioles sometimes to 6 cm. long, and the uppermost leaves subsessile to rarely sessile, these are gradually narrowed towards the petioles at base. In C. nervosum and C. microglochin the middle cauline leaves are sessile to subsessile and the upper ones distinctly sessile. In C. nervosum the upper cauline leaves are more or less narrowed, but in C. microglochin these are obviously rounded at base. The trichomes on both surfaces of the leaves of C. stewartii do not exceed 0.5 mm. in length, whereas in both the other species they are 0.5—1 mm. long. In form of the calyx and corolla, C. stewartii closely resembles C. microglochin, ex- cepting for some minor details given in the descriptions. SPECIES NOT SEEN C. schlagintweitii (Brand) Kazmi, comb. nov. Adelocaryum Schlagintweitii Brand, Repert. Sp. Nov. 13: 548. 1915; Pflan- zenr. IV. 252 (Heft 78): 78. 1921. Tyre: Tiset: Provinz Dras: Kargil, zwischen Suru und Tsringmat, Schlagintweit 7211 (Breslau). Perennial subshrubby herb, covered all over with short trichomes aris- ing from tuberculate bases. Basal leaves petiolate, oblong to lanceolate, (including the petioles) 5-6 cm, long, 5-8 mm. broad, obtuse, attenuate towards the base; lower cauline leaves short petiolate, the upper sessile, much reduced. Inflorescence long, lax, many flowered. Calyx lobes ob- long, obtuse, pilose. Corolla campanulate, 3 mm. long, lobes broadly ovate, tube short; faucal appendages broader than long, semilunar. Sta- mens inserted on the lower part of the corolla tube, filaments short. Style nearly half the length of the calyx. Nutlets (usually two abortive) ovoid, 4 mm. long, appendiculate at the margins, appendages glochidiate at apex, dorsally carinate, the ventral scar very small. DistRIBUTION: Kashmir, southwest Tibet. Kashmir: Kargil, between Suru and Tsringmat, Schlagintweit 7211 (Breslau). C. flexuosum (Brand) Kazmi, comb. nov. Adelocaryum flexuosum Brand, R WW. 252 (Heft 78): 78.1921. epert. Sp. Nov. 14: 145, 1915; Pflanzen 1971] KAZMI, BORAGINACEAE 353 Type: KasHMir: Gurez (nortlich von Srinagar, der Haupt Stadt von Kaschmir), 3 und 4. 10. 1856, Schlagintweit 7612 (Breslau). Annual (?) herb, stems delicate, flexuous, to 30 cm. or longer, cov- ered with spreading trichomes. Basal leaves not seen; cauline leaves short petiolate, the upper sessile, oblong to lanceolate, 3-4 cm. long, 10-16 mm. broad, attenuate at both ends. Inflorescence terminal, lax, many flow- ered. Pedicels of the lower fruits longer than the calyx. Calyx lobes ob- long-elliptic, obtuse, to 3 mm. long. Corolla cylindric-campanulate 4.5 mm. long, tube equalling the limb; faucal appendages subquadrate, apex emarginate. Stamens inserted at the middle of the corolla tube, anthers reaching the bases of faucal appendages, filaments short. Style equal or subequal to the calyx. Nutlets 3 mm. long, dorsally appendiculate, ap- pendages glochidiate at apex. DISTRIBUTION: West Pakistan. West Pakistan: Gitcir AGENCY: Gurez, Schlagintweit, 7612 (Breslau). Since I did not see the types of the two species cited above, the de- scriptions given here are according to Brand (1. c.). From the descrip- tions the taxa appear to belong to the genus Cynoglossum, and the species to be closely related to C. wallichii. On examination of the original ma- terial the species may, however, prove to be the same as or varieties of C. wallichii or C. lanceolatum. 26. Actinocarya Benth. in Benth. et Hook. f. Gen. Pl. 2: 846. 1876. Type species: A, tibetica Benth. Slender procumbent or diffuse annual herbs, sparsely strigose or nearly glabrous. Leaves alternate. Flowers minute, solitary on axillary pedi- cels. Calyx divided to the base, 5-lobed, hardly enlarged in fruit. Corol- la campanulate, lobes broad-ovate, slightly shorter than the tube, ob- tuse, spreading, imbricate in bud; faucal appendages 5, very small, trans- versely linear. Stamens 5, inserted in the lower part of the corolla tube, anthers included, small, ovate, obtuse; filaments short. Ovary deeply 4- lobed; style short, stigma small. Nutlets 4, obovoid-oblong, appendiculate, appendages glochidiate at apex, free or dilated at base, confluent, form- ing marginal wings or a coniform dorsal crest, attached apically or sub- apically to a plane gynobase, turgid at the margins. : Species 2, distributed in Kashmir, southwest Tibet and northwest India. KEY TO THE SPECIES me Leaves glabrous on the upper surface, with few short trichomes below; oa pals bearing scattered short trichomes externally; fruits densely covere with minute, frequently uncinate trichomes and scattered appendages. .... 1, A. tibetica. pat] a Oo & < oO wa = = a — ) i=) aq ~ =e a am ° 3 Oo wn ° 3 o 5 wn = a5 rtf a is] wa ry =} OQ FX) cr oo > fa] 5 9 ial w. =] n . wn s Ee a g ry 354 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 ing few scattered trichomes externally, ciliate at the margins; fruits glabrous, appendiculate, appendages forming a dorsal crest. ........... 2. A. acaulis. 1. A. tibetica Benth. in Benth. & Hook. f. Gen. Pl. 2: 846. 1876; C. B. Clarke in Hook. f. Fl. Brit. India 4: 154. 1883; Giirke in Nat. Pflanzenfam. IV. 3a: 101. 1893; Brand, Pflanzenr. IV. 252(Heft 78): 162. 1921. Type: West Trset: Nubra, alt. 13,000 ft., near Karsar village, Thom- son s.n. (K-holotype, GH-isotype) Icon.: Hooker, Icon. 23: t. 2256. 1893. Annual, decumbent or erect herb, to 15 cm. tall. Stem solitary or many, slender, weak, unbranched, usually glabrous in the lower parts, with few scattered, spreading, minute, trichomes above. Basal leaves few, obo- vate-spathulate, entire, sometimes slightly revolute at margins, apex round, narrowed towards the base into long petioles, (including petioles) to 25 mm. long, 1.5-5 mm. broad, glabrous on both surfaces, or the low- er surface bearing few scattered, retrorsely subappressed minute tri- chomes arising from tuberculate bases; cauline leaves usually with shorter petioles. Flowers axillary throughout the length of the plant. Pedicels short, in flower erect in the axils of the lower leaves, up to 25 ceolate to broad-lanceolate, obtuse to subacute, externally sparsely pi- lose, ca. 1.5 mm. long, 0.5 mm. broad in flower; erect, not at all to slightly enlarged in fruit. Corolla light blue, campanulate, tube usually shorter than the calyx, ca. 1 mm. long, limb equalling or slightly longer than the tube, divided nearly to the base, lobes broad-ovate, ca. 1 mm. long and broad, spreading; faucal appendages very small. Stamens 5,. anthers small, ovate, obtuse, included, filaments short. Ovary 4-lobed, style Short, stigma small, capitate. Nutlets 4, grayish, obovoid-oblong, divaricate, ca. 2 mm. long, bearing numerous minute, frequently uncinate trichomes as well as numerous appendages, appendages glochidiate at apex, 1-1.5 mm. long, scattered, not forming a dorsal crest. DISTRIBUTION: West Tibet, Kashmir, northwest India (?). ie a near Karsar village, Thomson s.n. (K, GH); Ladakh: Nanu, » hoe" 2358 (cH); Henus, 3300 m., Koelz 2558 (cH). Reported from: Kambaj i ‘ “a 5 . oS 1882 ( oe Prain s.n. (in Herb. Bornmiiller) ; without locality, King’s collector, . ae (W. W. Smith) I. M, Johnston, Jour, Arnold Arb. 21: 53- Eritrichium acaule W. W. Smith, Rec. Bot. Surv. India 4; 225. 1911. pies —_ La, Llonakh, Sikkim, top of Pass, 17,500 ft., 3 Aug. 1909, . ave 2811 (cat-holotype, cx-, K-isotypes). 1971] KAZMI, BORAGINACEAE 355 Annual prostrate compact herb, sometimes branched near the base; branches many, very short, fleshy, covered with many scattered, spread- ing trichomes to 0.5 mm. long, arising from tuberculate bases. Leaves obovate-spathulate, entire, te slightly revolute at the margins, apex round, narrowed at base into long, usually winged petioles + dilated below, (including petioles) i 20 mm. long and 4 mm. broad, bearing on both surfaces and margins crisped, spreading trichomes, to 0.7 mm. long arising from tuberculate bases; cauline leaves, if any, shorter. Flowers solitary in the axils of leaves or sometimes few axillary and congested on the top of the short branches. Pedicels short in flower, to 10 mm. long, fleshy, in fruit bearing many spreading trichomes. Calyx divided nearly to the base, ca. 1.5 mm. long in flower, lobes erect, ovate to ob- long, obtuse to subacute at apex, bearing few scattered trichomes dor- reap with ciliate margins, enlarged to 2 mm. long and + 1 mm. broad n fruit. Corolla blue, campanulate, tube usually equalling a calyx, limbs shorter than the tube, divided to the base, lobes rotundate, spread- ing; faucal appendages very small. Stamens 5, included, anthers ovate, small, filaments short. Nutlets 4, ovoid-oblong, glabrous, appendiculate, appendages to 1 mm. long, glochidiate at apex, dilated at base, confluent and forming a dorsal crown. DISTRIBUTION: Kashmir, northwest India, Sikkim. Kashmir: LapAKH: Rupshu, Kiangchu, Lachulung, 4500 m., Koelz 6650 (cH). aay from: Rupshu, Purga, Koelz 2184; Nanu, Koelz 2358. Both species of Actinocarya, widely distributed in the Himalayas, are very similar in gross habit. A. tibetica is apparently glabrous though the leaves are pustulate and strigose beneath. The dimorphic fruits bear numerous minute, frequently uncinate trichomes. In A. acaulis the stem and the leaves bear relatively coarse, elongate, loosely appressed hairs. The fruit is nearly glabrous and the secondary glochidiate appendages are fewer, tending to be crowded just outside the coarse dorsal margin of the nutlet. The fruits of A. tibetica are dimorphic. Those borne along the stems are long pedicellate and composed of apically attached nutlets bearing a small crown on the back towards the base. Outside this small crown are numerous scattered glochidiate appendages. The fruits produced about the base of the plant have short stout pedicels and more compressed nutlets. The dorsal crown is incomplete on these, or entirely absent. Nu- merous scattered glochidiate appendages, are however; always present. 27. Lithospermum L. Sp. Pl. 132. 1753; Gen. Pl. ed. 5. 64. 1754. Buglossoides Moench, Meth. 418. 1794; I. M. Johnston, Jour. Arnold Arb. 38: 38. 1954, Type spectrs: L. officinale L. (Lectotype species). Annual or perennial, erect or decumbent, hispid or rough herbs. Leaves 356 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 alternate. Flowers white, blue or yellow, axillary and solitary or in ter- minal bracteate racemes. Calyx divided to or nearly to the base, usu- ally enlarged in fruit. Corolla infundibuliform to hypocrateriform, either longitudinally plicate or gibbous at throat, lobes imbricate in bud, later suberect to spreading. Stamens 5, included; anthers oblong, obtuse or apiculate; pollen encircled with a single row of 6 to 9 pores. Ovary deeply 4-lobed; style usually filiform, stigma bifid. Nutlets 4, erect, ovoid, stony, smooth, rough or tuberculate, attached by a basal scar to the nearly flat gynobase. A polymorphous genus of 50 to 60 species, distributed mainly in Ameri- ca. About 10 species are known in Africa and about 10 in Europe and Asia, including three species from West Pakistan and Kashmir. KEY TO THE SPECIES 1. L. arvenis L. Sp. Pl, 132. 1753; Ledeb. Fl. Rossica 3: 129. 1847; Boiss. Fl. Orient. 4: 216. 1875; C. B. Clarke in Hook. f. Fl. Brit. India 4: 174. 1883; M. Pop. Fl. URSS 19: 165. 1953; Riedl in Rechinger, Fl. Iranica 48: 151. 1967. Margarospermum arvense Decne. in Jacquem. Voy. Inde Bot. 122. 1844. Type: “Habitat in Europae agris et arvis,” Herb. No. 181.4 (LINN). Icon.: Hegi, Ilustr. Fl, Mitt. Eur. 5(3): fig. 3112, a—c. 1927. ioles; upper cauline leaves, subsessile to a cup to hold the nutlets in fruit. Corolla white to light blue, tubular-in- tely strigose, corolla tube longitudinally pli- 1971] KAZMI, BORAGINACEAE 357 cate from the top of the anthers to the throat, glandular or minutely hairy, 4-5 mm. long, lobes ovate, spreading, to 1 mm. long. Anthers oblong, apiculate, roundish at the base, to 1 mm. long; filaments very short, in- serted 1-1.5 mm. above the corolla base. Nutlets 4, gray, pink, or light brown, ovate, usually abruptly narrowed in the upper half, tuberculate all over the surface, 3 mm. long, 2 mm. broad; areola basal, small. DistRIBUTION: Europe to Japan. West Pakistan: CHITRAL STATE: Oihor Gol, 3000 m., Bowes Lyon 841 (pm): Brumboret Gol, N. of Drosh, Stainton 2226 (BM). Hazara Dist.: Abbottabad, 1300 m., R. R. Stewart 13690A (cH, us). KHYBER AGENCY: Khyber Pass, 1100 m., H. H. Johnston 143 (r); Forsapper, Khyber Pass, 1700 m., H. H. John- Ston 61 (£). Quetta Dist.: Environ de Quetta, Schmid 1572 (c). RAwat- PINDI Dist.: Dhamyal near Rawalpindi, 500 m., R. R. Stewart 19389A (cH); Topi Park, Rawalpindi, Pinfold 10 (pM); Murree Hills, Ghora Gali, 1700 m., field weed, R. R. Stewart 12951 (GH), ir: Srinagar, 1700 m., R. R. Stewart 12952 (GH); near Srinagar, R. R. Stewart 63 (pm). Miscellaneous: Himal. bor. occ., Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (GH). 2. L. tenuiflorum L. f. Suppl. 130. 1781; DC. Prodr. 10: 75. 1846; Boiss. Fl. Orient. 4: 217. 1875; C. B. Clarke in Hook. f. Fl. Brit. India 4: 175. 1883; M. Pop. Fl. URSS 19: 164. 1953; Riedl in Rechinger, Fl. Iranica 48: 150. 1967. Myosotis tenuiflora Viviani, Fl. Libycae, 9. 1824. Type: “Habitat in Aegypto,” Latourette, Herb. No. 181.13 (LINN). Icon.: Jacq. Icon. Pl. Rar. 2: t. 313. 1793; Vivintls L ©.8.. 1 ee. 2 1824, Very similar to Lithospermum arvensis except that the plants are ro- bust; stems decumbent, many, usually branched, to 30 cm. long; leaves larger, cauline leaves usually sessile with a broad base; corolla 5-6 mm. long; areola of the nutlets broad and firmly attached to the gynobase. Disrripution: North Africa and southeastern Europe to Central Asia, Afghanistan, West Pakistan, Kashmir, northwest India. 73.1898, Duthie (kK), PESHAWAR Dist.: Islamia College, ca. 350 m., Kazmi 2528 (PEs); between Peshawar and Tarnab Farm, ca. 350 m., Kazmi 2481 (PES); P abbhi-Nowshehra, Kazmi 2470 (pes); Peshawar, ca. 350 m., Kazmi 21b (res), K U = = Lamond 1318 (E); Quetta, Duthie 8680 (pm). RAWALPINDI Dist: Rawalpindi, 358 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 10841 (kK); Tangmarg, 2200 m., R. R. Stewart 12953 (GH). Miscellaneous: Himal. bor. occ., Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n, (BM, cn), dated 4. 4. 1848 (x). 3. L. officinale L. Sp. Pl. 132. 1753; Ledeb. Fl. Rossica 3: 130. 1847; Boiss, Fl. Orient. 4: 218. 1875; C. B. Clarke in Hook. f. FI. Brit. India 4: 175. 1875; M. Pop. Fl. URSS 19: 162. 1953; Riedl in Rechinger, Fl. Iranica 48: 149. 1967. Margarospermum officinale (L.) Decne. in Jacquem. Voy. Inde Bot. 122. 1844. {This reference given in Index Kewensis, but the combination was not properly made by Decaisne. | Type: “Habitat in Europae ruderatis,” Herb. No, 181.1 (LINN). Icon.: Reichenb. Icon. Fl. Germ. 18: ¢. 112. J, 1-4. 1858; Hegi, Fl. Mitt.-Eur. 5(3): fig. 3115, a~f and 3116. 1927, Erect perennial herb, to 100 cm. tall. Stems solitary or many, branched, branches to 20 cm. long; stem and branches covered with subappressed to patent trichomes, to 1.5 mm. long, arising from minute or large tuber- culate bases. Leaves lanceolate, entire, sometimes slightly revolute, acute, to 8 cm, long, 20 mm. broad, covered with thin or thick, antrorsely sub- appressed trichomes with tuberculate bases, to 1 mm. long. Inflorescence bracteate, bracts leaflike. Pedicels very short in flower, erect, pubescent, elongate, to 4(-5) mm. long in fruit. Calyx divided to the base, lobes narrowly lanceolate, 3-4 mm. long in flower, not much enlarged, pubescent, trichomes spreading in fruit. Corolla yellowish, greenish or sometimes nearly white, hypocrateriform, 4-6 mm, long, corolla tube usually equal- ling, sometimes slightly longer or shorter than the calyx, 2.5-3.5(—4) mm. long, limb 4 mm. in diameter, lobes spreading, 1-1.5 mm. long, corolla throat with 5 trapeziform, intruded, distinctly invaginated appendages, densely velvety at apex, glanduliferous (especially on the sides). Fila- ments inserted at the middle of the corolla tube. Style 1-2 mm. long; stigmas 2, small.- Nutlets ovate, white, smooth, shining, 3-4 mm. long. DISTRIBUTION OF SPECIES: From Europe to central Asia. The plants from Afghanistan, West Pakistan, and Kashmir are dif- ferent from the specimens of Lithospermum officinale collected from the other areas of its distribution in havin longer, stouter, and spreading trichomes on the stems. They are distinguished here as a variety. Var. stewartii Kazmi, var. noy. Type: Kasi: Banehal Pass, 9000 ft.. Tul 1-4, 1931, R. R. Stewart 12182 (cH-holotype). hos Setar he A varietate officinali differt caulibus hispidulis, pilis gracilibus, rectis, divaricatis donatis. DISTRIBUTION OF VARIETY: Afghanistan, West Pakistan, and Kashmir. 1971] KAZMI, BORAGINACEAE 359 West Pakistan: KurrAM AGENCY: Kurram Valley, Aitchison 338 (GH). Kashmir: Banehal Pass, R. R. Stewart 12182 (GH-type); Sind Valley, Gan- darbal, 1800 m., Ludlow & Sherriff 7621 (Bm); Tilal Valley to Chorwan, ca. 3000 m., R. R. & I. D. Stewart 18599 (Gu); vicinity of Sonamarg, on the Srinagar, C. B. Clarke 29100A (BM). Miscellaneous: Kashmir, Kohli 1 (x); Himal. bor. occ., 6000 ped., Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (BM, GH). The new variety closely resembles Lithospermum erythrorhizon Sieb. & Zucc, (Abh. Bayer. Akad. Wiss. 4(3): 149. 1846) from eastern Asia in having spreading trichomes on the stem. In characters of the calyx and corolla, however, the plants of var. stewartii agree with typical Litho- spermum officinale. In L. erythrorhizon, unlike the condition in L. of- ficinale, the corolla limb is spreading and formed of rounded, equally long and broad corolla lobes, in diameter it is commonly equal to the total length (5-9 mm.) of the corolla. The fruiting calyx is usually 5-10 mm. long and several times longer than the nutlets. 28. Arnebia Forssk. Fl. Aegypt.-Arab. 62. 1775. TYPE SPECIES: Arnebia tetrastigma Forssk. Annual to perennial, erect to decumbent hispid herbs. Stems single to many, branched or unbranched, arising from a tap root. Inflorescence usually terminal, simple or forked, of scorpioid cymes, bracteate. Calyx divided nearly to the base, lobes accrescent, usually shorter than the co- Flowers heterostylic. Style slender, simple or simply forked or bifid, in- cluded to slightly exserted; stigmas 2-4, capitate, oblong or flabellate, simple or slightly bilobed. Usually one to four nutlets developing, these conic-ovoid, pyramidal or lance-ovoid, usually broadest at base, rough, tuberculate, verrucose, rugose or rugulose, surface usually dull; venter usually angulate, keeled, attachment basal, large, horizontal or suboblique. Gynobase flat or broadly pyramidal, attachment surface separated by lineal grooves, plane or somewhat concave. Species 20 to 22, distributed from North Africa to Central Asia and the Himalayas. KEY TO THE SPECIES a. Plants distinctly perennial; corolla tube without annulus at base. b. Leaves with midrib only; inflorescence globose or broader than long. 360 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 c. Plants gray or silvery, stem arising from the center of the basal leaf- cluster, less than 15 cm. tall, leaves less than 5 Iruiting calyx accrescent, the tube forming a flattene wee “Puliform disk 12-17 mm, in diameter, this often sericeous Be ber oh Be: sharply reflexed at an abrupt right angle, the lobes ovate veins aoe €, 5-10 mm. long, 7-12 mm. broad, striate with coarseé » obtuse or rounded at apex; fruit soon glabrate, glabose to oblate, 1971] SMITH, PACIFIC ISLAND PLANTS, XXIII 399 20-30 mm. long, 25-30 mm. broad, the seeds 8 (or perhaps 10), 10-15 mm. long. TYPE LOCALITY: Samoa; the type, without definite locality other than “Tutuila and Savaii,” is a U.S. Exploring Expedition specimen cited be- ow. DistRIBUTION: Eastern Fiji (Lau Group), Horne and Wallis Islands, Tonga, Niue, and Samoa. The species is also accredited to the Solomon Islands and New Guinea by Bakhuizen (2: 224), but it is likely that these records should be more carefully examined, since a logical western bound- ary for the taxon is noted in the Lau Group of Fiji. The species has been noted as a tree 3-25 m. high, with a trunk 8—45 cm. in diameter, oc- curring in often dry forest, woodland, or thickets, frequently near the sea on limestone but also on overgrown lava fields. Altitudes from sea-level to 600 m. have been recorded. The corolla is white to cream-colored or yellowish white; the fruit is yellow to pink, red, or dark brown, with blackish or dark brown seeds. A vaguely seasonal rhythm is noted in that flowers have been obtained from September to March, fruits from May to November. LocaL NAMES AND USES: In Lau the name kaukauloa has been recorded, and it is noted that the wood furnishes a hard black “ebony.” On Alofi (Horne Islands) the species is known as kau kaw uuli or tutunu, and the wood is used for construction. Common names in Tonga, recorded by Yuncker and other collectors, are tutuna, pekepeka, kolivao, and kaka- la’uli; there the species is used for firewood. On Niue both Yuncker and Sykes indicate the name as kieto; the hard black wood is used for spears. Variant names in Samoa are aoaoli, aoauli, auaouli, aw’auli, and au’a’uli, with a single record of anume (Christophersen 1000); the last more com- monly refers to D. elliptica. Fiji. VANUA Mpatavu: Southern limestone section, Malatta, Smith 1451 (BISH, GH, K, Ny, UC, US). FULANGA: In ridge forest, Bryan 448 (A, BISH, K, UC, US); on limestone, Smith 1226 (BISH, GH, K, NY, UC, US). ONGEA LEVU: . Wallis Islands. Uvea: Mt. Afala, McKee 19921 (P). Tonga. Nrvaro’ou: Jaggar, Oct. 1930 (BIsH). Vava’'u: Crosby 98 (K); be- low Leimatua, northwest of Neiafu, Yuncker 16191 (BISH, GH, US) ; Tuanikavale, Soakai 765 (x). “Vava’u and LiruKa:” Harvey, in 1855 (kK). Kao: Yumcker 15967 (BIsH, us). Torua: Northern side of island, Scarth-Johnson 162 (x). Eva: Near Lokupo, northeastern side, Yuncker 15521 (BISH, US); above me, northwestern side, Yuncker 15495 (BISH, GH, US), 15496 (BISH, us); iku terrace, Parks 16321 BISH, GH, K, NY, UC, US). : ; Niue. East of Alofi, yeu 10064 (A, BISH, - uc, us); Vaipapahi farm, Ykes 472 (x); Tuhia’atua-Hakupu track, Sykes 774 (K). i Samoa. Se Between re and Asau, Christophersen 3432 . . Tanga, Christophersen 2821 (bISH, K); Salailua-Lataitai, Christopherse 400 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Hume 2620 (BISH, K, NY, UC, US), 2624 (BISH, UC); Safotu, Vaupel 483 (BIsH, Ny), Christophersen & Hume 2531 (BISH, K, NY, UC, US); Safune, Christopher- sen 2396 (BISH, K, US); Matavanu lava field, Christophersen & Hume 1875 (BISH, us); Neiafu, Bristol 2203 (Bis); Gataivai, Bristol 2249 (Bisu, K), 2271 (BIsH). Upotu: Falese’ela, Lefaga, Bristol 2316 (BisH, K); vicinity of Apia, Reinecke 107 (BIsH, kK, us), Eames 121 (BIsH, Ny); Lanutoo, Christophersen 387 (BISH, K, NY); Vaea Mt., Christophersen 465b (BIsH, NY); below Malolole- lei, Christophersen 291 (IsH, uc, us). TuTUILA: Papatele ridge, Christopher- sen 1000 (BISH, K, NY, UC); vicinity of Pago Pago, Diefenderfer 16 (BIsH); Fa- The overly inclusive nature of some past interpretations of D. samoensts has been discussed above in my comments on Subgenus Diospyros. In fact, this species is sharply demarcated from its closest ally, D. vitiensis, and has a very natural distribution. 7. Diospyros vitiensis Gillespie in Bishop Mus. Bull. 74: 14. fig. 17. 1930; Bakh. in Bull. Jard. Bot. Buitenz. III. 15: 457. 1941. Tree 3-15 m. high, the young parts minutely strigose, the branchlets and foliage very early glabrate, the branchlets terete, usually copiously and conspicuously lenticellate; petioles semiterete, 3-14 mm. long; leaf- blades chartaceous or subcoriaceous, glabrous, often punctulate-glandular, ovate-oblong or elliptic or elliptic-lanceolate, 7-18 cm. long, 3.5-12.5 cm. broad, obtuse to cuneate at base, obtusely short-acuminate at apex, the mersed; inflorescences axillary, congested-cymose, the staminate short- pedunculate or glomerulate and short-branched from base, 3—10-flowered, pale-sericeous, the bracts deltoid, about 1 mm. long; calyx 3- or 4-lobed, sparsely strigillose without and soon glabrate, about 4 mm. long and in diameter, th 10-12 mm. etd ae calyx accrescent, glabrous, the tube forming 4 lobes ite ~15 mm. in diameter, the limb spreading or reflexed, the =p Re oho, 4 or 5 (very rarely 3), deltoid to oblong-ligulate, se ONG oe, broad, subacute to obtuse at apex; fruit soon 1971] SMITH, PACIFIC ISLAND PLANTS, XXIII 401 glabrate, ellipsoid to obovoid to globose or oblate, 20-40 mm, long, 20- 0 mm. broad, the seeds usually 8, about 15 mm. long. Although in my observation this species is a Fijian endemic, Bakhuizen (2: 457) cites in the synonymy, with question, D. acris Hemsl. (in Kew Bull. 1895: 136. 1895), typified by Comins 311, from the Torres Islands, New Hebrides. At the moment no evidence indicates the advisability of giving up the Fijian binomial for Hemsley’s older one. Points of separa- tion between D. vitiensis and D. samoensis are discussed above. I have not discovered a reliable means of separating the two suggested varieties of D. vitiensis in the absence of mature fruits, but the extremes of the fruiting calyx are such that it seems inadvisable to ignore them. Fortunately specimens of Diospyros are more often collected in fruit than in flower. It must be acknowledged that the fruiting calyces are some- what variable on a few collections, and some of these have been referred to one or the other variety without much conviction. The type of D. slobosa, with staminate flowers, is referred to var. longisepala because another Kambara collection, Smith 1272, has obviously elongate fruiting calyx-lobes. Future material from Lau will hopefully indicate whether only var. longisepala occurs there. KEY TO VARIETIES Calyx-lobes associated with mature fruit reflexed, deltoid, often about as broad as long, 5-10 (-12) mm. long, 5-12 mm. broad at base. .. 7a. var. vitiensis. Calyx-lobes associated with mature fruit spreading or reflexed, oblong-ligulate or deltoid-oblong, conspicuously longer than broad, 8-18 mm. long, 3-10 my broad at base... 0), BATE ae ee 7b. var. longisepala. 7a. Diospyros vitiensis var. vitiensis. Diospyros vitiensis Gillespie in Bishop Mus. Bull. 74: 14. fig. 17. 1930. Diospyros samoensis var. longisepala sensu J. W. Parham, Pl. Fiji Isl. 161, quoad fig. 61, A (err. legend as B), non sensu typi. 1964. : The typical variety, characterized by having the calyx-lobes sae with the mature fruit reflexed, deltoid, often about as broad as long, 5— (-12) mm. long, 5-12 mm. broad at base. peter Type Locatiry: Province of Namosi, Viti Levu, Fiji; the type 1s Gil- les pie 3083, cited below. Viti Distripution: Endemic to Fiji, and known definitely only Boz ne Levu and Vanua Levu, where it has been noted as an often Pay eS 3-15 m. high in dense, dry, or secondary forest at altitudes of re oe The fruit is greenish yellow to dull yellow, becoming pomuies and or brown. Fruiting specimens have been obtained between May December. ie bulumate, mbole. CAL NAMES: Kauloa, mbaumbulu, m ; orthern slopes Fiji. Virt Levu: Mpa: Fiji Dept. Agr. 14468 (BISH, SUVA); DOTEACE : ; : . s); vicinity of Mt. Namendre, east of Mt. Koromba, Smith 4543 (A, BISH, ™ vs) 402 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 of Nandarivatu, Gillespie 4207 (BIsH, uc), Fiji Dept. For. 262 (Watkins 784) (K, suvA); hills between Nggaliwana and Nandala Creeks, Smith 5817 (A, BISH, K, US); vicinity of Nandala, Degener 14835 (A). NANDRONGA & Navosa: North- ern portion of Rairaimatuku Plateau, between Nandrau and Rewasau, Smith 5446 (A, BISH, K, US). SeRUA: Hills between Waininggere and Waisese Creeks, between Ngaloa and Wainiyambia, Smith 9379 (BISH, GH, K, SUVA, UC, US); hills east of Navua River, near Nukusere, Smith 9086 (BISH, GH, K, SUVA, UC, US); Navua River near Namata Rapids, Gillespie 2950 (sisH). Namosi: Mt. Naitara- ndamu, at Navunitaruilau, on wooded ridge between the watersheds of the Wainimala and the Wainikoroiluva, alt. 900 m., Sept. 27, 1927, Gillespie 3083 (BIsH holotype; isotypes at K, Ny, UC, Us); east of Namosi Village, Gillespie 2829 (BISH, UC). VANUA Levu: Matuuata: Nakoroutari, Fiji Dept. Agr. 15239 (BISH, K, SUVA). THAKAUNDROVE: Hills between Vatukawa and Wainingio Rivers, Ndrekeniwai Valley, Smith 584 (BIsH, GH, K, NY, UC, US); hills west of Korotasere, Natewa Bay region, Smith 1946 (BIsH, GH, K, NY, UC, US). 7b. Diospyros vitiensis var. longisepala (Gillespie) A. C. Sm., comb. nov Diospyros longisepala Gillespie in Bishop Mus. Bull. 74: 14. fig. 16. 1930. Diospyros samoensis var. longisepala Fosberg in Bull. Torrey Bot. Club 67: 418. 1940; Bakh. in Bull. Jard. Bot. Buitenz. III. 15: 407. 1941; J. W. Parham, Pl, Fiji Isl. 162, excl. fig. 61, A. 1964 Maba globosa A. C. Sm. in Bishop Mus. Bull. 141: 121, sensu typi, excl. jig. 3. 1936 63. p Diospyros globosa Fosberg in Bull. Torrey Bot. Club 65: 612. 1939; J. W. Parham, Pl. Fiji Isl. 160, 1964. Scarcely separable from the typical variety in foliage or flowers, but characterized by having the calyx-lobes associated with the mature fruit often spreading, sometimes reflexed, oblong-ligulate or deltoid-oblong, con- spicuously longer than broad, 8-18 mm. long, 3-10 mm. broad at base. TYPE LOCALITY: Mba Province, Viti Levu, Fiji; the type is Gillespie 4300, cited below. The type of Maba globosa is Smith 1241, cited below, from the island of Kambara in Lau. DistRIBUTION: Endemic to Fiji, and thus far known from several islands, apparently being more widespread than var. vitiensis. This va- niety has been recorded as an often slender or spreading tree 3-10 m. high, occurring from sea-level to 970 m. in various types of forest, some- times on limestone. The corolla is noted as white and the fruit as green- ish yellow; flowers have been obtained in March and fruits more or less throughout the year, Loca NAMEs: Kaukauloa, kailoa, mbama, mulu, sawira. Fiji. YAsawas: Waya IsLanp: North of Yalombi, along Olo Creek, St. John 18121 (A, BISH). Vitr Levu: M (A, BISH, K, NY, UC), 1253 (a, BISH, K, NY, U Agr. 14810 (Bis, suva), 14812 (BISH, suva); Nandendeleva, Fiji Dept. Agr. 14851 (BISH, k, MASS, § 1971] SMITH, PACIFIC ISLAND PLANTS, XXIII 403 Webster & Hildreth 14255 (DAv, MASS); southern slopes of Mt. Ndelainathovu, on escarpment west of Nandarivatu, Smith 4943 (a, BISH, K, us); slopes of Mt. Lomalangi (Nanggaranambuluta), near Nandarivatu, alt. 950 m , Dec. 20, 1927, Gillespie 4360 (BisH holotype); hills between Nandala and Nukunoku Creeks, along trail from Nandarivatu iii barns Smith 6200 (A, BISH, K, NY, us); Nandala, Degener 14741 (A, BISH, K, us). NANDRONGA & Navosa: Horne 904 (k); Nausori Highlands, Fiji Desk ap: 12649 (a, K, suva). Nat- TASIRI: Nanggarathangithangi, Mendrausuthu Range, Fiji Dept. Agr, 15028 (mass, SUVA). OVALAU: Hills west of Lovoni Valley, on ridge south of Mt. Korolevu, Smith 7631 (BISH, GH, K, SUVA, UC, US); Lovoni Valley, Horne 195 (cu, K), Fiji Dept. Agr. 14504 (stsH, suva). Ncau: Hills east of Herald Bay, inland from Sawaieke, Smith 7775 (BISH, GH, K, SUVA, UC, US). pele: Levu: Msvua: Upper Ndama River valley, Smith 1608 (BISH, GH, K, NY, US KamparA: On limestone formation, in forest, alt. 0-100 m., Smith ai (BISH holotype of Maba globosa; isotypes at GH, K, NY, UC, US); same locality, Smith 1272 (BISH, GH, K, NY, UC, US). CULTIVATED SPECIES 8. Diospyros discolor Willd. Sp. Pl. 4(2) : 1108. 1806; Hiern in Trans. Cambr. Philos. Soc. 12: 260. 1873; Bakh. in Bull. Jard. Bot. Buitenz. TIT. 15: 145. 1938. Samoa, without further locality: Guest 5 (BISH). The cultivated occurrence of D. discolor in our area is indicated by a single sterile specimen. The species, placed by Bakhuizen in his subgen. Euptospyros sect. EBENASTER, is endemic to the Philippines, now widely cultivated for its edible fruit, usually with the common name madolo. REFERENCES 1, BAKHUIZEN VAN DEN BRI _C. Enumeration of Malayan Ebenaceae. Gard. Bull. Straits porns 7: 161-189. 1933. . Revisio Ebenacearum Malayensium. Bull. Jard. Bot. Buitenz. III. 15: 1-49, 1936; 50-368. 1938; 369- 515. 1941; i-xx, pl. 1-92. eee 3. Fosperc, F. R. Notes on plants of the Pacific Islands. J. Bull. Torrey Bo Club 65: 607-614. 1939. . Notes on plants of the Pacific Islands — IL Ibid. 67: 417-425. s: I, Kew Bull. 23: 337- 1940 5. GREEN, P. . a al.). Notes on Melanesian plant 346, 6. Hiern oe oy A ait ginbite of Ebenaceae. Trans. Cambr. Philos. Soc. 12; 27-300. 18 DEPARTMENT OF BOTANY UNIveRsITy of MASSACHUSETTS AMHERsT, MassAcuusettTs 01002 404 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 THE GENERA OF OROBANCHACEAE IN THE SOUTHEASTERN UNITED STATES * Joun W. THIERET OROBANCHACEAE Ventenat, Tabl. Régne Vég. 2: 292. June 1799, “Orobanchoideae,” nom. cons. (BROOMRAPE FAMILY) Annual or perennial, root-parasitic, achlorophyllous herbs, often fleshy, with 1 to several flowering shoots arising from a more or less thickened base,” pubescent with glandular or sometimes also nonglandular hairs. Leaves reduced to scales, alternate, the upper passing imperceptibly into bracts. Inflorescence racemose or spicate, or flowers rarely solitary. Flow- ers perfect, zygomorphic, hypogynous, axillary to bracts, with or without bractlets, long-pedicelled to sessile. Calyx persistent, synsepalous, nearly actinomorphic to strongly zygomorphic, with 1-5 lobes or teeth. Corolla sympetalous, tubular, the limb mostly 2-lipped, the upper lip 2-lobed to entire, the lower usually 3-lobed. Androecium of 4 didynamous stamens inserted on the tube of the corolla and alternate with its lobes; filaments elongate to very short, included or exserted; anthers dorsifixed, with pla- centoids, 4-sporangiate, bilocular at maturity, the locules dehiscing lon- *Prepared for a generic flora of the southeastern United States, a joint project * the Arnold Arboretum and the Gray Herbarium of Harvard University made pos- sible through the support of the National Science Foundation (Grant GB-6459X, given in this paper b I am indebted to Dr. Wood for his careful review of the manuscript and for other aid always so unstintingly gi : y given; to Dr. B. arts of the manuscript; to Dr. ¥. D. Tina’ r. B. L. Burtt for gag on Pp peci i Virginia Sava Devore and D ge (Ficure 1), and by Sydney B. r poten oe eran g 2) from dissections by Dr. Wood. Specimens for the i r. Wood and De x ol the Generic Flora of the Southeastern United States by *I am aware of A. Wilson, Mr. R. J. Eaton, and myself. base of the plant : cho ‘ame (to me) morphological term to apply to the swollen of Conopholis For te pac Orobanche or to that — seemingly quite different — hocks ° Tms such as ‘tuber,” “tubercle,” and “gall,” my enthusiasm knows 1971] THIERET, GENERA OF OROBANCHACEAE 405 gitudinally; pollen triaperturate or rarely nonaperturate. Gynoecium syn- carpous, the carpels 2 or rarely 3; stigma simple, variable, often capitate, crateriform, or bilamellate; style single, elongate, included or exserted; ovary usually unilocular, with 4 or rarely 5 or 6 [2 or 3] parietal pla- centae; ovules many, unitegmic, tenuinucellar, mostly anatropous; em- bryo sac development normal (Polygonum type); endosperm ab initio cellular, with terminal haustoria. Fruit a 2- or rarely 3-valved capsule, typically dehiscing anteroposteriorly, the valves usually lateral, each bearing 2 [1] placentae; seeds numerous, endospermous, with a minute undifferentiated embryo. Type GeNus: Orobanche L. The Orobanchaceae comprise about 150 species in 17 genera (14 rec- ognized by Beck-Mannagetta; three — Mannagettaea, Necranthus, and Tienmuia — described since). In the conterminous United States four genera occur; three of these, represented by four species, are found in the Southeast. A majority of the genera and about 90 per cent of the species of Oro- banchaceae are native to the Old World. Conopholis, Epifagus, and Kopsiopsis are confined to the New World; Boschniakia and Orobanche are native in both hemispheres. The family is primarily one of the North Temperate and warm zones. Only about 10 per cent of the species occur in the tropics; only one species reaches the arctic. Orobanchaceae are one of only about a dozen families of angiosperms in which the parasitic habit is known. They are the largest family of holoparasitic flowering plants. The range of hosts is large, including both woody and, more commonly, herbaceous species; both gymnosperms (rare- ly) and angiosperms; monocots and dicots. Reports of Orobanchaceae on ferns require verification. Some Orobanchaceae are confined to a single host species or genus; at the other extreme, some parasitize many species and genera, especially Compositae, Labiatae, and Leguminosa One earlier worker (Wiesner) claimed to have detected chlorophyll in Orobanche, but modern studies, utilizing electron microscopy (Laudi; Laudi & Albertini) and chromatography (Baccarini & Melandri; cf. Oro- banche) found this pigment to be completely lacking in Orobanche and Lathraea. There seems little doubt that Orobanchaceae are achlorophyl- ous. Orobanchaceae germinate only in he host plant —a phe- However, the need for where the cells are undergoing The substances may diffuse into the It would appear that many Orobanche seeds within this zone of diffusion 406 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 are stimulated to germinate but are unable to form a parasitic attach- ment because their radicles usually do not exceed 2 mm. in length. The “axiom” that host-stimulants are always required for germination of Oro- banche seeds was recently called into question by Krenner, who reported that fresh seeds of O. cumana germinated in plain water in fair numbers. After overwintering, however, the seeds germinated only in the presence of host root exudates. The embryo of members of the Orobanchaceae is quite undifferentiated and may consist of as few as 45 cells. Its proximal end gives rise to the radicle. After germination of the seed and subsequent penetration of the host by the primary haustorium, the part of the seedling outside the host root enlarges to form the small, so-called “tubercle.” At the apex of the “tubercle,” which corresponds to the distal end of the embryo, is the plumule. From the proximal part of the “tubercle” secondary roots may arise. Continued enlargement of the “tubercle” — often incorporating tissues of the host, which are stimulated to abnormal growth — may re- sult in a large, rather regularly shaped “tuber” (e.g., Epifagus) or in a large, misshapen “tuber” (e.g., the “gall” of Boschniakia and Conopholis). The “tubercle” in some species enlarges but little, the bulk of the sub- terranean part of the plant being secondary roots. Parasitic haustoria in Orobanchaceae are of two kinds. The primary haustorium, terminal in origin, develops from direct transformation of rk parasite being connected by a slender strand of xylem that is distal y uregular or even double (Kuijt). In Orobanche, many of the T cells of the primary haustorium are said to be multinucleate. rari ao of Orobanche are exogenous, but “they are unmistakably (Kuijt). In init those of Epifagus are endogenous. Conoph- oots cn ou ah Accordingly, each of the resulting four placentae 1s suggested by B fn one carpel. That Orobanche has four carpels was first of the traditibend lagi & Sankhla, then denied by Guédés in a reaffirmation attd-elabeaiien interpretation of the gynoecium, and finally defended t a Y. D. Tiagi. According to Tiagi, Orobanche has on median and two lateral; each carpel bears a placenta gin. The marginal placentae of adjacent carpels are fused, De, i C onopholis, FAA-preserved speci : 0 j i : Synoecium of Epifagus is basically identical to that of Conopholis and 1971] THIERET, GENERA OF OROBANCHACEAE 407 Orobanche and so would certainly be regarded by Tiagi as four carpel- late also. Two lines of gynoecial evolution in the Orobanchaceae are recognized by Y. D. Tiagi. In one line, exemplified by Boschniakia, the number of carpels is reduced from four to two. In the other line, exemplified by Aeginetia, Cistanche, and Orobanche, all four carpels are retained, but the lateral ones are smaller and are devoid of midrib bundles. It is un- fortunate that Tiagi’s scheme does not account for those species of Oro- banchaceae (the “Orobanchaceae tricarpellatae” of Beck-Mannagetta) that have six placentae. Are these six-carpellate? The relationships of Orobanchaceae are controversial, Three points of view are held: the broomrapes are most closely related to the (1) Scrophulariaceae, (2) ‘Solanaceae and allied taxa” (Y. D. Tiagi), and (3) Gesneriaceae. Also questioned is the maintenance of Orobanchaceae as a separate family. The majority of workers favors a close relationship between Orobancha- ceae and Scrophulariaceae. Boeshore concluded that these taxa form a continuous morphological and physiological series “from non-parasitic through semi-parasitic Scrophulariaceae to the most degraded parasites of the family [e.g., Striga orobanchoides, Harveya, Hyobanche|, and that these again show direct continuity with the still more degraded and con- densedly parasitic types of Orobanchaceae” and that the Orobanchaceae and Scrophulariaceae, “alike logically and biologically . . . should be treated in continuous descending series from the highest to the most de- graded genera.” After a study of the stomatal apparatus of various heterotrophic flow- ering plants, Linsbauer & Ziegenspeck concluded that Orobanche is the culmination of an evolutionary series beginning with nutrient-salt para- sites, such as Striga, among the Scrophulariaceae-Gerardieae | Buchne- reae|. On embryological grounds, relationship between Orobanchaceae and Scrophulariaceae was first suggested by Schnarf. Several years later Glisié pointed out that the two families are so similar embryologically that they could well be united. More recently, B. Tiagi (1963) concluded, again on the basis of embryology, that the Orobanchaceae are “derived from the family Scrophulariaceae (Rhinanthoideae) through forms like Striga orobanchoides.” : ws In anther structure, certain Orobanchaceae and certain parasitic Scrophulariaceae are similar in that one longitudinal half of the anther has become reduced or otherwise modified. Some of these genera were placed by Livera in his proposed family Aeginetiaceae, distinguished mainly by characteristics of the anthers: “Of the two anther cells [i.e., halves] one only is fertile, the other either wanting . . . or in the form of ee en Kuijt suggested that, on the basis of anther evolution and “other evolu- tionary trends, there appear to be no serious objections to the possibility that Christisonia, and perhaps even Aeginetia [both of the Orobanchaceae | ; are derivatives of Harveya | Scrophulariaceae]. Should this conception be 408 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 52 valid, the familial status of Orobanchaceae would be further undermined.” Nikiticheva found the similar development of microspores and pollen in Orobanchaceae and in certain parasitic Scrophulariaceae particularly strik- ing. Gynoecial differences are the traditional ones separating Orobancha- ceae and Scrophulariaceae, the former having a unilocular ovary with parietal placentae, the latter, a bilocular ovary with axile placentae. Lo- cule number is, however, not an infallible characteristic. Species of Chris- tisonia (including Campbellia) of the Orobanchaceae may have either a unilocular or bilocular ovary; in at least one species the ovary is bilocular below and unilocular above. Among the Scrophulariaceae a few genera (Dopatrium, Limosella, Micranthemum) have unilocular or imperfectly bilocular ovaries. Members of Lathraea, a genus that has been shifted back and forth between Orobanchaceae and Scrophulariaceae, have uni- locular ovaries. Arekal reported a tendency toward unilocularity in the distal portion of the ovary of the scrophulariaceous genera Orthocarpus and Awreolaria (“Gerardia”) and concluded that a clear transition from the two-celled [two-locular] ovary of most Scrophulariaceae to the one- celled [one-locular] ovary of Orobanchaceae exists. Levyns found the ovary of Hyobanche, a genus of South African Scrophulariaceae, to be imperfectly bilocular, with axile placentation below and parietal above; considering that this “discovery . . . breaks down the principal character Separating the families,” she included Orobanchaceae in Scrophulariaceae. 4 two families had been combined even earlier by Bellini and by Hal- That placentation in Orobanchaceae and Scrophulariaceae is funda- arc ey sdb ailing questioned by Y. D. Tiagi, who concluded lonsitatih aud car ay acentation is anatomically the same. In Veronica t Scrophulariaceae studied by him, he found the pla- ce S a gap could be seen between the placentae. Tiagi ae that placentation in both Orobanchaceae and Scrophularia- ceae is parieta Ac i : fede er slid Tiagi, “the occurrence of a fundamentally tetra- .» Synoecium in the family Orobanchaceae must rule out any Scrophulariaceae can be more logically lel evolution rather than indicative of rela- the ancestors of the Orobanchaceae may be better anaceae and allied taxa where the multicarpellary €clum is quite often met with.” * The Scro i formly ilaspae ae Rica Ds a considered as examples of aral tionships. Search for : made among the Sol condition of the gyno be better characterized as “almost” uni- pelary: Phillips (Gen. So, Af genus Bowkeria apparently can be tricar- - So. Afr. Fl, . i bered” and the capsule se Poy esp vid describes the ovary as “2-3 cham 1971] THIERET, GENERA OF OROBANCHACEAE 409 The Orobanchaceae, because of their typically unilocular ovary, have been included by some authors (e.g., Baillon, Eichler, and Warming) as a parasitic offshoot of the Gesneriaceae. The idea of close relationship between these two families has been revived by Crété, who asserted that embryologically the Orobanchaceae and Scrophulariaceae do not have “direct affinities” but that the broomrapes are so closely related to the Gesneriaceae that they could well be united with them. The Orobanchaceae seem especially similar to Gesneriaceae in their accumulation of oro- banchin, in their lack of aucubin-like glucosides, and in their high silicic acid content (Hegnauer). The biology of most Orobanchaceae — their ecology, morphology, de- velopment, host relationships, longevity, pollination, dispersal, etc. — is not well known. Even such basic information as host plant is seldom de- termined and noted by collectors. Orobanchaceae, because of discolora- tion and brittleness, usually make rather unpleasing exsiccata. The dis- coloration, at least in Orobanche, has been attributed to the oxidation of orobanchoside, a glucoside. Cleistogamy in the family appears to be a regular feature only in Epifagus although subterranean cleistogamous flowers have been reported in Cistanche. Peloric flowers are known to occur in Orobanche. Orobanchaceae, especially Aeginetia and Orobanche but also Christi- sonia, are of economic importance primarily as parasites of various crop plants. A few species are used in folk medicine. REFERENCES: Acatt, J. A., & J. P. Tan. Controlling the Aeginetia indica in cane fields. Sugar News 12: 852-854. 1931a. _ The effect of Atlacide on Aeginetia indica. Ibid. 82-89. 1931b. ArEKAL, G. D. Embryological studies in Canadian representatives of the tribe Rhinantheae, Scrophulariaceae. Canad. Jour. Bot. 41: 267-302. 1963. BAILLoN, H. Gesnériacées. Hist. Pl. 10: 59-112. 1888. [Orobanchaceae, as a “série” of Gesneriaceae, 73-77, 108-112.] BECK von MANNAGETTA, G. Orobanchaceae. Nat. Pflanzenfam. IV. 3b: 123- £325:1393, Orobanchaceae. Pflanzenreich IV. 261(Heft 96): thor’s name given as G. BEcK-MANNAGETTA. } BELLINI, R. Criteri per una nuova classificazione delle Personatae (Scrophu- lariaceae et Rhinantaceae). Ann. Bot. Roma 6: 131-145. 1907. [Includes Orobanchaceae in Scrophulariaceae. ] , BENTHAM, G., & J. D. Hooke Orobanchaceae. Gen. PI. 2: 980-986. 1876. Borsnore, I. The morphological continuity of Scrophulariaceae and Oro- banchaceae. Contr. Bot. Lab. Univ. Penn. 5: 139-177. 1920 Borsster, E. Orobanchaceae. Fl. Orientalis 4: 492-518. 1879. Cézarp, R. Germination in-vitro de quelques Orobanchacees en présence des racines de différentes plantes. Bull. Ecole Nat. Super. Agron. Nancy ef 64-69. 1965a.* ee saat . Orobanchacées: germinations observees in-vitro en présence de stances naturelles, 1. Présence d’agents stimulants dans les liquides de culture de Centaurea Scabiosa L. [bid. 153-168. 1965b.* 1-348. 1930. [Au- 410 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 Cuatin, A. Anatomie comparée des végétaux, plantes parasites. Vol. 1. 560 pp.: vol. 110 pls. Paris. 1892. [Orobanchaceae, 42-130, pls. 7-19. CreETE, P. L’application de certaines données embryologiques a la systématique des Orobanchacées et de quelques familles voisines. Phytomorphology 5: 422-435. 1955, Davis, G. L. Systematic embryology of the angiosperms. x + 528 pp. New York. 1966. [Orobanchaceae, 196, 197; references. | Don, G. Orobanchaceae. Gen. Syst. Garden Bot. 4: 627-636. 1837. ErcuHier, A. W. Bliithendiagramme. Vol. 1. 347 pp. Leipzig. 1875. [Oro- banchaceae, 220, 221, a “Gruppe von Gesneraceae.” ErpDTMAN, G. Pollen morphology and plant taxonomy. Angiosperms. 539 pp. Waltham, Mass. 1952. [Orobanchaceae, 301, 302. | Farnswortu, N. R., R. N, Biomster, M. Y. Quimpy, & J. W. SCHERMERHORN. The Lynn Index. A bibliography of phytochemistry. Monograph VI. Order, Tubiflorae. 274 pp. Pittsburgh. 1969. [Orobanchaceae, 165-167.] FLEISCHER, E. Zur Biologie feilspanformiger Samen. Bot. Arch. 26: 86-132. my t. R. . LATTER. Observations on the pollen development of two species of Lathraea. Jour. Roy. Microscop. Soc. III. 47: 280-325. 1927. Gur, A. Necranthus: a new genus of Orobanchaceae from Turkey. Notes 968. Guisic, L. M. Uber die Endosperm- und Haustorienbildung bei Orobanche he- derae Duby und Orobanche gracilis Sm. (Zugleich ein Beitrag zur Phylo- genie der Orobanchaceae). Bull. Inst. Jard. Bot. Univ. Belgrade 1: 106- . 1929 GRANEL DE SoLicnac, L. Contribution critique 4 la connaissance de Vhistoire _Systematique des Orobanchacées. Nat. Monspel. Bot. 20: 119-134. 1969. GUEbEs, M. Remarques sur la placentation des Orobanchacées. Bull. Soc. Bot. France 111: 257-261. 1965. [Reaffirms the two-carpellary interpretation < ee chromosome counts in Orobanchaceae. Ibid. 177: 438, 439. . hromosome numbers in some members of the family Orobanchaceae. Jour. Linn. Soc. Bot. 55: 772-777. 1958. Harri, D. Das Placentoid der Pollensicke, ein Merkmal der Tubifloren. Ber. Deutsch. Bot. Ges. 76: 71, 72, 1963. es G. Orobanchaceae. Illus. Fl. Mittel-Europa 6(1): 132-155. [1914.] ae R. Chemotaxonomie der Pflanzen. Band 5. Dicotyledoneae: Mag- mop 506 pp. Basel & Stuttgart. 1969, [Orobanchaceae iaber ersbeis R. Recherches sur l’appareil végétatif des Bignoniacées, Rhinan- ‘acees, Orobanchées et Utriculariées. 765 pp. Paris. 1888. [Orobancha- ior long and detailed abstract in Bot. Jahrb. 11(Lit-ber.): 65- 1971] THIERET, GENERA OF OROBANCHACEAE 411 Caractéres anatomiques généraux des organes végétatifs des Rhinan- thacées et des Orobanchées. (Abstr.) Bot. Jahrb. 14(Lit-ber.): 42. 1892. [ Original tee said to be in Bulletin de la Société d’Etudes Scientifiques de Paris, ] Hu, H. H. On me genus Gleadovia in China. Sunyatsenia 4: 1-9, . Tienmuia, a new genus of Orobanchaceae of aaa China Bull. Fan Mem. Inst. Biol. Bot. 9: 5-7. 1939b. Jvuttano, J. B. Anatomy and morphology 8 the bunga, Aeginetia indica Lin- naeus. Philip. Jour. Sci. 56: 405-451. 1935. Kine, L. J. Weeds of the world. Biology and control. 526 pp. London & New York. 1966. [Orobanchaceae, 61-64. Knutu, P. Handbook of flower pollination. Vol. 3. (Transl. J. R. A. Davis.) iv + 644 pp. Oxford. 1909. [Orobanchaceae, 232, 233.] KOHLER, E. Parasitische Samenpflanzen. Pp. 866-897 in P. SoRAUER, Hand- buch der Pflanzenkrankheiten, vol. 3. 948 pp. Berlin. 1932. [Orobancha- ceae, 893-897. ] Kurjr, J. The biology of parasitic aes plants. 246 pp. Berkeley. 1969. [Orobanchaceae, 81-103, 181-184, passim.] Laupr, G. Ricerche infrastrutturali sui olastidi delle piante parassite. II. Eathriea squamaria. Caryologia 19: 47-54. 1966. [No chlorophyll de- | tected . ALBERTINI. Ricerche infrastrutturali sui plastidi delle piante parassite. III. Orobanche ramosa. Ibid. 20; 207-216. 1967. [No chlorophyll ected. | mi Ler, A., & F. Goseco. Studies of the sugar-cane root parasite, Aeginetia indica. Int. Soc. Sugar Cane Technol. Bull. 101: 1-12. 1932. [Includes data on insect parasites (moth larvae in seed capsules); bound with Proc. Int. Soc. Sugar Cane Technol., vol. 4.] Levyns, M. R. Scrophulariaceae. In: ApaMson & T. M. Satter, Fi. Cape Peninsula. xix + 889 pp. ay ae 705-721; includes Orobanchaceae. | sh tee Linc, K. C. Bunga. Taiwan Sugar 2(4): 21. 1955. [Aeginetia indica. ] LinsBauer, K., & H. ZIEGENSPECK. Das Vorkommen von Spaltéfinungen bei heterotrophen Bliitenpflanzen im lichte der Physiologie und Stammes- geschichte. Biol. Gen. 17: 511-565. 1943. [Orobanchaceae, 554-557.] Livers, E. J. Aeginetiaceae, a new natural family of flowering plants. Ann. Roy. Bot. Gard. Peradeniya 10: 145~159. 1927. Lo, T. T. N:Co 310 highly resistant to the root parasite bunga. Taiwan Sugar 2(4): 18-20. 1955. [Sugarcane and naib indica. | Lopez, M. E. Aeginetia indica — Roxb., a phaner rogamic parasite of — Philip: Sugar tas Quart. 9(2): 51-61, 1963. [Taxonomy, biology, — Martin, A. C. The comparative internal ponte of seeds. Am. Mi a 36: 513-660. 1946. [Orobanchaceae, Mercatre, C. . : asia Anat. Dicot. 2: 725-1500. 1950. [ Orobancha- ceae, 988-991. NixirtcHeva, Zi z Anther development and microsporogenesis in oe sentatives of Scrophulariaceae oe (In Russian; Eng summa Bot. Zhur. 53: 1704-171 Ozenpa, P. wieoseeain sur les phanérogames parasites. I. ue 0 “aes : récents. Phytomorphology 15: 311-338. 1965. [Orobanchaceae, ux 412 JOURNAL OF THE ARNOLD ARBORETUM [vox 52 PEDERSEN, A. Scrophulariaceernes og Orobanchaceernes udbredelse i Danmark Bot. Tidsskr. 59(1, 2): 1-140. 1963. [Distribution maps. PetcH, T. Campbellia aurantiaca, Wight, and Christisonia albida, Thwaites. nn. Bot. 38: 679-697. 1924. ; wei a cytinoides Wight. Ann. Roy. Bot. Gard. Peradeniya 11: 269-275 QuisumBING, “a ‘On Christisonia Wrightii Elmer, a parasite of sugar cane. Philip. Jour. Agr. 11: 397-401. 1940.* Rancan, T. S., & N. S. Rancaswamy. Mo hogenic investigations on para- sitic angiosperms I. Cistanche tubulosa (Orobanchaceae). Canad. Jour. Bot. 46: 263-266. 1969 Rao, T. S. Pollen morphology of two species of Orobanchaceae. Curr. Sci. Bangalore 32: 557, 558. 1963. [Orobanche nicotianae, Aeginetia indica. | Raun, W. Die Bildung von Hypokotyl- und Wurzelsprossen und ihre Bedeu- tung fiir die Wuchsformen der Pflanzen. Nova Acta Leop. 4: 395-553. 1937. [Orobanchaceae, 410-416. ] Razt, B. A. An annotated list of phanerogamic parasites oar a and Pak- istan. Lloydia 20: 238-254. 1957, parry ns Reuter, G. F. Orobanchaceae. DC. Prodr. 11: 1-45. 1847 Ross- CRAIG, S. Drawings of British plants. foci Scropihisladiceal (2), Orobanchaceae, <5 Sabra Verbenaceae. 38 pls. London. 1966. [Oro- banchaceae, pls. 2 30.] Roxas, M. L. age indica on sugar cane. Sugar News 12: 89-91. 1931. SCHMUCKER, T. Hohere Parasiten. Pp. 480-529 in W. RuuLanp, Handbuch fa _Plancenphsiloi vol. 11. 1033 pp. Berlin. 1959. [Orobanchaceae, Scunarr, K. Zur Entwicklungsgeschichte von Plantago media. Sitz-ber. Akad. Wiss. Math.-Naturw. Wien. Abt. 1. : 927-950. 1917. [Suggests rela- tionship, on etubiyologies) grounds, between Orobanchaceae and Scrophu- Sen, A. B., & S. P. SivcH. Chemical examination of Christisonia bicolor Gardn. Jour. Titian Chem. Soc. 41: 228. 1964. » S. Buarracuaryr, & S. P. StncH. Chemical examination of Christi- sonia bicolor Gardn. Ibid. 40: 925-928. 1963. SmitH, H. Plantae Sinenses. XXVI. Orobanchaceae. Acta Horti Gothoburg. : 8: 127-146. 1933. [Mannagettaea H. Smith, gen. nov., 135-139.] OLMS-LauBacH, H. De Lathra raeae generis positione systematica. Dissertation, Friedrich Wilhelm Univ. 42 pp. Berlin. [1865.] SRINIVASAN, A. R. & C. L. SUBRAMANIAN. A review of literature on the pha- nerogamous italia: 96 pp. Indian Council of Agricultural Research, New elhi. 1960. Perera 2-15; many references. | > N. re of the seeds of angiospermous a _ py dys in 7 L. Harper, ed., The biology of weeds. xv + 2 . Ox- ord. 1960a. [Chiefly Proleuche and Striga. | Petace eee of the Striga and Orobanche germination arc by € number and Exper. Bot. 11: 236-245. 1960 and variety of stimulants. Jour. Exp ———.. The production of the Striga and Orobanche es stimulants by Pike: ize roots, . I. Conditions of synthesis in the root. Jbid. 356-366. 1960c. since E. S. Embryology of Orobanchaceae. I. — pai (M.B.) Poir. (in Russian.) Bot. Zhur. 51: 1601-1607. 1971] THIERET, GENERA OF OROBANCHACEAE 413 THIERET, J. W. riprrreese: In: C. L. Lunpett, Fl. Texas 2: 331-337. pls. 26-28. 1970 (“19 Tract, B. Studies in the aoe Orobanchaceae. III. A contribution to the em- bryo logy of Orobanche cernua Loeffl. and O. aegyptiaca Pers. Phyto- morphology 1: 158-169. 1951. ———. Studies in the family Orobanchaceae. I. A contribution to the embryol- ogy of Cistanche tubulosa Wight. Lloydia 15: 129-148. 1952a. Studies in the family Orobanchaceae. II. A contribution to the em- bryology of Aeginetia indica Linn. Bull. Torrey Bot. Club 79: 63-78. 52b. —. Studies in the family Orobanchaceae. IV. Embryology of Boschniackia himalaica Hook. and B. tuberosa (Hook.) Jepson, with remarks on the evolution of the eile. Bot. Not. 116: 81-93. 1963. . Studies in the family Orobanchaceae. VI. Development of the seed in Conopholis americana (L. fil.) Wallr. Acta Bot. Acad. Sci. Hungar. 11(1/2): 253-261. & N. S. SANKHLA. Studies i in the family Orobanchaceae. V. A contribu- tion to the embryology of Orobanche lucorum, Bot. Mag. Tokyo 76: 81- 88. 1963. Tract, Y. D. Anatomical studies of the vascular equipment of the flower of some species of the families Orobanchaceae and Scrophulariaceae. (In Russian.) Vestnik Moscow Univ. 6. Biol. 2: 29-52. 1962. Trasut, M. L. Fleurs cléistogames et souterraines chez fe Orobanchées. Bull. Soc. Bot. France 33: 536-538. 1886. [Cistanche lutea with subterranean cleistogamous flowers. | VaLLey, K. R., & T. S. Cooperriper. The Orobanchaceae of Ohio. Ohio Jour. Sci. 66: 264, 265. 1966. WaL.rortn, F. W. Orobanches generis diaskeue. 80 pp. Francofurti ad Moenum. 1825. Warmtne, E. A handbook of systematic botany. 620 pp. London. 1895. [“Oro- banche (Broom-rape) is allied to [Gesneriaceae] as a parasitic form,” 528. Wiesner, J. Untersuchungen iiber die Farbstoffe einiger fur pear asain gehaltenen Phanerogamen. Jahrb. Wiss. Bot. 8: 575-594. 1872. [Claime to have detected chlorophyll in Orobanche. ] cs WorspELt, W. C. On the comparative anatomy of certain species 0 risti- sonia, Ann. Bot. 9: 103-136. 1895. a . On the development of the ovule of Christisonia, a genus of the Oro- banchaceae. Jour. Linn. Soc. Bot. 31: 576-584. 1897. Key To THE GENERA OF OROBANCHACEAE IN THE SOUTHEASTERN UNITED STATES General characteristics: root parasites lacking chlorophyll; leaves scale- ey crowded to clearly alternate; flowers yg zygomor phic ; - g wh epipetalous; ovary superior, 1-locular, with 4 parietal placentae, each bearing many anatropous ovules; fruit capsular ; embryo minute, undifferentia A. Flowers in large part with calyptriform corollas, es sometimes a distal few tubular and chasmogamous. .....-------- "7" "°° ; A. Flowers all with tubular corollas, chasmogam B. Flowers bibracteolate; calyx split sacra serted. ‘spathaceous”’; pmo S ex- 2. Conopholis. 414 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 B. Flowers ebracteolate; calyx split anteroposteriorly or campanulate and i d 3-0 5-cleft to 5-parted; stamens included. ................ . Orobanche. 1. Epifagus Nuttall, Gen. N. Am. Pl. 2: 60. 1818, nom. cons.4 Low, puberulent or glandular-puberulent, yellow, black-purple, purple- brown, yellow-brown, or brown, simple or branched annuals. Base 0 plant an enlarged “tuber” bearing leaf-scales and thick, short, coralloid, usually branched secondary roots. Stem with scattered, alternate leaf- scales. Inflorescence racemose, each flower bibracteolate, short pedi- celled. Calyx obliquely broadly campanulate, nearly equally 5-toothed to 5-lobed. Cleistogamous flowers: usually fertile; corolla calyptriform, per- sistent for a time on the enlarging capsule, 4-lobed, the upper lobe inter- nal, entire to slightly notched; stamens included, the filaments very short, the anthers at first free but becoming hard and firmly adnate to the stigma, anther halves somewhat divergent, mucronulate basally, those of adjacent stamens becoming connate; style declined anteriorly, very short, included; stigma capitate, slightly bilobed; style, stigma, corolla, and stamens deciduous as a unit, Chasmogamous flowers: usually sterile; co- often widening distally, soon deciduous, 4-lobed, the upper lobe internal, erect, rounded, entire or slightly notched, the lower lobes erect to some- what spreading, acute; stamens included to slightly exserted, the fila- ments elongate, anthers free, anther halves somewhat divergent, mucro- nulate basally; style elongate, soon deciduous, usually exserted; stigma capitate, slightly bilobed. Capsule somewhat compressed laterally, thick- est below the middle, more or less reniform in lateral view but with a truncate base, 2-valved or rarely 3-valved, dehiscing anteroposteriorly. Seeds minute, very numerous, narrowly ovoid to ellipsoid; testa finely reticulate. (Leptamnium Raf.) Type species: E. americanus Nutt., Aah : ris i E, virginiana (L.) Bart. (Orobanche virginiana jh om Greek, epi, upon, and Latin, fagus, beech, in allusion to the host plant.) — BEEcHDROPs, CANCER-ROOT. A genus of one Nor (Leptamnium virginianum io.) : Raf), beechdrops, ranges from Cape (Fagus grandifolia var. grandifolia), and ‘a Var. mexicana), are more or less identical, been found in Oklahoma or in Puebla. t reported from Mexico until 1939; Epifagus, les share, with a number of others, a charac- 4 c . onservation Superfluous ; see Rickett & Stafleu, Taxon 9: 113. 1960. 1971] THIERET, GENERA OF OROBANCHACEAE 415 Ficure 1. Epifagus. a-k, E. virginiana: a, plan t with fruits, cleistogamous i ea and two cleistogamous partly diated fruit, the corolla already fallen, X 1 X 5; e, same, with alf of c alyx. and corolla and two © stamens removed — note ovary, showing | 4 corolla being pushed o by ge of ovary, X 5; h, same, in vertical section — note developing seeds, X 5 early mature fruit seen from abaxial side, X 5; j, open capsule with seeds, ‘atlasial side r left, X 5; k, two seeds, X 40. 416 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 teristic bicentric distribution pattern: eastern United States and the high- lands of eastern Mexico. The distribution of these may have been con- tinuous during the early Pleistocene, a controversial matter. Decreased rainfall in Texas and northeastern Mexico later may have brought about the present disjunction. Although it has been suggested that Epifagus is saprophytic (Hill) or that it is “self-sustaining after an initial period of parasitism” (Meehan), ‘ made with beeches of all ages (even one-year-old seedlings), is effected by means of a primary haustorium that pierces the beech root (Brooks). Earlier workers (Cooke & Schively) had reported that the organ of con- nection between beech and beechdrops arises from the beech and grows into the tissues of the parasite. As the beechdrops tuber grows, its tis- sues develop around the beech root, which eventually becomes buried therein. No secondary haustoria are formed. Beechdrops is usually purple-brown, but variant color forms occur; The mature plant is differentiated into two regions: an enlarged base, the “tuber,” and an aérial “stem,” the inflorescence. The “tuber” is considered by Boeshore to represent a “fused primary root below and a is above.” The curious, short, branch- t may be simple. All branches are axillary to scale-leaves. Some additional branches may arise from the apex of the “tuber.’ Stomata are abundant and wides st a 1s 8 . lud- ing the style. The rapid wiltin pread on the aérial parts, even inc 8 of beechdrops after it is severed from due largely to excessive transpiration through Cleistogamous flowers ar humerous than chasmoga ground. Some plants, or e the first to appear and are usually more mous ones. They may even develop under- €ven whole colonies, may bear no chasmogamous 1971] THIERET, GENERA OF OROBANCHACEAE 417 Map 1. Documented distribution of Epifagus virginiana. Inset shows north- eastern Mexico and southern Texas. For distribution of Fagus grandifolia see Jour. Arnold Arb. 52: 169. 1971. flowers. When produced, these flowers are proterogynous and are distal ducing cleistogamous flowers. Above firmly connate to each other an apparently is caused, at least in par grow through the anther wall into the 418 JOURNAL OF THE ARNOLD ARBORETUM [voL, 52 together. More pollen grains per anther are produced by the cleistogamous flowers than by the chasmogamous ones. Nothing is known of pollination in Epifagus. The presence of a nec- tary (in, curiously enough, both kinds of flowers) suggests insect pollina- tion, but the lack of fruit from most chasmogamous flowers may indicate that the cross-pollination possibly required for fruit development is at best infrequent. The means of dispersal of Epifagus seeds is not known. When the cap- Sule opens, the minute seeds are at first firmly attached by the long funiculi, Even though soon free, they tend to cohere, thus reducing the effectiveness of wind dispersal, The seeds can be dislodged from the cap- sule by drops of water that fall into its opening, which faces upward and is fully exposed to the impact of raindrops. When the seeds are shed, the embryo is several celled and undiffer- entiated, but by the following spring it appears as a “many-celled body enlarged at one end and tapered abruptly into a nearly filamentous struc- ture at the opposite end” (Brooks). According to my field observa- tions (in Indiana and Louisiana), germination takes place usually in whitish “tubercles” 1/16 to 1/4 inch in diameter. Such tubercles ap- parently can be found in fall (Schrenk), suggesting that beechdrops may behave sometimes as a “winter annual.” By July, the tubercles are about sign of life, their tuberous bases being in Th j . : an- nual plant ere seems little doubt that Epifagus is an Cc sa vascular System is a dictyostele, consisting of usually bicollateral bi ar bundles, with the Phloem exceeding the xylem in extent. In the orescence the bundles are regularly arranged in a ring, but in the a ee confused and irregular arrangement, running . Planes apparently” (Cooke & Schively). However, aise noted, in the tuber, that the phloem “follows a very irregular et rt pattern” but “the xylem masses . . . are arranged in a ring.” ae nics 6a are Steatly developed in the inflorescence, giving it its Epifagus. we Batare, Brooks found no evidence of a cambium in Biri: ti “staplers” by Fergus, have neither root caps nor vascular bund] « cortical parenchyma cells closest to their small central ar bundle have greatly thickened cell walls; it is these cells that account largely for the stiffness and rigidity of the roots. 1971] THIERET, GENERA OF OROBANCHACEAE 419 Cooke & Schively reported that endosperm formation, after fusion of the polar nuclei, begins before fertilization in E. virginiana and that the second male gamete is non-functional. This report of “precocious albumen” requires confirmation. Beechdrops was once of some notoriety as a medicinal herb. Three of its common names — cancer-root, clap-wort, and flux-plant — attest to the variety of its uses. REFERENCES: Under family references see BECK voN MANNAGETTA (BECK- iprmaeanpa BENTHAM & HOOKER, BOESHORE, ren Kurjt, Martin, METCALFE HALK, REUTER, THIERET, and WALL LRO Attarp, H. A. Epifagus gotacp (L.) Py forma pallida Weatherby in West Vicsinia. Castanea 20: Austin, C. F., & E. A. ee Bibi virginiana, Bart., var. rauana, Austin. Bull. Torrey Bot. Club 6: 65, 66. et, J. H. Nomenclatural a Bull, Torrey Bot, Club 24: 409-411, 897, [Epifagus, 409, 410.] Fee A. E. A preliminary sou genie 9° Nee of Epifagus virginiana (L.) Bart. Proc. oe Acad. Sci. 70: 73-78. 1961. es & A. F. SCHIVELY. eee on ath structure and mpeg pidieelt virginiana. Contr. Bot. Lab. Univ. Penn. 2: 352-3 1904. he S. T. Epiphegus virginiana. Bot. Gaz. 8: 154-156. 1883. “Thtiecel. HernAnvez X., E., H. Crum, W. B. Fox, & A. J. Suarp. A unique vegetational sland mexicana near Gomez Farias; list of associates includes E. vir- oe E. I. Flora of the White Lake region, Michigan, and its ecological rela- tions. Bot. Gaz. 29: 419-436. 1900. [E. virginiana, 430, 431; saprophytic relationship between beechdrops and beech suggested. | Hoim, T. Medicinal plants of North pen aA jae panned 1912. [Morphology an ots os ne pea ae Bot. Centralbl. 44(1): 1-89. 1927. [Epifagus, 63; leaf-scale anatomy. | Korpren, A. C. E ip. egus virginiana. virginiana Bart. Am. Jour. Pharm. IV. 23: 274-276. * Epiphegus. Bot. Gaz. 33: 376. 1902. S ’ [Reprinted i in Plant World 5: 114. 1902. ; MEEHAN, T. The native flowers and eae of e nie States. Vol. 2. v + 200 pp. Boston. 1879. [E. virginiana, 93-9 MitispaucH, C. F. esi te (beech drops). Re aii Recorder 4: 10-20. 1889. [Medicinal u Mrranpa, F., & ApS HARP. Chineselbeisilcs of the vegetation perate regions i eastern Mexico. eid 31: 313-333. giniana near Zacualtipan, Hidalgo, 317-] r Patme_r, S. C. Epifagus wicginiine. Rhodora 57: 71, 72. 1955. [A colony of E. virginiana that is “butter yellow will Pease, A. 9 A color-form of seek. tirone, Rhodora 54: 140. 1952. [Forma bd atropurpurea. | in certain tem- 1950. [E. vir- 420 JOURNAL OF THE ARNOLD ARBORETUM [vox. 52 SCHRENK, H. Parasitism of Epiphegus virginiana. Proc. Am. Microscop. Soc. 15: 91-128. 1894. SHarp, A. J., E. HernANnvez X., H. Crum, & W. B. Fox. Nota floristica de una asociacién importante del suroeste de Tamaulipas, México. Soc. Bot. México Bol. 11: 1-4. 1951. [Stand of Fagus mexicana near Gomez Farias : list of associates includes E. virginiana. | STEYERMARK, J. A. Epifagus virginiana in Missouri, Rhodora 36: 352,308. Poa THIERET, J. W. Notes on Epifagus. Castanea 34: 397-402. 1970 “1969.” [ In- cludes distribution map. WEATHERBY, C. A. Epifagus virginiana (L.) Bart., forma pallida, n. f. Rho- dora 36: 59. 1934, 2. Conopholis Wallroth, Orobanches Generis Diaskeue 78. 1825. each flower with [1] 2 bractlets [or none], short- to long-pedicelled. Calyx tubular, split anteriorly (‘“spathaceous”), 2- or 4—5-toothed or y the greater length-width ratio of its Bietiens ; by its denser and thicker inflorescence, and by its y e eing usually deciduous from the fruit. Haynes, the recent mono- slg er of Conopholis, recognized two varieties of C. alpina — var. alpina and’ Var. mexicana (Gray ex Watson) Haynes — distinguished by range, vestiture. Conopholis americana and C. alpina that they could well be considered conspecific. ed to treat the eastern and western populations their reproductive isolation, their morphological “H/ Ficure 2. Conopholis. a-j, C. americana: a, fruiting plant detached from root of Quercus rubra, which was to right — note scaly buds on “tuber,” < 1/2; b, bud, from below to show calyx with two bractlets and aestivation of petals, X 4; c, d, lateral views of two flowers, X 4; e, f, two stamens, ; g, diagram- Matic cross section of ovary with five placentae, X 6; h, diagrammatic cross section of fruit with four placentae, x 2; i, j, two seeds, X 12. distinctness, and their apparent host specificity. So far as is known, the two taxa approach no closer than about 800 miles to each other. The oaks available as hosts for the two species differ. Only Quercus Muehlen- bergii and Q. virginiana, neither known to be parasitized by Conopholis, occur sympatrically with both species of the genus. Whether one species of Conopholis could parasitize the hosts of the other is not known. While genera other than Quercus have been reported to be hosts for Conopholis, in all cases where the parasitized root has been traced to its origin, it has been found to belong to an oak. Boeshore and Percival indicated the hosts to be members of the red oak group (Quercus subg. 422 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Q. Shumardii, and Q. velutina. Many of these records require confirmation. Collectors of Conopholis should attempt to ascertain — not always an easy task — the host plant. If this cannot be done, at least a piece of the parasitized root should be collected to permit identification, by means of wood anatomy, of the oak group to which the host belongs. The mature plant of Conopholis americana consists of one or more flowering shoots that arise endogenously from an enlarged “gall” at- tached to an oak root. The root does not usually extend past the “gall.” The flowering shoots are annual, but the “gall” is perennial, perhaps liv- ing up to several years before flowers are produced. ‘“‘Galls” are re- corded to range from 0.5 to 10 inches in diameter and are irregularly Shaped. Large conglomerate masses may be formed by the fusion of two or more (to at least 18) “galls.” The “galls” may be completely sub- terranean or be partially exposed at the soil surface. The “gall” is cov- ered with a thick, porous dark brown bark and is composed of “‘innu- merable granules of sclerenchyma” (Wilson) and parenchyma, The aggregates of stone cells are its most prominent feature. The nature of the “gall” is uncertain. Wilson regarded it as “in the main, a modification of the host’? caused by the “irritant action of the parasite” that brings about “swelling up of the host root, and enormous multiplication of its sclerenchyma patches.” Boeshore concurred. Perci- val, however, interpreted the “gall” as consisting “almost entirely of the stem tissues of the parasite, apparently in more or less disorganized ar- b excessive hypertrophy.’”’ The vascular undles can be traced directly from the shoot through the “gall” to their » the stomata being either open or closed (Percival). Boeshore . alpina var. alpina in Tamaulipas. Evi- at least occasionally, be self-pollinated in the : . — of pollen from dehisced anthers ran “Preserved plant of C. americana showed that abou a of the pollen grains had already seocurioaery : ae ries: of Conopholis are large and few relative to those of other nenaceae. Their highly diversified shapes develop as a result of areas exerted by the enlarging ovules. Double ovules and seeds are requent. Nothing is known of seed dispersal in the genus. 1971] THIERET, GENERA OF OROBANCHACEAE 423 The early stages in the life history of Conopholis are almost unknown. Percival tried various ways to germinate seeds of C. americana, both in the field and in the laboratory. His study, which involved about 22,000 seeds, showed negative results except where the seeds were in proximity to oak roots. The embryos of several of these seeds showed “a modification of cell content and absorption of food from the endosperm.” These changes may have represented the first stage in germination. A single seed was found to have developed a radicle that had attained a length of 3 mm. before it came into contact with and penetrated an actively growing oak root tip. The vascular system of the shoot is a dictyostele composed of two (Boeshore, Wilson) or three (Chatin, Percival) rings of bundles. That the inner ring is much less prominent than the others may account for the different interpretations of ring number. Wilson described the bundles as collateral, with the xylem and phloem inverted in the bundles of the inner ring; Boeshore interpreted the bundles as being bicollateral. Percival, whose account probably is the most reliable, regarded the bundles as col- lateral, with those of the middle ring having inverted xylem and phloem. He identified a cambium in the bundles of the outer and middle rings. Percival was unable to locate sieve tubes and companion cells in the phloem of Conopholis americana, although Wilson had reported the pres- ence of both these cell types. The phloem, according to Percival, is made up of phloem parenchyma — which he regarded as “apparently quite use- less as a conductive tissue, the more so since it was often completely crushed in the basal section of the flowering shoot. The xylem vessels were numerous and well developed; they evidently are the main pathways for the conduction of food and water from the host.” Percival considered it logical to assume that any Conopholis cells could obtain dissolved foods from con- tiguous cells of the host because of the difference in sap concentration re- ported by Doak. ; i th An interesting reaction of the host root to C. americana 1s the gradua infiltration of tannin into the root tissues adjacent to the tissues of the parasite. The older a “gall’’ becomes, the more completely the se saere tissues of the host are infiltrated and plugged until, at length, the “ga dies and disintegrates, leaving an ok root well protected by tannous deposits against attack by decay-producing organisms. So far ng ere G hon is of no economic importance. An sane of the whole plant of C. americana is said to possess insecticidal properties, REFERENCES: Under family references see BECK VON MANNAGETTA “eoormtneecemieg 4 BENTHAM & HOooKeER, BorsHorE, CHATIN, KUIJT, — I CHALK, REUTER, THIERET, B. TIAGI (1965), and eeproaeit an Ciute, W. N. Rarity of Conopholis. Am. Bot. a: a seibagy odlid Doax, K. D. Parasitism of Conopholis americana Wallr. on F : 9: 102. 1929. ° bicolor Willd. (Abstr.) Phytopathology 1 ; 4 FEvELL, A. J. ae Lieferung 4 in J. von WiEsNER, Die Rohstoffe des Pflanzenreichs, ed. 5. 244 pp. Weinheim. 1965. [Conopholis, 115.] 424 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Haynes, R. R. Conopholis alpina Liebmann var. mexicana (Gray ex Watson) Haynes, comb. nov. (Orobanchaceae). Sida 3: 347. 1969. . A monograph of the genus Conopholis (Orobanchaceae). 75 pp. M. S. thesis, University of Southwestern Louisiana, Lafayette. 1969. A monograph of the genus Cono pholis (Orobanchaceae). Sida (in press). JENNINGS, O. E. Rarity of Conopholis. Am. Bot. 26: 29. 1920. Lewis, W. H. Chromosome numbers of phanerogams. I. Ann. Missouri Bot. Gard. 53: 100-103. 1966. [C. americana, 100. | LIEBMANN, F. M. To nye arter af slaegten Conopholis Wallr. Forh. Skand. Naturf. Christiania 4: 184-136. 1847. [C. alpina, C. sylvatica.] McVaucu, R. Suggested phylogeny of Prunus serotina and other wide-ranging phylads in North America. Brittonia 7: 317-346, 1952. [C. mexicana, 342, Map 25.] Mortier, D. M. On certain plastids, with special reference to the protein bodies of Zea, Ricinus, and Conopholis. Ann. Bot. 35: 349-364. 1921. Norton, A. H. New stations for Peltandra virginica and Conopholis americana in Maine. Rhodora 4: 168, 169. 1902. PEeRcIVAL, W. C. The parasitism of Conopholis americana on Quercus borealis. Am. Jour. Bot. 18: 817-837. bls. 55-58. 1931. [= Q. rubra.] Witson, L. L. W. Observations on Conopholis americana. Contr, Bot. Lab. Woopson, R. E., Jr., & R. J. Sewerr. Contributions toward a flora of Panama. II. Miscellaneous collections during 1936-1938. Ann. Missouri Bot. Gard. 25: 823-840. 1938. [C. panamensis, 835, 836.] re -Orobanche Linnaeus, ep. Pl 2: 632. 1753; Gen. Pl. ed. 5. 281. 1754. _ Low, glandular-pubescent to nearly glabrous, yellowish to tan, some- times reddish- or bluish-tinged, simple or branched annuals or perennials. Base of plant slender to enlarged; secondary roots present. Stem covered stigma variable, commonly crateriform, bilamellate, or 2-lobed. Capsule ovoid to ellipsoid, dehiscing anteroposteriorly, the valves sometimes re- maining attached at their tips. Seeds minute, numerous, cylindric, ovoid, or ellipsoid; testa finely reticulate. (Including Aphyllon Torr. & Gray in Gray, M yzorrhiza Phil., and Thalesig Raf.) Lecrorypr sprctes: O. major L.; see Britton & Brown, Illus, Fl. No, U. S. ed. 2. 3: 234. 1913. (Name from Greek, orobos, vetch, and anchein, to strangle, in allusion to the plants’ parasitic habits, — BRoomrapr, 1971] THIERET, GENERA OF OROBANCHACEAE 425 The largest genus of Orobanchaceae, with about 100 species, a majority (about 90 species of sects. OROBANCHE and TRIONYCHON) native to Eurasia and Africa, especially the Mediterranean region, about 10 species (sects. EUANOPLON and MyzorruizA) American. Several Old World species are widely distributed weeds. Each of the four sections of Oroban- che has been accorded generic status by various workers; Beck-Manna- getta’s inclusive interpretation of the genus is almost universally accepted today. In the southeastern United States, Orobanche is represented by one native and one introduced species, which belong to different sections. Section EuaNopLon (Endl. ex Walp.) Thieret ® (§ Aphyllon (Torr. & Gray) G. Beck and § Gymnocaulis Nutt.), with two species, is character- ized by its long-pedicelled, ebracteolate flowers and by its campanulate, subregular, 5-cleft to 5-parted calyx. It is represented in the southeastern United States by O. uniflora L. var. uniflora (Aphyllon uniflorum (L.) Torr. & Gray in Gray; Thalesia uniflora (L.) Raf.), 2n = 36, 72 (in a diploid-parthenogenetic form). The species ranges, in several too con- fluent varieties, over much of extreme southern Canada (including New- foundland) and the conterminous United States but is not yet recorded from Manitoba, Saskatchewan, South Dakota, Arizona, New Mexico, and Louisiana. Its habitats include moist to dry deciduous or mixed woods, stream banks, rocky glades, and grassy roadsides; it has been recorded as parasitic on Artemisia, Aster, Coreopsis, Lithophragma, Potentilla, Quer- guished two kinds of roots: those forming parasitic connections, and the “soil roots,” which do not. The soil roots seem functionless in absorption, probably serving as supporting organs. Neither kind of root possesses root hairs or root cap. The vascular tissues are disposed in a diarch or, less often, a triarch arrangement. The phloem exceeds the xylem in extent. Sieve tubes seem to be absent. aoe ‘ The vascular system of the stem is a dictyostele composed of a ms o collateral bundles, with the phloem better developed than the xylem. ( tubes are lacking; the phloem consists of * elongated elements hear granular contents, which are sometimes nucleated.” The xylem ~~ s “tracheae.” Smith could not, with assurance, identify a cambium. He one cluded further that the xylem has no conductive function but serves solely to support and strengthen the plant. ; ees recognized ie species in sect. EUANOPLON, = sp ppey mole and O. fasciculata Nutt. Usually rather easily distinguis : Kotapdiirs connected by occasional specimens of intermediate morphology. aay fasciculata var. subulata Goodman, described from Oklahoma, 1s pore A placed in either species, at least on the basis of herbarium ma nes biosystematic study of sect. EUANOPLON 1S a poe op a Section OROBANCHE (§ Ospreolen Wallr.), with abou sauitad : © Orobanche sect. Euanoplon (Endl. ex Walp.) a comb. nov. Anoplan thus sect. Euanoplon Endl. ex Walp., Repertorium 3: 480. 1844. 426 JOURNAL OF THE ARNOLD ARBORETUM [vou. 52 tion. To what extent other broomrapes follow this pattern is not known, although flowering in “several strains” of Orobanche occurs on hosts that are in purely vegetative condition (Kribben). ae An additional section, TRIONYCHON Wallr., characterized by its bibrac- teolate flowers and its mostly 4-lobed calyx, is represented in the United States by the introduced O. ramosa L., 2n = 24, a Eurasian species that has been collected on Cannabis in Kentucky and that may yet be found in The flowers of Orobanche are homogamous or, more rarely, proterogy- nous. They may or may not secrete nectar. Pollination in some species 1S by insects, especially Hymenoptera. In other species, the flowers appear to be self-pollinated. The seeds are distributed mainly by wind and by rain wash, They have been shown to pass unharmed through the digestive tract of cattle but subsequently to be rendered inviable during fermenta- tion of the feces. Chromosome counts are available for members of all four sections of Orobanche: EUANoPLon, 2n = 36, 72 (one species); Myzorruiza, 2n = 24 (one species) ; OROBANCHE, 2n = 38 (12 species) or 38, 40 (one species); and TRIONYcHoN, 2n = 24 (three species). Gardé suggested that six is the basic number for the genus and that those species having 38 chromosomes are heteroploids, 6n 3 ae ; As _angiospermous Parasites of agricultural plants, various species of Orobanche rank in importance with the witchweeds (Striga and Alectra, tr. “Catch crops,” plants that serve as hosts for the parasite, may be sown and then, before the parasite has produced - “Trap crops,” plants that do not serve as hosts but do stimulate the germination of Orobanche seeds, have been used with some success. Control with chemi resistant strains of crop plants 1s under investigation, especially in Helianthus. 1971] THIERET, GENERA OF OROBANCHACEAE 427 Successful use of an extract of Orobanche crenata to treat kidney stones has been reported (Ibrahim). REFERENCES: Under family references see BECK vON MANNAGETTA (BECK-MANNAGETTA), BENTHAM & HOooKEr, BoErsHORE, CHATIN, Davis, GuiSi¢, GuEpEs, HAMBLER (1954, 1956, 1958), HoveLacque (1888, 1892), Kinc, Knutu, KOH Ler, Kurt, AUDI & ALBERTINI, METCALFE & CHALK, NIKITICHEVA, OZENDA, RAO, RAuH, REUTER, Ross-CrAIc, SCHMUCKER, SRINIVASAN & SUBRAMANIAN, SUNDERLAND (1960a, 1960b, 1960c), THteret, B. Tract (1951), Tract & SANKHLA, Y. D. TIAcI, WALLROTH, and WIESNER. AcHEy, D. M. A revision of the section Gymnocaulis of the genus Orobanche. Bull. Torrey Bot. Club 60: 441-451. 1933. ALBRECHT, H. Der Kleewiirger (Orobanche minor). Prakt. Blatt. Pflanzenbau 4: 98-101. 1906.* Anonymous. Broom-rape. Jour. Board Agr. London 23: 478-481. 1916. [ Notes . minor. BAccarint, A., & B. A. MELANDRI. Studies on Orobanche hederae physiology: pigments and CO, fixation. Physiol. Pl. 20: 245-250. 1967. [Includes chromatography. | Baroy, Identification des anthocyannes d’Orobanche minor Sutt. Ann. Physiol. Vég. 5: 141-149. 1963. , BEcK von MannacettA, G. R. Monographie der Gattung Orobanche. Bibliot. Bot. 4(19). 275 pp. 4 pls. 3 maps. 1890. pCa be See 1: 73-81. maps 61-69. 1927. [Author’s A Bec, M. U., S. Prakasu, M. Sincu, K. K. Tewanrr, & P. S. KrisHNAN. Bio- chemical aspects of parasitism by the angiosperm parasites: III — Phytic acid & other forms of acid-soluble phosphate in angiosperm parasites & hosts. Indian Jour. Biochem. 5: 157-160. 1968. 5 BENNETT, J. C. Orobanche uniflora L. Jour. Roy. Hort. Soc. 59: 397, 398. iten. Ber. Deutsch. Bot. Ges. la 1934. Bercpott, E. Uber die Saugkrifte einiger Paras Lathraea squamaria, L. clan- 45: 293-301. 1927. [Includes O. speciosa, destina. A BIFFEN, R. as Weeds. Jour. Roy. Agr. Soc. England 74: 376-379. 1913. [ Notes on O. minor. : BLANCHARD, M. suman d’un composé organomercurique sur le apo de lorobanche du pois (Orobanche speciosa D.C.). Compt. Rend. Acad. Sci. Paris 233: 1224-1226. 1951la. Contribution a l’étude de la biologie de l’Orobanche et 4 sa destruction. ce 37: 582-584. 1951b. , : étude de la biologie de YOrobanche et 4 sa destruction. ; i is 178: cules de VOrobanche rapum Thuill. Compt. Rend. Acad. Sci. Paris 1 oes oe le processus du noircissement des orobanches au Sagi * leur deseiceation, ibid. 180: 387, 38. 1929- [Attributed to ox orobanchoside, a glucoside. ] 428 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Brown, R., A. D. GREENWoop, A. W. Jounson, A. R. Lanspown, A. G. Lone, & N. SUNDERLAND. The Orobanche germination factor. 3. Concentration of the factor by counter-current distribution. Biochem. Jour. 52: 571-574, 1952. ) ———, ——, & A. G. Lone. The stimulant involved in the germina- tion of Orobanche minor Sm. 1. Assay technique and bulk preparation of the stimulant. Biochem. Jour. 48: 559-564. 1951. , » ——, ——, & G. J. Tyter. The stimulant involved in the germination of Orobanche minor Sm. 2. Chromatographic purification of crude concentrates. Biochem. Jour. 48: 564-568. 1951, Burkart, A. Las Orobancaceas, especialmente Orobanche chilensis y su dis- tribucién geografica en la Republica Argentina. Darwiniana 4: 303-310. 1942, CAPPELLETTI, C. Ricerche sulla germinazione dei semi di Orobanche. Nuovo Giorn. Bot. Ital. I. 41: 441-443. 1934. [O. gracilis. ] - Osservazioni sulla germinazione dei semi di Orobanche gracilis Sm. e Orobanche crenata Forsk. Ibid. 43: 263-266. 1936. Ibid. 44: 331-334. 1937. CARPENTER, T. roomrape on tomato and other hosts in southern California. (Abstr. ) Phytopathology 47: 518, 1957. [O. ludoviciana. | Carter, K. M. A contribution to the cytology of the ovule of Orobanche minor. Jour. Roy. Microscop. Soc. III. 48: 389-403. 1928. Caspary, R, Samen, Keimung, Species und Nahrpflanzen der Orobanchen. Flora 37: 577-588, 593-603. 1854, Cessera, J. D. Origin and development of the female gametophyte, endosperm ime egret in Orobanche uniflora. Bull. Torrey Bot. Club 62: 455-466. . CHABROLIN, C. La germination des graines d’Orobanche. Compt. Rend. Acad. Sci. Paris 198: 2275-2277. 1934, — Germination des graines et plantes-hétes de l’orobanche de la féve (Orobanche speciosa DC.). Ibid. 200: 1974-1976. 1935. La germination des graines de l’Orobanche speciosa, Ibid, 206: 1990- - 1938, : 1992 CLuTe, W. N. The meaning of plant names, LXII. Bladderworts and broom rapes. Am. Bot. 41: 60-64. 1935. Curter, E, G. Anatomical studies on the shoot apices of some parasitic and saprophytic angiosperms, Phytomorphology 5: 231-247. 1955. [Includes Lathraea., | Davies, W. E. Experiments on the control of broomrape in red clover. Pl. Pathol. 8(1): 19-22. 1959. [0. minor. the branched broomrape and its occurrence in Califor- nia. Pl. Disease Rep. 37: 136, 137. 1953. [O. ramosa. | Evans, D. C. What about broomrape? Agr. Gaz, New South Wales 73: 200- 202. 1962. [O. minor, 0. australiana, | s ARDE, A. Nota cariologica sobre tres Orobanchacees Portugesas. Genét. Ibér. 5: 133-142, 1 pl. 1951, [Reports chromosome number of O. crenata, 0. ramosa, Cistanche Phelipaea. } Garman, H. The broom-rape of hemp and tobacco (Phelipaea ramosa, Linn.). “sep. Aentucky Agr. Exper. Sta. 1890: 57-73. 1890. [O. ramosa; also as Kentucky Agr. Exper. Sta. Bull. 24. he broom-rapes, Kentucky Agr. Exper. Sta, Bull. 105: 1-32. 1903. 1971] THIERET, GENERA OF OROBANCHACEAE 429 Git, L. S. Broomrapes, dodders, and mistletoes. Yearb. U.S. Dep. Agr. 1953: 43-17. GILLI, A. Bestimmungsschliissel der air iaig tears Varietaten und Formen von Orobanche. Verh. Zool.-Bot. Ges. Wien 105/106: 171-181. 1966. GoopMAN, G. J. A new variety in Forts a Leafl. West. Bot. 5: 36. 1947. [O. fasciculata var. subulata in Oklahoma and Texas. | ———. The genus Orobanche in Oklahoma. Proc. Okla. Acad. Sci. 33: 173. 1954. Gossett, D. M., & L. SHaw. Broomrape on burley tobacco in a Carolina, 1963. Pl. Disease Rep. 48: 508. 1964. [Apparently O. mino GraHAM, R. A. A new combination in Orobanche. Kew Bull. 1955: 467. 1955. O. ramosa var. brevispicata (Ledeb.) Graham. | GRESHNOVA, V. N., . B. NATALYINA. On the susceptibility of broomrape (Orobanche cumana) parasitic on corn to corn smut (Ustilago zeae). (In Russian.) Bot. Zhur. 49: 599. 1964. Guarp, A. T., & W. H. Srtver. A new station for Orobanche ludoviciana. Proc. Indiana Rigi sci. 51; 116, 117.,1943, GuEGUEN, F. Recherches sur l’Orobanche. Ann. Epiph. 1: 433-436. 1913. 0. minor. Gurtmarigs, J. p’A. Monographia das Orobanchaceas. Brotéria 3: 5-208. pls. 1-14. 1904. [In Portugal; Orobanche, Cistanche.] Gupta, S. C., L. C. SHarma, & G. G. Dateta. Occurrence of Orobanche cernua on Pluchea lanceolata, an obnoxious weed in Rajasthan, India. Pl. Disease Rep. 53: 43, 44. 1969. HALLER, R. Sher den Character der Zellwandsubstanz eee Pflanzen. Ber. Schweiz. Bot. Ges. 59: 155-161. 1949. [Includes O. m —. Untersuchungen an Heterotrophen-material. Ibid. 63: 384-389. 1953. eae eae includes Orobanche. Hatstep, B. D. A study of weed roots. New Jersey Agr. Exper. Sta. Ann. Rep. 1891: 13- 1891. [Includes O. ramosa. : i 318 18 parasites of ‘i red clover. Bull. Torrey Bot. Club 25: 395-397. 1898. [Includes O. mino Branched acy -rape upon peters Rhodora 3: 295. 1901. [O. ra- ane E. J. Owen, & J. K. SHaw. Broom-rape on Pelargonium. = Jersey Sinte ‘Agr. Exper. Sta. Ann. Rep. 1905: 509. 1 pl. 1905. [O. m ‘ yt fe Hyetmouist, H. Nagra Orobanche-notiser. Bot. Not. 116: 34 [Notes on host species. | HotpswortH, M., & P. S. NUTMAN. Flowering responses in banche minor. Nature 160: 223, 224. ais be HooreseEkKE, C.-J. vAN. Mémoire sur les orobanc tion a la culture du ag dans les communes ou St ache nuit ture. 22 pp. Gand. 1 Horner, F. D. Ae oe Garde “branched tuber.” ] Hovetacgue, M. Développement et valeur banches. Compt. Rend. Acad. Sci. Paris . Sur le développement et la structure des je 530-533. 1887b. IsraHim, F. D. Preliminary studi sclinns of the extract of Orobanche crenata. a strain of Oro- r servir d’instruc- a sa cul- n London 61: 31. 1902. [Plant with morphologique du sucoir des oro- 105: 470-473. 1887a unes orobanches. /bid. hysiological and pharmacological epg Jour. Egypt. Med. Assoc. 51: 430 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 939-948. 1968. [“Extract causes an initial inhibition, followed by rapid recovery and strengthening of the frog’s heart”; it “seems to augment the rhythmic activity of rabbit’s small intestines.” “The use of the watery ex- tract in patients suffering from renal calculi seems to give marked relief of symptoms and rapid discharge of the calculi in some patients.”’] Izarp, C., & T. Berc. Une nouvelle méthode d’identification des virus TMV et x. Application 4 l’étude des relations entre l’Orobanche et le tabac, en pré- sence de ces deux virus. Compt. Rend. Acad. Sci. Paris 247: 1526-1529. 1958. [QO. ramosa. —— & H. Hirer. Obtention de la germination “in vitro” des graines de l’orobanche du tabac. Compt. Rend. Acad. Agr. France 39: 567-569. 1953. [O. ramosa. | & —. Action du 1.3-dichloropropane 1.2-dichloropropéne, de la rindite et de la gibberelline, sur la germination des graines de ]’Orobanche, parasite du tabac. Compt. Rend. Acad. Sci. Paris 246: 2659-2661. 1958. [Generic name only. JeENseN, H. W. The normal and parthenogenetic forms of Orobanche uniflora L. in the eastern United States. Cellule 54: 135-142. 1951. [Diploid par- henogenesis. Jeppsson, L. Ett fynd av Orobanche minor Sm. in Skane. Bot. Not. 120: 488, . 1967. Jonsson, B. Iakttagelser dfver tillvaxten hos Orobanche-arter. Acta Univ. Lund. 31(5). 23 pp. 1895, RY, A. E. R., & H. Tewric. A contribution to the morphology and anat- pire of seed germination in Orobanche crenata, Bot. Not. 109: 385-399. a. . - Seed germination in Orobanche crenata Forssk. Sv. Bot. Tidskr. 50: 270-286. 1956b. Kamet, S. H. Etude chimique et toxicologique d’une plante égyptienne: Oro- ae ee Sutton. Revue Elevage Méd. Vétérin. Pays Trop. II. 9 Kanna, S. K., P. N. VISWANATHAN, C. P. Tewart, P. S. KrisHNaN, & G. G. ANWAL. Biochemical aspects of parasitism by the angiosperm parasites: ar in parasites and hosts. Physiol. Pl. 21: 949-959. 1968. [O. aegyb- iaca, Kocn, L. Ueber die Entwi cklung d : hrb. Wiss. Bot. 11: 218-261, 1878. menS der Orobanchen. Jahr ich . obanchen, mit besonderer Beriick- sichtigung ihrer Beziehungen zu den Kulturpflanzen. 389 pp. Heidelberg. : 1887. [Abstract in Bot. Jahrb. 9(Lit-ber.): 15-17. 1888. | ORFF, G, Der Kleeteufel (Orobanche minor Sutt.) und seine Bekampfung. Prakt. Blatt. Pflanzenbau 4: 109-114. 1906.* Krenner, J. A. The natural history of the sunflower broomrape (Orobanche cumana Wallr.). © morphological anatomy of the sunflower ne nn ; n and the infectio hanism of its germ. Acta Bot. Acad. Sci. Hungar. 4: 113-144 1958 ies cae ee a J. Die Bliitenbildung von Orobanche in Abhiangigkeit von der ntwicklungsphase des Wirtes, Ber. Deutsch. Bot. Ges. 64: 353-355. 1952. [O. hederae, O. ramosa, O. speciosa.) . . 1971] THIERET, GENERA OF OROBANCHACEAE 431 Kumar, L. S. S. Flowering plants which attack economic crops. II. Orobanche. Indian Farming 3: 638-640. 1942. [Hand weeding recommended. Lanspown, A. R. Chemistry of the Orobanche and Striga germination factors. Abstr. Diss. Univ. Cambridge 1953-1954: 234. 1956. [O. minor. ] Locayono, M. Criterii sui caratteri delle Orobanche ed enumerazione delle nuove specie rinvenute in Sicilia, Nat. Sicil. 1: 45-48, 53-56, 90-93, 131- 135, 162-165, 169-175, 198-202, 209-216, 255-258. 1881-1882. Ibid. 2: 11-15, 37-41, 59-64, 80-84, 105-110, 132-136. 1882-1883. [Orobanche, Phelypaea in Sicily.] Lone, A. G. Chemical factors involved in certain host-parasite systems. Abstr. Diss. Univ. Cambridge 1952-1953: 202-204. 1955. [Orobanche germina- tion factor. | Lonc, H. C. Broom-rape. Jour. Min, Agr. Great Britain 39: 311, 312. 7 pl. 1932. [O. minor. ] Matix, S. A. Study of the efficacy of various chemicals for the control of Orobanche in the form of their application at different intervals in rela- tion to yield of tomatoes (No. 37 crop). Pakistan Jour. Sci. 15: 197-200. 1963. [Generic name only; species not noted. ] MarTELLI, G. M. Nota preliminare sui parassiti animali dell’orobanche della fava “Orobanche speciosa” D.C. Rivista Patol. Veg. 23: 233-240. 19353. . & A. Estoppry, L’Orobanche. Chron. Agr. Canton Vaud 15: 402-404. 1902. [Notes on O. minor.] MaruparajAn, D. Note on Orobanche cernua Loefl. Curr. Sci. Bangalore 19: 64, 65. 1950. [Seeds remain viable after passing through cattle and goats. ] Marxovsxu, S. T. Control of Orobanche with the aid of Phytomyza. (In Russian.) Sovet. Agron. 9(8): 92, 93. 1951.* Aen! MEEHAN, - On Aphyllon as a root parasite. Proc. Acad. Nat. Sci. Philadel- phia 1887: 154. 1887. [O. uniflora, O. fasciculata. mee Merry, D. M. E. Broomrape invades cowgrass in Nelson District. New Zea- land Jour. Agr. 74: 308. 1947. [O. minor. | aoe ee a Moreno Marguez, V. Observaciones sobre la diseminacién del “jopo a ae banche crenata Forsk.). Bull. Patol. Veg. Entomol. Agr. 14: sacar a . La poliploidia como posible medio de obtener Mes haraehe e wae sistentes al “jopo,”’ (Orobanche crenata Forsk.). Bull. Patol. Veg. : Agr. 16: 243-252. 1949.* MUENSCHER, W. C. Coleus para 165. 1924a. [O. ramosa.] . Orobanche ramosa on a Coleus. Munz, P. A. The North American acer Orobanc Bull. Torrey Bot. Club 57: 611-624. ‘ ; Murray, M. CA new variety of the lesser broomrape Solera cage mad in Scotland. Ann, Scot. Nat. Hist. 1907: 253. 1907. [Fo Beck. se NA M. J. THIRUMALAC , : 54, rar SF ee Phytopath. Zeitsch. 2 421-428. 195 [Causes wet rot of the parasite but does not affect ho ‘seed germination. Nasu, S. M., & S. WiHeLM. Stimulation of Ss ludoviciana var. Phytopathology 50: 772-774. 1960. [0. ramos®, ™ Cooperi. | sitized by broom-rape. Gard, Chron. Am. 28: Rhodora 26: 133-135. 1924b. he, section Myzorrhiza. ar. A sclerotinia disease of Oro- 432 JOURNAL OF THE ARNOLD ARBORETUM [vou. 52 Oxiverra, M. De L. D’., and M. De L. V. Borces. Um parasita da “Orobanche crenata” Forsk. Brotéria Ci. Nat. 15: 95, 96. 1946. [An agromyzid fly. ] PALMGREN, O. Chromosome numbers in angiospermous plants. Bot. Not. 1943: 348-352. 1943. [Includes O. lucorum Persipsky, D. Zur Embryologie der Orobatiche cumana Wall. und der O. ra- mosa L. Bull. Jard. Bot. Kieff. 4: 6-10. 1926.* - On the formation of endosperm and haustories in Orobanche ramosa L. (In Russian; English summary.) /bid. 16: 89-101. 1933. PILLAI, . N., & N.S. Murty. Control of Orobanche by allyl alcohol. Indian Jou r. Agr. Sci. 38: 216-220. 1968, PRASAD, N, Control of Orobanche on tobacco by, 2 . ae sul- phate. Bansilal Amritlal Agr. Col. Mag. 6: 12, 1 . Control : pal in tobacco by Crag Herbicide I. Indian Tobac- co “4(3): 139, Price, J. M. Sat in Aphyllon uniflorum. Trans. Kansas Acad. Sci. 14: 132. 1896. [O. uniflora.] AT, G. Isolement et caractérisation de l’acide chlorogénique des _ tissus d’Orobanche Soe: Duby. Compt. Rend. Acad. Sci. Paris 249: 456, 457. 1959a. rae Duby). Ibid. 156-158. . Recherches sur les phanérogames parasites (étude d’Orobanche hede- rae Duby). Ann. Sci. Nat. Bot. XII. 1: 721-871. 1960. Pucstey, H. W. Notes on Orobanche L, Jour. Bot. 78: 105-116. 1940. [Brit- ish species. ] Pusuxareva, K. V. The characteristics of the seeds of different biological races of broomrape, Orobanche cumana. (In Russian.) Izv. Optyn. Severn. Kavkaza 19: 155-166. 1930. [Races cannot be distinguished by seed char- - Sur la germination in vitro de l’orobanche du lierre (Orobanche hede- 1959b. » A. Carum ajowan Benth, et Hook., une nouvelle et intéressante plante nourriciére de !Orobanche ramosa L. ‘Jour. Agr. Trop. Bot. Appl. 5: 491-496. 1958. Nouvelles observations sur ramosa L. Ibid, 6: 111-114. | 59. Rao, D. New Pee of Orobanch nua, O. indica. | & N. C. JHA. Some new hosts of Orobanche. Ibid. 23: 405. 1954. [O. indica. | Rao, P. G. A rapid method for studying the germination of the seeds of the root parasite Rabe cernua Loefl. var. desertorum (Beck). Sci. Cult. 21: 258-261. 1955. pect ¥. See la germination des graines d’Orobanche e. Curr. Sci. Bangalore 22: 311. 1953. [O. cer- S. AMBASHT. Root relations of Orobanche and its hosts. rr. Sci. Bangalore 27: 445, 446. 1958. Seineicasans D. & N.R du calcium, ‘de que Soc. Bot. France 113: SCHULTz, F. W. B AYNAUD. Taux et formes des éléments, en particulier Iques_phanérogames parasites nord-africaines. Bull. 439-448. 1966. [Orobanche sp., Cistanche violacea.] Beitrag zur Kenntniss der deutschen Orobanchen. 12 pp. 1 Pl. - [Showing peloria in O. caryophyllacea. | - Hatsscutu. Uber den Anschluss einiger héherer Para- siten an die Siebréhren der Wirtspflanzen. Ein Beitrag zum Plasmodesmen- problem. Jahrb. Wiss. Bot. 87: 324-355. 1938. [Orobanche, 336-347.] 1971] THIERET, GENERA OF OROBANCHACEAE 433 SHARMA, D. K. The control of Orobanche. Allahabad Farmer 34: 98, 99. 1960. SHARMA, S. L. Orobanche on wheat (N. P. 52). Curr. Sci. Bangalore 22: 56, 953 SHAw, F. J. F. Orobanche as a ape in Bihar. Mem. Dept. Agr. India Bot. Ser. 9(3): 107-130. 3 pls. 191 Stmmon, E. Anomalie florale de “aber Epithymum. Bull. Soc. Bot. Deux- Savres Vienne ide 14: 214-216. 1902. SINGH, M., D. V. Sincu, P. C. Misra, K. K. Tewart, & P. S. Kr1sHNAN. Bio- chemical aspects bee parasitism by the angiosperm parasites: starch accu- mulation. Physiol. Pl. 21: 525-538. 1968. [O. aegyptiaca. | . K. Tewart, & P. S. KRISHNAN. Biochemical aspects of parasitism by angiosperm parasites: Part V — Water relations in metabo- Polish angiosperms. Sixth contribution. Acta Biol. Cracov. Bot. 9: 31-58. 1966, [Includes O. flava. ] corte A. C. The structure and parasitism of Aphyllon uniflorum, Gray. Contr. t. Lab. Univ. Penn. 2: 111-121. 1901. [O. uniflora. ] betecnntt M. Composizione chimica dei semi di Orobanche speciosa DC. (= 0. lca Forsk.). Ann. Sperim. Agr. II. 1: 97-114. 1947. [Many references. | Starr, O. naan a uniflora {. nana. Bot. Mag. 153: pl. 9194. 1930 (“1927”). Starr, G. H. A new parasite ba tomatoes. Phytopathology 33: 257, 258. 1943. 0. ludoviciana in Wyoming. | Stout, G. L. Recurrence a eine Orobanche ramosa L., on tomato plants in seer Bull. Dep. Agr. Calif. 27: 166-171. 1938. & K. Wacnon. Branched broomrape, Orobanche ramosa L., a pest of tomato and certain other crops. Bull. Dep. Agr. Calif. 42: 45-51. 1953. Tate, P. On the anatomy of Orobanche eee Duby and its attachment to its host. New Phytol. 24: 284-293. 192 TERIOKHIN, E. S., & G. I. Ivanova. A contribution to the taxonomy of the Caucasian broomrapes (Orobanche). (In Russian; English summary). Bot. Zhur. 50: 1105-1112. ; i VaucueEr, J. P. Monographie des Orobanches. 69 pp. 16 pls. Genéve. 1827. Vyas, B. P. Control of Orobanche. Indian Farming 16(2): 51. 1966. WERNECK, H. L. Der Kleeteufel (Orobanche minor S.) in Oberésterreich und seine natiirlichen Feinde. Neuheiten Gebiet Pflanzenschutzes 29: 226, 227. 1936. [ Diptera. poner bik Tee hiie! Bedeutung von Orobanche minor ce. donau. (Ein Beitrag zur ieee pee ne hk seers der Bekamp ung des Schmarotzers.) Ang. Bot. ais ic S. Deleterious a. of drying eer of broomrape seed in soil. isease Rep. 38: 890-892. 1954 ees ai aig alate with special reference to the broomrapes. Calif. Weed Conf. . 10: 13-18. . History 7 gecesi (Orobanche r their control by preplant soil injection Sixteenth Int. Hort. Congr. ee 6 ve : & L. Benson. Vertical distributio field soil. Pl. Disease Rep. 38: 553, 554. 1954. [0. 7 amosa and O. ludoviciana) and with methyl bromide solutions. —399. 1963. f broomrape seed in tomato amosa. 434 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 —. Weeds as important interim hosts of broomrape. /bid. 39: 1955. [O. ramosa.] Wanner S., L. C. Benson, & J. E. SAcEN. Methyl bromide gives promising trol of broomrape. (Abstr.) Phytopathology 47: 537, 538. 1957. [O. ramosa. | —. Studies on the control of broomrape on tomatoes. Soil fumigation by methyl bromide is a promising control. Pl. Disease Rep. 42: 645-651. 1958. [O. ramosa.] , R. C. Srorxan, J. E. Sacen, & T. CarPeNTER. Large-scale soil fumi- gation against broomrape. Phytopathology 49: 530, 531. 1959. [O. ra- mosa, O. ludoviciana var. Cooperi.| DEPARTMENT OF BIOLOGY UNIVERSITY OF SOUTHWESTERN LOUISIANA LAFAYETTE, LOUISIANA 70501 1971] FULFORD ET AL., ELFIN FOREST, 15 435 THE ECOLOGY OF AN ELFIN FOREST IN PUERTO RICO, 15. A STUDY OF THE LEAFY HEPATIC FLORA OF THE LUQUILLO MOUNTAINS MARGARET FULFORD,!? BARBARA CRANDALL, and RAYMOND STOTLER THE LuQquitto Mountains, a part of the 28,000 acre Luquillo Na- tional Forest, are situated in the northeastern part of the island of Puerto Rico (FicurE 1). At low elevations the mountain slopes are covered with a luxuriant montane rain forest, while at somewhat higher elevations palm forests are dominant. The summits of some of the mountains are cov- ered by dwarf forests of stunted trees and shrubs and are known as “mossy” or elfin forests. In all three vegetation regions, moderate tempera- ture (20-24°C range during the course of a year), frequent (usually daily rains), and fairly low incident global radiation (Gates, 1969) pro- vide climatic conditions conducive to the development of a luxuriant leafy hepatic flora. The leafy Hepaticae are abundant and widespread throughout the Lu- quillo Range, covering the ground, tree trunks, branches, shrubs and leaves at all elevations, but it was not known whether the distribution of all the species was uniformly widespread, or whether elevation, micro- climate, and other edaphic factors which vary from place to place were reflected in restricted patterns of distribution. To ascertain whether changes in composition of the hepatic flora do occur with changes in elevation and microclimate, etc., a total of 740 random collections were made in ten fairly distinct areas of the Luquillo Mountains during Febru- ary of 1967. A permanent slide was made for every species found in each collection, and a list of species present was compiled for each sie Treating each of the collections as a sample quadrat, the erage each species was determined for each of the ten areas by dividing t : number of collections in which a species was found in an area, by the tota number of collections made in that same area (TABLE 1). In — the ten areas were compared, using Sgrensen’s ( 1948) sapvinlp re) “9 munity, to assess the degree of floristic similarity among them ( . . In this method the number of species common to two areas is simply €x- e.g., 37.2% of the species found in areas } | areas, while the remaining 62.8% of the species are found in only one of Science Foundation Grant, is assistance as our guide to Pico e : j for the del Oeste and to El Yunque. He has also supplied the lists of ae ates study areas. Both Dr. Carroll Wood and Dr. Howard ag 10 * - Te eaewasile Pico del Oeste thus giving our survey a more comprehensive coverag study area here. 436 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 ATLANTIC OCEAN @ RIO GRANDE PALMER \ @EL VERDE C) 7 e A EL YUNQUE J) Am MT. BRITTON One TORO ® © NAGUABO oO San Juan PICO DEL OESTE A PICODEL ESTE Palmer N PUERTO RICO FicurE 1. Map of Puerto Rico wj jon i : : with an enlargement of the study region In 7 Luquillo National Forest. The numbers 1 through 10 indicate the ten areas -* abate. (1) low elevation roadside; (2) rain forest around the La Pet Pape sahosgr area; (3) Mt. Britton Loop or Route 191; (4) U.S.D.A. La G) conn ©) aye to - El Toro trail; (6) high elevation rere ico del Oeste; (8) summit . Bri mmit 0 Yunque; and (10) summit of El ee ee Oh 1971] FULFORD ET AL., ELFIN FOREST, 15 437 the two (TABLE 2). Finally, utilizing the unweighted pair-group method with arithmetic averages of Sokal and Sneath (1963), a cluster analysis was made of the similarity coefficients of the ten areas or communities and a dendrogram based on percent similarity was constructed (FIGURE 2). DESCRIPTION OF THE TEN STUDY AREAS Area 1. This area is defined as a region of relatively low elevation (200 meters) approximately 8.0 kilometers north of the La Mina Recreational Area (Ficure 1). The 28 collections made here were restricted to the more or less disturbed vegetative areas along the highway. Although the rainfall of this region appears to be approximately equal to that found at the La Mina Recreational Area, the microclimate of the hepatic flora is much drier because along the roadside the tree trunks and soil are much more affected by drying breezes and increased solar radiation than they are in the dense, shaded forest. The adventive vegetation in this area consists of shrubs and seedling trees. Heterotrichum cymosum, Lantana camara, Eupatorium odoratum, Inga vera, Cecropia peltata, Casearia guianensis, Cordia globosa, Piper hispidum with Rubus rosaefolius and Gleichenia bifida forming tangles. The most often encountered species, in the order of their frequency of occurrence * are: Lopholejeunea sa- graeana and Ceratolejeunea maritima, both of which are widespread on the trunks of large Cecropia trees. Lejeunea flava, Cyclolejeunea convexistipa, Euosmolejeunea trifaria, Harpalejeunea subacuta, Alobiellopsis domini- censis, and Calypogeia laxa are also significant elements * of the hepatic flora of this region with the latter two being common on soil and the others, on the trunks of large specimens of Cecropia, bamboo, and tree ferns. In this exposed area of roadside plants were found two species, Archilejeunea viridissima and Ptychocoleus polycarpus, which were not encountered in any of the other nine areas of study. Area 2. The second area comprises the montane rain forest surrounding the La Mina Recreational Area and Restaurant (FIGURE 1). At “ vel vation of approximately 400 meters, the average annual apy o i ” area is 300 cm. per year (Recher & Recher, 1966). The rains sie ea i and usually heavy with rapid clearing following each torrential s . : go ; Be i be highly diversified, with buttress a 40 to 50 feet or more above the branches of the trees allow for y. The principal trees and ndens, Guarea ramiflora, Xanthoxylum martinicense, Nectandra patens, Myrcia —— ject latifolia, Sapium laurocerasus, Croton poecilanthus, Me oe whe and Jxora ferrea, A small stream with banks lined wi - ple sane boulders, traverses the rain forest just below the restauran 1) * For percent frequency of occurrence see TABLE 438 JOURNAL OF THE ARNOLD ARBORETUM [von. 52 utes to a fairly large waterfall in the northern part of the collecting area. A total of 86 species of leafy Hepaticae were found in the 116 collec- tions made from the trees, shrubs, vines, leaves, boulders, and soil in this area. The most often encountered species in the order of their fre- quency of occurrence * were Macrolejeunea subsimplex, Cyclolejeunea convexistipa, Trichocolea elliottii, Ceratolejeunea maritima, and Plagio- chila divaricata. Other significant species * in the hepatic flora include Prionolejeunea aequitexta, Drepanolejeunea fragilis, Bazzania schwanec- kiana, Lophocolea martiana and Ceratolejeunea cornuta. Of particular interest is the fact that here in the rain forest 14 species were found that were not encountered in any of the other nine areas. Listed in the order of their frequency of occurrence * these include C eratolejeunea rubiginosa, Plagiochila amoena, Odontolejeunea lunulata, Plagiochila tamariscina, Ceratolejeunea cubensis, Cheilolejeunea decidua, Leptolejeunea elliptica, Stictolejeunea squamata, Lophocolea guadeloupensis, Drepanolejeunea lichenicola, Bazzania eggersiana, Crossotolejeunea bermudiana, C. boryana, and Lopholejeunea miilleriana. Area 3. The roadside regions of the Mt. Britton Loop road or Route 191 are designated as study area 3 (Ficure 1). The overall elevation is 750 meters and the rainfall is approximately equal to that of area 2. As in area 1, collecting was restricted to more or less disturbed habitats such as roadside rocks, trees and shrubs, a large pile of decaying palms, a Cyclolejeunea convexistipa, Ceratolejeunea maritima, Taxilejeunea sul- phurea, Cyclole jeunea chitonia, Drepanolejeunea inchoata, and Macrole- jeunea subsimplex. In addition, Herberta divergens, Plagiochila rutilans, Trichocolea flaccida, Diplasiolejeunea pellucida, and Drepanolejeunea fragilis also form a Significant part of the flora. Only one of the species found here, Calypogeia caespitosa, was not found elsewhere. ve shrubs are dominated by coin ' , aceous plants, especially on the pestenos banks, consisted of Sida carpinifolia, Polygala paniculata, Ele- ; antopus scaber, Sauvagesia erecta, Synedrella nodiflora, Borreria rachysepala, Blechnum occidentale, and Lycopodium cernuum. The ele- * For percent frequency of occurrence see TABLE 1, 1971] FULFORD ET AL., ELFIN FOREST, 15 439 vation is only slightly higher and the microclimate more open and exposed to the winds of the mountain pass, so that the area is considerably more xerophytic than Area 3. Only 32 species of leafy hepatics occurred in the 55 collections made in this area, a fact which might be related to the larger number of hepatics here occurring on the soil bank along the highway. Dominant species in the order of their frequency of occurrence * include Ceratole- jeunea maritima, Syzygiella perfoliata, Taxilejeunea sulphurea, Neurole- jeunea breutelii, Cyclolejeunea convexistipa, Ceratolejeunea patentissima, and Leucosarmentum portoricense. Other species of significance are Neesioscyphus bicuspidatus, Harpalejeunea uncinata, Leucolejeunea xanthocarpa, and Omphalanthus filiformis. Only one species, Microle- jeunea monoica was found only in this area. Area 5. A total of 92 collections was made in this area, which consisted of rain forest and disturbed roadside vegetation approximately 5.0 kilo- meters south of the La Mina Recreation Area, 2.0 kilometers south and down the mountain from area 4. (FicurE 1). In the vicinity of the en- trance to the El Toro Trail, this area has an elevation of approximately 700 meters and a microclimate quite similar to that described for area 3. The trees in this area are primarily species of Nectandra, with Cyrilla racemiflora, Sapium laurocerasus, Alchornea latifolia, Croton poecilanthus, and Calycogonium squamulosum. Those of smaller stature included M:- conia pubescens, Ditta myricoides, Casearia decandra, and Hedyosmum arborescens. Fifty-one species were collected in the area. The most often encoun- tered species in the order of their frequency of occurrence * were Cera- tolejeunea maritima, Cyclolejeunea convexistipa, N eurolejeunea ate Cyclolejeunea chitonia, Lejeunea flava, Bazzania gracilis, and Plagiochila abrupta. Within this area were three species, Symbiezidium ate Euosmolejeunea clausa, and Lopholejeunea howet which were not foun in the other areas. Area 6. This area includes a large grassy ridge along the U.S. a then has been populated with numerous and a few scattered trees and shrubs. } and Daphnopsis americana are the ce is due to several species * For percent frequency of occurrence see TABLE 1. 440 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 The leafy hepatics found in this region occur on leaves, branches, and trunks of the scattered trees, on blades of the grasses and sedges, and on soil. Of the 43 species * found in the 41 collections made here Ceratole- jeunea maritima, Taxilejeunea sulphurea, Frullania atrata, Ceratolejeunea patentissima, and Omphalanthus filiformis were the most widespread. Other species of significance included Drepanolejeunea fragilis, D. inchoa- ta, Herberta divergens, Plagiochila bursata, and C yclolejeunea convex- istipa. Unlike all of the other areas studied, no species were found which appeared to be restricted to this area alone. 0.8 cal. cm.—2 min-! as compared to an incident global radiation of 1.3 cal. cm.—? min-! in San Juan (Gates, 1969). Lepidozia patens, and Radula Saccatiloba, Fou species were on this peak which were not found on any of the peaks or in th lypogeia elliottii, Nowellig isma prostratum. Pi yee Britton comprises study area 8 (Ficure 1), Reaching an ele- 10n of 941 meters, Mt, Britton is the lowest and hence the least exposed of occurrence see TABLE 1, n the forests at the lower elevations: Ca- dominicensis, N. caribbeania and Odontosch- * For percent frequency 1971] FULFORD ET AL., ELFIN FOREST, 15 441 of the mountain peaks studied. According to Recher and Recher (1966) the average annual rainfall on the summit is 500 cm. per year and the average temperature is 22°C. The phanerogamic vegetation of Mt. Brit- ton is best described as a Sierra palm forest (Recher & Recher, 1966) with Euterpe globosa comprising 80% of the canopy. Even at the sum- mit the trees, which are mostly palms, are 20 to 40 feet in height and are more or less openly spaced. The peak is generally windswept and often enveloped in clouds, but not to the same extent as Pico del Oeste. A total of 112 collections were made of the hepatic flora of Mt. Brit- ton; some of the samples were taken from the region around the obser- vation tower on the summit of the mountain, others were from an old road cut somewhat below the summit, and still others were from lower elevations along the Mt. Britton trail above the Loop Road. Among the 88 species found here, the more abundant ones listed in the order of their frequency of occurrence * are Ceratolejeunea maritima, Cyclolejeu- nea convexistipa, Neesioscyphus bicuspidatus, Prionolejeunea aemula, Plagiochila bursata, Cystolejeunea lineata, Bazzania stolonifera, B. hook- eri, Odontoschisma longiflorum, Trichocolea flaccida, Bazzania gracilis, and Trichocolea elliottii. Four species, Symbiezidium transversale, Baz- zania bidens, B. cubensis, and Marchesinia brachiata were found only in this area. Area 9. This area comprises the forest at the summit of El Yunque (Ficure 1). With an elevation just 50 feet higher than Pico del Oeste the microclimate and vegetation of this area are very similar to those of Pico del Oeste. According to Brisco (1966), the mean wind speed on El Yunque is 13 mph, the mean relative humidity is 98%, — the ery global radiation at noon ranges from 0.3 cal. cm~* min~ to 0.8 cal. cm—? min-1, The peak is covered by clouds, mist, and rains most of the time. The vegetation of this peak, as described by Gleason and Cook | = ”), is that of the mossy or elfin forest with four species constituting Si bal the canopy: Weinmannia pinnata, Ocotea spathulata, Eugenta ged quensis, and Tabebuia rigida. Although the overall — fy) vee stunted, twisted trees and dense undergrowth is similar to that descri for Pico del Oeste, there are some differences in the two - as 4 example, the presence of Podocarpus coriaceus on El Yunque (Dansereau, 1966) and its absence on Pico del Oeste (Howard, 1968). apes While most of the 85 samples made in this area were taken along also made on El] Yunque Rock, an exposed granite outcrop eight species of leafy hepatics were ae La of ppmcaps eect or tA ip pat Plaviochila chinant- Syzygiella perfoliata, it * For percent frequency of occurrence see TABLE 442 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 one species, Herberta elliottii, was found here and not in any of the other nine study areas, Area 10. El Toro, the tenth study area, is not only the highest peak he the Luquillo Range with an elevation of 1074 meters, but is also rd farthest inland of the peaks studied (Ficure 1). Because of its higher ele- _ vation and its geographical location, the summit of El Toro is much ies xerophytic than the other peaks. Rains are less frequent, and the St is free of clouds for at least 90% of the daylight hours. The winds, whic blow constantly, and a fairly high solar radiation have a pronounced dry- ing effect on the exposed portion of the peak. ‘At the summit iteelf the vegetation is more like that of a noi scrub than of an elfin forest (Dansereau, 1966). Sarmentose trees am shrubs, among the large rocks and exposed outcroppings devoid of sou, reach a height of only three to eight feet, and the forest lacks the lianas ascular epiphytes so common to El Yunque and Pico del Oeste. Mitropsidium sintenisii, Miconia foveolata, Toralbasia cuneifolia, wee: borinquensis, and Jlex sintenisii occur along with the more common Cyr. : a racemiflora, Tabebuia rigida, Ocotea spathulata, Symplocos micrantha, and Haenianthus salicifolius. ; ‘ct In a few more protected areas below the summit a more typical a elfin forest occurs. The general aspect of this latter forest resembles tha of the summits of El Yunque and Pico del Oeste. At still lower eleva- tions this more or less stunted forest een feet in height. i tation types, montane scrub, elfin forest, palm tor ere sampled in the 78 collections made on El Toro. d, the more abundant, listed in the order of their €a inchoata, Frullania atrata, Syzygiella perfoliata, DISCUSSION In the ten areas treated in this study a total of 156 species of leafy hepatics were encounter, maritima, C. patentissima, Cyclolejeunea convexistipa, Drepanolejeunea Sragilis, and Frullaniq subtilissima were present in all ten study areas * For percent frequency of occurrence see TABLE 1. 1971] FULFORD ET AL., ELFIN FOREST, 15 443 (TABLE 1). Drepanolejeunea inchoata and Taxilejeunea sulphurea were found in nine of the ten areas, while Calypogeia peruviana, Neesioscyphus bicuspidatus, Ceratolejeunea valida, Harpalejeunea heterodonta, Macro- lejeunea subsimplex, Neurolejeunea breutelii, and Prionolejeunea aemula were present in eight. Only these fourteen species, however, appear to be dispersed generally throughout the Luquillo bryoflora, apparently with- out regard to elevation or microclimate. The remaining 142 species have various more restricted distribution patterns. For example, roughly 18% of the species were found only in the rain forest and roadsides of the lower elevations, areas 1 through 6, and not at all in the palm, elfin, or montane forests of the four mountain peaks. These species and their areas of distribution are as follows: * Archilejeunea viridissima — 1 Bazzania eggersiana — 2 Bazzania schwaneckiana — 1 and 2 Calypogeia caespitosa — 3 Ceratolejeunea brevinervis — 3, 5, and 6 Ceratolejeunea cubensis — 2 Ceratolejeunea rubiginosa — 2 Cheilolejeunea decidua — 2 Crossotolejeunea bermudiana — 2 . Crossotolejeunea boryana — 2 . Drepanolejeunea lichenicola — 2 . Euosmolejeunea clausa — § . Frullania exilis — 3 and 4 . Leucosarmentum portoricense — 4, 5, and 6 . Leptolejeunea elliptica — . Lophocolea guadeloupensis — 2 . Lopholejeunea howei — . Lopholejeunea miilleriana — 2 et ee ee ee ht — x ee S. oe S = sc 8 8 aS S d . Plagiochila bicornis — 2 and 6 . Plagiochila distinctifolia — 2, 4, and 5 . Plagiochila tamariscina — 2 . Ptychocoleus polycarpus — 1 . Stictolejeunea squamata — 2 . Symbiezidium laceratum — 4 hd MH Bb bd bo TOO BB Ww ht ed only in the forests on one or more 16 e encounter i fp OL ee ee l in the rain forest below. In this of the mountain peaks and not at al group are included the following: * 1. Bazzania bidens — 8 2. Bazzania cubensis — 3. Bazzania stolonifera — 7, 8, 9, and 10 4. Calypogeia cellulosa — 7, 9, and 10 5. Calypogeia elliottii — 7 * Numbers following each name indicate the areas of distribution of each species. 444 JOURNAL OF THE ARNOLD ARBORETUM [vox. 52 6. Calypogeia lophocoleoides — 10 7. Cephalozia subforficata — 8 and 10 8. Diplasiolejeunea unidentata — 7, 8, and 9 9. Herberta elliotii — 9 10. Jungermannia dominicensis — 8 and 9 11. Lepidozia patens — 7, 9, and 10 12. Leptoscyphus gibbosus — 7 and 9 13. Marchesinia brachiata — 8 14. Micropterygium carinatum — 7, 8, 9, and 10 15. Nowellia dominicensis — 7 16. Nowellia caribbeania — 7 17. Odontoschisma prostratum — 7 18. Plagiochila chinantlana — 7, 8, 9, and 10 19. Plagiochila dominicensis —7 and 10 20. Plagiochila simplex — 8, 9, and 10 21. Prionolejeunea exauriculata — 7 and 10 22. Radula fendleri —7, 8, and 9 23. Symbiezidium transversale — 8 24. Syzygiella rubricaulis — 7 , 9, and 10 25. Trachylejeunea inflexa —7 and 10 Finally, still other species (approximately 20% of the flora) were re- stricted to only one of the ten study areas as noted previously in the de- scriptions. The flora of each of the study areas then, is composed both of wide- spread elements which are more or less characteristic of the entire Lu- quillo bryoflora, and of more specific elements which do reflect the ele- vation, microclimate, and possibly even the geographical isolation of the Area 1 (low elevation roadside) has fairly low similarity coefficients areas, suggesting that it has a more or less distinct ( 3, 4, 5, and 6, all more or less recently disturbed habitats of higher elevation than area 1, appear to possess numerous common species with similarity coefficients of 45.3% to 55.2%. Area 3 while Mt. Britton (area 8) most closely resembles El Toro (area 10) with a similarity coefficient of 69.3% 1971] FULFORD ET AL., ELFIN FOREST, 15 445 Percent Similarity ° 50 60 Area 3 Area 7 26 30 | te: SS CU f the Luquillo with arithmetic area 2, to the rain forest around the La Mina ——— aoe et Magy Mt. Britton Loop; area 4, to the U.S.D.A. La Mina s anna; area 7, to Pico entrance region to El Toro; area 6, to a high elevation “ah oe te weitere. del Oeste; area 8, to Mt. Britton; area 9, to El Yunque; an The dendrogram prepared from a cluster analysis te - amen coefficients substantiates the relationships aig by : — a ae community (Ficure 2). Again area 1 is seen to “aig ine areas. Areas 3, tinct flora as it does not group with any deg pres ane Loop) being 4, 5, and 6 form one cluster group with area 3 (Mt. [voL. 52 JOURNAL OF THE ARNOLD ARBORETUM 446 oe e- Wey “STADIL LT ca = CyOFD Y ‘quapury) vunyaynaiq %6'0 "SIAQIT, (S98N) suapiq pluDss0g AVAOVIZOdIAA'T 097 678 MPL %VI ‘adg (‘Av \L) sytsuad %6'¢ %6'0 IE 60'S aA (“AS) nb ae cans! %S'¢ ‘dg m4qomja V9 WLLL 6S Wirz Beet %0'01 ‘219H (‘ydays) suaSzaarp D449Q43 avaovLaaqayy 97 bet BIZ BI bre 6IT 8 ~6EL %6°9 amorg (-ydas) wa Ssh hi SNYGAISOIS 3A N AVAOVALHOVLOST %9'z sdvios Wo Ac oq" Ot %0'OI %6'°9 ‘ydais ® yor (ads) vpis90¥¢ tie! a ae 4 %S°¢ ed tibiae : y Yass 14410118 Bet: eet. fee ol St St %8'6I ‘ ‘ eo) Oy ydajs pssiftaasq El Lz %8'I ee is pipodoyias AVANVATOOOHIIA T, wad WAN VAN OT 6 8 L 9 s v 3 é Vv surejunopy oymbn’y ay} jo svary Apmig vay, ur avonedazy Ayea’y JO 99u911NnIIQ Jo AJuanba1q “| ATAV FULFORD ET AL., ELFIN FOREST, 15 447 1971] Ve YS lv %8'l %6'0 %0'S MlL°7z Y's Ms'z1 %6'0 %6'0 %I 1 % LOT %6'0 I 11 MV'I e's %8' 91 MBit M11 Me's %L'v1 WMv'Sz %v'z MV %S°9 Sz Sz MSZ Ltt %6'0 %6'°9 Wes ‘ydays maponpa "ydayg (ads) ndijsojako "[YOStg] DIDINUAAD ‘ydayg (Buaidg) psojnyzaa ‘ydayg Cads) pvsozidsana piasodg app) AVAOVIADOdA TV.) Pury “Ss (CaysBuy) pz2unlas DAUDAD]A J, ‘ydais wnyoypxa “Way (AVIt)) twnzDULEDI wni3h£4ajd 0dr PW ‘SINT CMS) seanpndos pizopigajo4a Py ‘quopury suazod mzopigaT sdeios “SIAQIT, (‘MS) Daéafiu01098 %9'¢ ‘SIAL, (11905 y aduieyy) vunryaaunmyos upseg (‘05)) sisuaqna JOURNAL OF THE ARNOLD ARBORETUM [von. 52 448 %9 72 MET %9'7Z 6S %Y7Z %e't %e'I Yr Zz S'S %6'0 %I IT %6'0 %1'I MI M17 VT %1'T Wt %6'O "ads s1j149n09 sisdomy vay AVAOV TAVOIGAANAY “SIAIIT, CMS) wngnaysosdg Yav'b MSE ‘yda}s (ARYL) wnsoygrsu0} Mee ‘wing (saan) wnzopnuap DUSLYISOJUOPE) pz Ct 660. hr! ‘JIN (adg) stsuastnuop sisd onarqoyy AVAOVWSIHOSOLNOGQ) "ydaqs sisuaotuzmop fing viuvaqqiav DEJAN N ‘ZIIF 0IVIY40{ gus %9'e WMA (0D ‘quapury) vyofissp42 %L'T ‘FIN vievag guava pizojvygay AVAIVIZOTVHdd,) Wry Ws'z sb MOS ‘QUO! FY saaN Yuviansad ‘ydais saproajo20yd 07 Wes St -GhLOl ‘quapury “10H vxoy 6 L 9 S b ¢ Z I wIaWAN vary (panutjuo2) sureyunoy oymMbny ayi Jo seary Apmg voy, ut aeonedayy Ayeay JO 29ua11NI9Q Jo Aduanbarq “| ATAV], 449 FULFORD ET AL., ELFIN FOREST, 15 1971] Wey Mest Mel MEI vz Wil MVS N%y'Z S's %6'0 %6'0 %6'0 WS'y %Le VI MES 6e'v MV “ads srsuaomruop DIMUDWAITUN fF AVAOVINNVWAAOND [ ‘ydayg (SeaN) syMDIHAQns "adg (MG) v7oyof4ed Dyaisasag SOG. et 6S aVaOVIZ0NA0'] Yab'b , ye ig | S9ON UDI Mit ‘ydais stsuadnojapons najo20yqoT ‘WHA CARL LL) snsogqus snyqdxasojgay %6'0 uyseg (Cydeis) mouse snyq 980499 H AVANVATOOOHAO’T yt WEE *Wi6 "yng aswaot40z404 UNIUIUWLADS OINAT AVAOVOLLSVWOAOVAV J %e'b “UMTS (1105) Housny 011919O1V AVAQVTTAIAOTY SV ‘ydaig vunyugun sisq007 aVAOVAISdOOZ JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 450 %Y2Z %0'8 ze S72 Wt zz %S'6 LY AD sdeids Lz Writ %6'0 ‘quapury simuayZ %s'¢e “YydaIs PulsuDmD; D6 Bbe2 Bes By %V'I %t's [ACL “L 070;uapiqqns 06 %S'¢ %0'S wind, (Mg) xapduts %Ee'l YZ Mit %6't %zz %$'I 60°01 ‘quapury Suppryns Wel %2"3 %E'”> %®e's Mrz aduiepy 2 ‘quapury srsuarrpmnl %gz %¢'9g ACL L Sisuazmimop WL %¢'9 %0°S %z LT quapury] oyp2tzparp EE %9'¢ %S°¢ ‘quapury orofezzuspsrp %6'0 %2'z s¢ ‘HOD Y “quepury] suapun{uor %0'6 QZ12 6S'b %®E EZ ‘2205) DuDizupury? Co a, eae erm MS1Z *Wsst *®6rer %E'9 %z'Z1 %6S°Z 266'°0 quapuryT (‘ASaq]) 270sang %$'t %S2 %60 ‘Quapury] PuDyaynaesg Wr1's Mv'zZ %8'I %V'1Z %S°L ‘OD suapig YbyzZ WL} HON W edwey srus02g %97 lz %6'b 7S %9'E ‘quepury 970N240 %2e's "ydajs pusomn Ly SZ ‘ung (‘MS) Saptozunrpo Merl 21 %8't %6'r %S$°9 SL 6SE “‘quapury QW ‘wyeyT oydnsgn DILYIOIS DI IVAOVTIIHIOIOVIg MLL hiv 660 %L'E %9'¢ “WIA CYOOH) suaidi2ap snyjunjapy AVIOVHLINVTIACy OL 6 8 r 9 s + ¢£ 4 I wad WAN vary (panuzjuo2) sureumoy oymbn’y ayp Jo seary Apnig ua. ur seonedayy Ajea] yO 29u911NI90 Jo Aouanbary “| wavy, 451 FULFORD ET AL., ELFIN FOREST, 15 1971] %6'¢e MEeOor Wel Mel Yr St MLL MET %9‘7Z %2'S MIV'L yy aa | %vy'Z WI pl Vz I %6'0 %9'¢ %E'9 %B' st %E's Wee Eb 6b %sS ot Mvp'zZ VI %S°Z rv VT %9'€ %8' I %9'e %8't MSZ SL MS‘z %9'S MSL %S°2 %Lt %t'b %0°9 %8'L 9 "ydays (quepury) pwissrprna paunalap1yo4p SUuvAY CUSSIAT) Dz4afuog paunatajojdoup avaovanoalay VL ‘quepury (‘juoW WW SION) Duassyiyqns “quopury 2g “UlYe’] tazuny ‘APL LL sytxa IPPEY sisuarpispaq “wNnd (MS) 070470 DMEDIIN4 J AVAOVINVTTON] ‘dds ‘ydaisg Dgopiyn200s ‘Yydays yofisuoy ‘ydajg xa “}}0r) myaSay ‘ydaysg x9 “3305 vragfus “HOD tapuas DINPON avVaovIAavy ‘ON W adurezy svswaors0zs0d piupd 02> AVAOVINVAVIS [voL. 52 JOURNAL OF THE ARNOLD ARBORETUM 452 092 %V'S Mel %9°7Z MZ iz yan | WSs %6'9Z Wives yz %6' LI %gz suvay (105) suapasan paunalajojIKg %6'0 ‘ydayg (qUOy,) PUuDA10G %G'0 SUvAY DUDIPNU19q paunal ajo, OSSOLD Se 660 B606T 6%FL %o'9 WsV-1AO0f Psoydoyoucya Sti bV2 %0'S sy-yoA0f (‘UYIYIS X9 “W04)) v4as1AD]9 banjo) WezZ suvay vyjundjod MSE ‘adg vnpisap naunal aqojiayD Me bey BMvrtr BS OPS Bit sdesos M212 a oy sie MI %0S %60 suvagy opyoa %8'T bv e's ‘ydais ((101)) vsourds %0°9 ‘"ydays vsoursiqna ett biz Wer Witt Ore WE0l wet . 60 <“9e sueagq (110) yy odurepy) Diussszquaqod Wr, WeLe 686 68TIS Wz BOOw BSiZ Bist B6LI "ydayg (1dg) pe wid rt Gls Bry Mry 69'S er cde Ms" ‘ups (UOT) szsuagna %8'T MV't %Ve1 yds Cquapurs) Bresio? Beh att %$'t suvay pag Cee 6 8 l 9 ¢ ¢ ¢ Z I waAWAN VIAY (panutjuo?) sureyunoy ol[mbn’yT ay} Jo seary Apmis ua, ut seonedazy Ajva’y JO 29U911NIIQ Jo Aguanbarg ‘| aTAVL 453 FULFORD ET AL., ELFIN FOREST, 15 1971} W's Wz! MLL Met Mbv’b Mr'l Met %Oo's Wel Wey zl %YS1 Lv %6'0 MIL %s'l %B't %s'y My et %6'0 Wzez ®S'b Wg'e Weer %6'9T MI >t MY vl My Z Wor Mvy'z EL WzZ1 Hel ey A A %Mz2'z Me°z %S'St %6'O1 %8'l ss %B IY MSs %Lz1 %9'¢e MSZ %O'Ss MS ZI %O'OI %O'Or Moz %S'% %O'OE %O'St Mry've %8'L BLT %LOr Wil “UBS (SIAN) MLADst4} ‘ydais (‘UOJ FY SPN) DNpIIt4 suvagq (‘JUOP BY SPAN) vsnv79 paunalajowmsony sdvios ‘ydayg Cads) vjoomeyoty suvay (‘[AR], 1) pyauniansg ‘ydas (ads) py2unjdoun paunalajound a4(7 ‘UIs (quepUT] paunalajowsnjqid suvayq (‘quapury “WYya']) D7au2) paunalajojsKy suvagq (ads) vdysojoy paunalajoj4hy sueay (‘quepury sueaq (‘ydaqs) ndugsynsuv ae JOURNAL OF THE ARNOLD ARBORETUM [VoL 454 Mery %6'¢ 6S Mrz Mvy'zZ %yzZ %v'6 %s'¢ Lv Br'L Mrs %g'e %V'1 Wit %6'0 %p'Z %6'0 Sb %6'O1 %6'0 %6'0 %6'0 WEES M1 MS Zz MEL %Q'e Mel IE MS ve MSZ %O'S %g'e %6'9 %6'0 %6'0 LT S'S %Y7Z %6'0 “ERMPS (UOJ 3 SION) xaz¢msqns vaunalajo41ov Jy ‘UPIIS (‘WUOP_) DuDaDasvs ‘UBTS (HOD) Duvrsayniu SURAY 2amoy paunalajoygoT Wy tz sueaq (‘quapuryT Q Uys) v¢sv204jUuUvX paunalajomaT “UpTIS DIDINWUDY ‘ugigos (quepury “wy vaiadiya vaunalajojgaT SION (MS) Dany vounalaT %E'rI ‘ds paunalajo4sa ‘ydais pyoutaun SUBAY DINIVGUS SUBAY DzMOpOsLazaYy vaunalaypg Av FT b e c I aaaWAN vary (panuiquo2) sureunoy oppmbn’y aya yo seary Apnig uaz, ur avoneday Ajeay JO 29ua1InI9Q Jo AQuanbai1g ‘| ATAV[, 455 FULFORD ET AL., ELFIN FOREST, 15 1971] Mel ®6'¢ %L‘6E ME'l Wey Wel Mvy‘Z stl %e'8 %6'0 %6'0 M'L WL‘z MV Ee vl 62'L Yst %6'0 %6'0 MES %2'e WT Me vz VV Mv'z %M6'> vy‘ z MULT Ms rz Mit %S"l, %2°z Me VI AHO'Cl VT 6t'z %9'E %8' I %E'l, Mr Or Zt %L°9 MST %0'S %0'S MSL %S7Z %S'e BLT %6'0 %6'0 671 %S°6 Mes Ev %z's S's %L'T SsuBAY (}O4)) PUDOLAI4Aq paunalajoqjIay ‘ups Cydeig) my2aunayos paunalaj]ourK J Woe *‘SIAQIT, (SIAN) sndavakjod snajor0yrkid sdvios SURAT DIDLOUUL WOE SUPA] pyxayinban suvAy (13}05)) DpNWaD paunalajouolsd S99N (MS) srmsofyy snyqunjpydug ‘UBS CqIM) v20;nun] vaunalajozuopa VL suragq (05) magnasg paunalajoina ‘TOSI, DII0U0M %9'¢ suvag (‘UOJ WY SIaN) Suasaazan paunalajo1ni py ‘UBMS vpoty904q DIUISAYIAD JY [voL. 52 JOURNAL OF THE ARNOLD ARBORETUM 456 ET Wo %I7 M's %zS MIL v's %6'0 %3'T %6'LT %8'T M2 %2'y We'v M7 *®WE'S %002 %S°L1 ML Mvr'7Z WES %L'T %9'¢ Sb'b Woz %S'¢e WL M6'b MezZ %s'l %0'S %6'9 %I7Z %S°S S57 ‘ydaisg pxayut paunalajXy Iva J ‘ydaig (‘quepury N ‘wya]) vainydjns suvAgq (‘1dg) nyojnsunsnjzqo paunal apixv J ‘SIAQIT, (MS) appszaasun.g ‘quapurT) pire mnipizargudS "ydaisg pynjoaut vaunalaisdaas ‘ydais (PIM) 7ompnbs paunalajojIu¢ sueayq (‘JUOPT BF S99N) Mogorjcyd 9 $s v e c T (panutjuo7) sureyunoy opmbn’y ay} Jo svary Apnig vay, ur avonedazy Ajeay Jo 29ua1nI9Q Jo AQuanbary “[ ATAVL, aad WAN vauy 1971] FULFORD ET AL., ELFIN FOREST, 15 457 most similar to area 4 (U.S.D.A. La Mina Station), and area 5 (entrance to El Toro trail) most closely resembling area 6 (high elevation savanna). The four mountain peaks (areas 7 through 10) and the La Mina Recrea- tional Area rain forest (area 2) form a second cluster group. In this method of analysis, the hepatic flora of the rain forest (area 2) groups with those of Mt. Britton (area 8) and El Toro (area 10), and the flora of Pico del Oeste (area 7) groups with that of El Yunque (area 9). The two most closely related hepatic floras again are those of Mt. Britton (area 8) and El Toro (area 10) with a percent similarity of 69.3%. From this study it would seem that the composition of the hepatic flora of the Luquillo Range is highly dependent on the microclimate, and to a lesser extent, on the elevation and geographical isolation of the re- gion in question. The similarities of the microclimate in the rain forest proper (area 2) and the palm forests of Mt. Britton (area 8) and El Toro (area 10) might indeed explain why the hepatic flora of area 2 is more like those of areas 8 and 10 than those of its neighboring areas, ie., areas 3, 4, 5, and 6. It may be that the hepatic flora of Pico del Oeste is most like that of El Yunque because the microclimate of Pico del Oeste is more like that found on El Yunque than that found on either Mt. Britton or El Toro. The fact that area 6 (high elevation savanna) has 52.2% of its species in common with El Toro (area 10) may be re- lated to the fact that some of the collections made in area 10 were made along a dry grassy ridge very much like the savanna of area 6, and the close relationship between the floras of Mt. Britton and El Toro may well have its basis in the similar palm forests found on both. Characteristics of the microhabitat alone, however, cannot explain Com pletely the differences which occur in the hepatic flora in different parts of the Luquillo Range. Certainly, the restriction of species to one or : few specialized regions must be related to the dispersal mechanisms 0 TABLE 2. Coefficients of Community for Leafy —— Floras in Ten Study Areas of the Luquillo Mountains * AREANo. 1 2 3 4 5 6 7 8 ete 1 2 37.2 3 39.0. 514 4 441 339 S52 5 41.0 482° "455 pe - 6 40.0 388 SiO 4 7 25.3 37.0 43.5 34.8 34.2 48.5 : 8 36.5 66.7. S17 886° Be Se a 1 . 253 44.2 52.0 38.0 43.7 48.6 oh ie ae 10 305 588 61.6 40.7 41.9 52.2 67. : hey ae , ted following e coefficients represent percent species similarity and were calcula se Wy Sgrensen’s (1948) method. 458 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 the plants as well as to differences in elevation and climate. Just how much the composition of the hepatic flora is dependent on dispersal and how much it is dependent on climatic conditions remains problematical. LITERATURE CITED BaynTon, H. W. The ecology of an elfin forest in Puerto Rico, 2. The micro- climate of Pico del Oeste. Jour. Arnold Arb. 49: 419-430, 8 fig. 3 tab. 1968; 3. Hilltop and forest influences on the microclimate of Pico del Oeste. Ibid. 50: 80-92. 7 fig. 1 tab. 1969. Briscor, C. B. Weather in the Luquillo Mountains of Puerto Rico. Forest Service Res. Pap. ITF-3, Inst. Trop. Forestry, Rio Piedras, Puerto Rico. 1966. DANSEREAU, P. Studies on the vegetation of Puerto Rico, 1. Description ‘“ integration of the plant-communities. Inst. Caribbean Sci. Spec. Publ. pp. 1-45. Mayagiiez. 1966. Futrorp, M., B. CRANDALL, & R. STOTLER. The ecology of an elfin forest in Puerto Rico, 11. a leafy de tttag of Pico del Oeste. Jour. Arnold Arb. 51: 56-69. 13 fig. Gates, D. M. The aati of an elfin forest in Puerto Rico, 4. Transpiration tates and temperatures of leaves in a cool humid environment. Jour. Arnold Arb, 50: 93-98. 2 jig. 1969. Gieason, H. A., & M. T. Coox. Plant ecology of Porto Rien Pts. 1 and 2. Sci. Surv. Porto Rico Virgin Islands 7: 1-173. 50 pl. Howarp, R. A. The ecology of an elfin forest in Puerto Rico, 4 ae ai and composition studies. Jour. Arnold Arb. 49; 318-418. 14 fig. Lyrorp, W. H. The ecology of an elfin forest in Puerto Rico, 7. a roo and earthworm relationships. Jour. Arnold Arb. 50: 210-224. 7 fig. REcHER, H. F., & J. T. RECHER A contribution to the knowledge of “ihe avi- fauna of the Sierra de Luauillo, Puerto Rico. Carib. Jour. Sci. 6: 151-156. 966. P. H. A. Snearu. Principles # Saencl Taxonomy. 359 pp. Pa Freeman, San Francisco & London. SORENSEN, T. A method of establishing “tic _ equal amplitude in plant sociology based on similarity of species content. Danske Vid. Selsk. Biol. Meddel. 5: 1-34, 1948. M. F.] DEPARTMENT OF vasa art UNIVERSITY oF CINCIN CINCINNATI, OHIO 15201 [B.C. and R.S.] DEPARTMENT OF Botany SOUTHERN ILLINOIS UNIVERSITY CARBONDALE, ILLINOIS 62901 1971] TOMLINSON & FISHER, CORDYLINE, I 459 MORPHOLOGICAL STUDIES IN CORDYLINE (AGAVACEAE) I. INTRODUCTION AND GENERAL MORPHOLOGY P. B, ToMLINSON ANp J. B. FISHER THIS PAPER PRESENTS THE RESULTS of an investigation into the habit and vegetative morphology of the genus Cordyline which was undertaken as a preliminary to anatomical, developmental, and experimental studies. Later papers will restrict the enquiry to Cordyline terminalis, which is typical for the whole genus in most respects but is a particularly useful species for experiment. These studies will be a further contribution to the investigation of tree-like monocotyledons, particularly those within the Agavaceae which have secondary growth (Tomlinson & Zimmermann, 1969; Zimmermann & Tomlinson, 1969; 1970). The first of these papers summarized much of the extensive earlier literature on these plants. Information specific to Cordyline is quite limited, however, and in par- ticular the post-seedling phase of development in this genus (even though it was illustrated as early as 1852 by Gaudichaud-Beaupré), has never been described fully. In making good this deficiency we will show how an understanding of this phase of individual development is necessary in explaining the growth habit of Cordyline in possible evolutionary terms, TAXONOMY Earlier botanical literature does not distinguish clearly between C ordy- line and the closely related Dracaena, which is understandable in view of their similar habit. Many species of the former genus were described within the latter. However, more recently accumulated information re- veals that Cordyline is distinct from Dracaena in several exomorphic features of which fruit structure is the most constant and readily demon- strated (Brown, 1914). : The two genera are also clearly circumscribed in their vegetative anat- bes shown in considerable detail by Rothert (1909). Further investigations continue to support these distinctions. Secondary a tissues are present in the stems of both genera, but only e Dracaene ne it also develop in the roots. Cordyline has a distinct oe Phology, described below, for which there is no equivalent in racae - Leaf anatomy in Dracaena is distinctive because of the unusua ere ment of vascular bundles, as has been confirmed more recently by saat derlich (1950). Major differences between the two genera are summari2 in Taste I. n the two genera is otherwise accepted 460 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 TABLE I. Major diagnostic differences between Cordyline and Dracaena CoRDYLINE DRACAENA Ovules many (2 or more) per loculus Ovules one per loculus Fruits with several small seeds Fruits with 1-3 large seeds Germination epigeous (e.g. C. aus- Germination hypogeous tralis) or hypogeous (e.g. C. ter- minalis ) Post-seedling stage rhizomatous Post-seedling stage not rhizomatous Roots without secondary tissue Roots able to develop secondary tissue Primary vascular bundles of stem amphivasal Vascular bundles of leaf normally* orientated (xylem adaxial) Primary vascular bundles of stem collatera Some vascular bundles of leaf always horizontally orientated* (xylem an phloem perpendicular to normal orientation) Crystal cells (other than Crystal cells (other than raphides) present raphides) absent Ultimate inflorescence unit Ultimate inflorescence unit often a solitary flower a cincinnate cluster Bracteole+ bifid (often Bracteole+ not bifid (evidently apparently two) single) * We verified this differen ce by comparing leaves of Cordyline with leaves from several species of D racaena cultivated at Fairchild Tropical Garden. nates as a single encircling structure x d by subsequent growth of the flower it encloses (see p. 475)- This suggests that bracteole morphology is not fundamentally different in the two genera. umber of species j . ; ently understood, includ peciés it contains. Cordyline, as pres New Zealand is 0 be distinct and 1971] TOMLINSON & FISHER, CORDYLINE, I 461 species, C. dracaenoides in South America and two more in eastern Aus- tralia. In New Guinea existing collections do no more than hint at poorly understood forms. Other species are described from Norfolk Island, Mauritius, the Mascarenes, and New Caledonia. These in part form the genus Cohnia. Hybrids between natural populations of distinct species in New Zealand have been described (Cockayne & Allan, 1934). One species which offers particular difficulties is the familiarly cultivated Cordyline terminalis which seems to have a wide natural distribution in the Polynesian area, a range which has been so much extended by use of the plant for food, fiber, medicine, and as an ornamental that its original home has become obscure. Many cultivars are known and this has lead to nomenclatural problems where the more striking varieties have been given specific names. The solution to all these taxonomic complexities lies outside the scope of this present work, although it should be mentioned that there are important diagnostic features of leaf anatomy which may be of ultimate value in clarifying the situation (Tomlinson, unpublished). None of these problems is discussed further because we are dealing only with a limited number of the more clearly circumscribed forms. The background of morphological information we provide should, however, prove helpful in the future taxonomic revision which is still necessary. MATERIAL AND METHODS Only a minority of species has been investigated, but in view of the naturalness of Cordyline it is reasonable to assume that observations are applicable to the genus as a whole. The most extensive investigations were carried out on natural populations of the four commonest —_— native to New Zealand (Cordyline australis, C. banksii, C. indivisa, and C. pumilio) together with varieties of C. stricta, an Australian es cultivated in Auckland. Additional material both of wild and cultivated specimens from Fiji and New Guinea, referable to C. terminalis has been examined. More extensive observations were also made on old specimens representing several cultivars of C. terminalis grown In the Pa cha Early stages of seedling development of C. australis and C. sae sit ee observed in plants grown in pots in a greenhouse from seeds fe 8 i the wild. Studies of post-seedling stages have been made on p are 3 a wide range of sizes dug up in the wild. For studies of mgd sag Pie in C. terminalis, seeds from cultivated plants were sown and a - on fed seedlings were raised. Localities and sources of material are lis 1 tud APPENDIX. Plants for continuing anatomical and cman ; iin. have been preserved either in formalin-acetic-alcohol or 4 pare sg - Habit drawings in the figures were all made from either fresh or The fifth species, C. kaspar, is very rare since it is endemic - ~ ae aoe Islands just off the northern tip of the North Island of New Ze see Night, 1643 were discovered and named by Abel Tasman on the eve of phe idered more (Oliver, 1956). Presumably the particular specific name kaspar was cons euphonious than melchior or balthazar! 462 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 preserved specimens. Photographs were used to produce the habit dia- grams of mature specimens shown in Ficures 12 to 18. cotyledon __,cotyledonary onde mnt plumular leaves S 1-4. Cordyline australis. : 1 e Seedling: 1, early stages in germ ination; : t x aves = 9 extended but carved: right, at stage with sible and cotyledon erect, all X 3: 2 = e with cotyledon modes unextended: paige Sheeateican (detail rag 1971] TOMLINSON & FISHER, CORDYLINE. I 463 SEEDLINGS The genus Cordyline is unusual amongst monocotyledons because ger- mination may be either epigeous or hypogeous, depending on the species. For each species the type of germination seems constant. Epigeous germi- nation has been observed in C. australis and C. banksii ; hypogeous germination in C. terminalis, Epigeous. Germination of seeds and early stages of seedling growth were studied in Cordyline australis and C. banksii. The two species seem identical in all respects (FicuREs 1 to 6). Germination is epigeous, the slender cotyledon becomes erect and green, usually carrying the seed with it (FicureEs 1, 2). Extension of the radicle precedes that of the cotyledon. The hypocotyl is always elongated up to a length of 5 mm., and since it remains glabrous it is clearly delimited from the radicle which has a dense covering of short hairs. The cotyledonary node becomes somewhat swollen and marks the distinct upper limit of the hypocotyl (Ficures 2, 3). The first adventitious root normally may appear either at this node (Ficure 4) or at the base of the hypocotyl (Ficure 3). The plumular leaves are initially enclosed by the base of the cotyledon but later grow through the mouth of the cotyledonary sheath (FicurE 1). Internodes between early leaves may remain short (FicurE 3) but are successively wider in the manner characteristic of monocotyledons in their “establishment” phase of growth (Tomlinson & Zimmermann, 1966). Frequently, however, one Or more internodes above the hypocotyl becomes extended in varying degrees as is shown in Ficure 4. This produces a long seedling axis and may be a mechanism for adjusting the depth in the soil at which the plant becomes established. Hypogeous. Cord line terminalis is distinctive in that germination is hypogeous, the greater part of the cotyledon remaining within the seed and both remaining beneath the soil surface. Details will be presented in the second paper of this series since it deals solely with this species. ia All seedlings are alike in subsequent stages of growth. oe roots in seedlings with extended internodes may or may not be restrict to nodes. This restriction is not obvious in later stages because internodes are congested. POST-SEEDLING DEVELOPMENT General morphology. Further development continues = ? aa nt shorter period, in a way normal for large woody monocotyledons. umivel are successively larger (longer and wider) and the axis ee hy raabined by establishment growth. The resulting axis is bluntly —, eee 5, seedling axis). Leaves each subtend a bud although t ae u ime strongly inhibited. There is then an abrupt change oe Pe Aiton a way which represents a marked departure from normal monocoty 464 JOURNAL OF THE ARNOLD ARBORETUM [vox 52 C.australis ; ; : bit Ficures 5 and 6. Cordyline australis. Origin of seedling rhizome Fac te of seedling, x 1/4, enlargement of axis showing rhizome bud isc erae axis to ing axis, X 3; 6, habit of seedling, x 1/6, enlargement of rgeneey d), show well-developed r izome (the Parental seedling axis almost o ~*~ . id seedling growth. One of these axillary buds develops as a shoot co growth, at first obliquely outward and then vertically downward ( leaves. uces only fleshy scale-leaves, never foliage le who called it a y, and it is part of the la the underground stolon.” This distinction is ae tger problem involved in ee ) Parts of monocotyledons (cf. Tomlinson, 1970; p. 1971] TOMLINSON & FISHER, CORDYLINE, I 465 Subsequent growth of the seedling rhizome is so rapid that the existing axis is displaced laterally. It may persist as a more or less distinct append- age. The age and thus the size of the seedling on which the rhizome is initiated varies appreciably and this may account for the variation in size of this appendage itself. Ficures 9 to 11 show the base of three different C. pumitio 7 C.austratis — (adult) (post-seed|ing) 3 % Pitre imac Z age S oe tan, : a aA -_ Caustralis (post-seedling) qustralis: 7, C. pumilio, adult speci- wares pe a single rhizome, X 1/12; 8, : transition hs fe dling 142: Sige of sepaleny re terns | australis, details ¥ ope : : : at ing si x igs actge™ thizomes with original seedling re) by question — zs ageeas tvaet in 4 . inverted. It is po 7 tinct but small: 11, seedling axis a dad and has been replaced 466 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 saplings of Cordyline australis of approximately equal size, the right-hand one (FicuRE 11, but see legend) with a conspicuous inverted seedling axis, the left-hand one with the original axis only tentatively identified by the question mark (Ficure 9) and the middle one recognizable but small (arrow, Ficure 10). In subsequent growth the sapling retains a single axis without obvious indication that the lower part is a branch of the upper part (Ficure 8). Both axes grow in diameter at about the same rate. Branching of the seedling axis by development of more than one rhizome bud is possible but the only examples we have seen have been in C. terminalis. Branching of the underground parts in later stages of growth seems to involve branching of the rhizome itself. This seems more common in some species (e.g. C. pumilio, C. terminalis) than in others (C. australis, C. banksit) Growth in length of the rhizome at first much exceeds that of the aérial stem. Elongation of the aérial stem does not begin until the axis has reached a diameter of 1 to 2 cm. when the rhizome may be 15 cm. long. Subsequently this aérial axis is the dominant part. Rapid growth of the thizome in early stages of sapling development is indicated by its smooth white surface and obvious fleshy scale leaves (FicuRE 6) but later the surface becomes wrinkled because cork develops to within a few milli- meters of the scaly apical portion (Ficures 9-11). We have no infor- mation about the maximum length to which a rhizome may grow, but in saplings with 2 to 3 m. of visible trunk the rhizome is scarcely longer than Ina well-established seedling with the leaf rosette still at soil level. Rhizome and aérial stem continue their overall diameter growth at growth of the crown is long contin i : ued. The internal features which demonstrate this are described below (p. 468), g axis of Cordyline is unusual, if not unique a shoot apex at upper and lower extremity, g : co mstanices, Buds originate singly hey be bavi: scale leaves and persist indefinitely although y. rongly inhibited. Since they are primary in origin thelr trace Fy vascular system in the same way as has 1971] TOMLINSON & FISHER, CORDYLINE, I 467 been described for Dracaena (Zimmermann & Tomlinson, 1970). Sup- pressed buds develop into sprouts on stumps of damaged or felled trees, Later papers will indicate how growth of buds is determined and expressed in Cordyline terminalis. 6). Root primordia are also produced in larger numbers over the whole Phyllotaxis. As is usual throughout the monocotyledons leaf arrange- ment in the seedling is distichous, but there is a rapid change to a spiro- distichy in which the angular divergence between successive leaves is somewhat less (or more) than 180°. By the time the seedling rhizome is well established the original distichy is obscured by the pronounced helix of the two “orthostichies.” Subsequent development may take either one of two possible courses depending on the species. In the first group of species a true spiral is established, with divergences of the order of 3 /8 or higher, as in Cordyline australis, C. indivisa, and C. kaspar. Otherwise, in the second group, the spirodistichy persists as a feature of the adult plant as in C, banksii, C. pumilio, C. stricta, and C. terminalis. In these species leaf arrangement may be described as a‘persistent juvenile feature, in much the same way as distichy in other woody monocotyledons can re beyond the seedling phase, notably in Strelitziaceae, but also in a lew palms. This feature is obvious in such plants, because the poe is Precise, whereas in Cordyline it requires close examination to show t i there are two ranks of leaves. Spirodistichy is otherwise seen as a abe e feature in palm seedlings. It should be emphasized that the reg ee spirodistichy, with two “orthostichies” to a phyllotaxis wit nN orthostichies involves a major change, namely the addition of ee new orthostichies to the system of leaves. Cordyline thus offers iat — material for the investigation of changing phyllotactic patterns along Single shoots in monocotyledons. Initiation of secondary tissues. Since Cordyline develops tina tissues it is of some interest to know at what stage of paper gees : Seedling axis initiates secondary tissue and what relation this esate rhizome initiation. Secondary tissues can be recognized in so — section by their uniformly amphivasal bundles dati ao! six to radial arrangement of ground tissue, and the se damiectatta ee ermann 4 meristematic cambial zone (Cheadle, 1937; Tomlinson & damm 468 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 adventitious roots which connect to both primary and secondary ae These preliminary observations, however, do show that secondary v tissues are initiated prior to the development of the rhizome. ; ‘ail n The secondary tissues of mature rhizomes are much less lignified tha those of aérial stems the di . In this sense establishment (but not mere is dependent on the development of secondary : his brief analysis of the seedling axis is that it 1S a suitable experimental subject. HABIT Since earl habit of each fn naturally grown trees so that saplings ranched. Buds are inconspicuous since 1971] TOMLINSON & FISHER, CORDYLINE, I 469 TABLE ITI. Distribution of primary and secondary tissues in saplings of Cordyline* A. CORDYLINE AUSTRALIS LEVEL ABOVE DIAMETER PRIMARY AVERAGE WIDTH OF TOTAL STEM SOIL SURFACE VASCULAR CYLINDER SECONDARY TISSUE+ DIAMETER 0 2 62.5 137.5 5 7 42.5 EL Beas 28 10 S15 95 175 1g o25 88 255 18 re Pe 82.5 290 22 2135 75 390 27 14.0 66 620 31 ce: 56 B CORDYLINE BANKSII 0 8 61 140 0.5 12 49 120 17 19 43 115 5 23.5 ol 95 54.5 27 25.9 87.5 83 29.5 22 83 113.5 31.5 19 80 145.5 34 14 72.5 176 37.5 14 75 * Measurements in mm. 7 Average width of cortex is 5 mm. they consist of a prophyll enclosing a few primordia, the whole organ embedded in the surface tissues of the parent stem, but continuous with the primary vascular cylinder via a few traces which remain largely at a procambial stage of differentiation. Branching is usually the result of flowering although branching in the absence of flowering is not uncommon. When the axis produces a terminal inflorescence a lateral bud close beneath the inflorescence grows out and replaces the previous terminal bud (Ficure 26). This method of sympodial growth is normal for many woody monocotyledons (Schoute, 1903; Tomlinson & Zimmermann, 1969; Tomlinson et al., 1970). So long as only one bud grows out, the sym- podium remains simple. This is the usual condition in Cordyline indivisa, whence its name (Ficure 16), although in a strict morphological sense, the name indivisa is misapplied, since the axis, in fact, branches each time It flowers. Mature specimens of this species do occasionally produce divided trunks with more than one crown. : In other species more than one bud may grow out beneath an inflo- rescence so that the axis as a whole may bifurcate. Three or even four branches are less common beneath a single inflorescence. Forked axes are not dichotomous although there are some misleading reports which suggest this (for a discussion see Tomlinson et al., 1970). 470 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 it sketches of aérial parts, not to of the general appearance of the plant, the leaf, and a silhouette of the in- ' ther rom herbarium specimens or photographs. 12, terminalis; 18, C. australis ng al 15, C. pumilio; 16, C. indivisa; 17, C. to right and left. » With illustrations of two successive sapling stages oe ecieips fork with different frequency. Cordyline australis produce a oa alse ar (Ficure 14) branch relatively frequently and branched (inset fi nee Wn. The saplings are comparatively little- ingly and even nis i Ficure 18), Cordyline banksii branches spat- 12) This waa : sii has a slender, little-branched crown (FIGURE Species al ‘ di e date to the shaded forest habitat it occupies. also differ in their tendency to branch from the base. Cordyline 1971] TOMLINSON & FISHER, CORDYLINE, I 471 australis and banksii normally retain a single trunk and are tree-like. Cordyline terminalis is more of a shrub since it branches frequently from the base, the aérial shoots in old specimens becoming quite diffuse (F1icurr 17). Cordyline stricta is somewhat intermediate between these two ex- tremes and may or may not retain a single trunk (Ficure 13). The illus- tration of C. kaspar (Ficure 14) is of a cultivated specimen, the only mature one we have seen. The ability of older plants to branch from the base reflects a tendency for basal buds to develop. This is possible because branching may occur in the absence of flowering, buds being released from inhibition because they are remote from an inhibiting shoot. This type of branching can also occur in distal parts and produces forking stems which resemble sympodial branches. However, monopodial forks can be distinguished in their anat- omy from sympodial forks because primary vascular tissues of branch and parent axis are not obviously continuous in the former, whereas they are in the latter (Tomlinson e¢ al., 1970). Cordyline pumilio is a low-growing herb (FicuRE 15) with one or more rosettes of linear leaves, each rosette supported by a short stem. Under- ground parts of C. pumilio include one or more descending rhizomes which branch readily to produce both new rhizomes and new erect shoots, ac- counting for the clustered habit of older specimens. In this habit it is seemingly anomalous, compared with the shrubs or trees which otherwise characterize the genus. The anomaly can be explained if C. pumilio is equated with the post-seedling stage of development of other, larger species (cf. Ficurrs 7 and 8). This suggests that C. pumilio is a neotenous form, flowering in the post-seedling stage of development which then persists by branching and without ever forming a tall trunk. It is interesting to speculate to which other species C. pumilio is most closely related. The fact that leaves of C. pumilio correspond more closely in their anatomy to leaves of seedlings of C. australis of approximately the same size (as in Figures 19 and 23) than to seedling leaves of any other species may be suggestive. The hypothesis that C. pumilio represents a orga form of C. australis could be tested by further observation and pe designed to make C. australis flower precociously, or to prevent C. pumiio from flowering. The above ee describes habit in Cordyline in only a very gs Way. In a subsequent paper we will describe the development * me shoot of C. terminalis in more detail with especial reference to the dis bution of leafy and rhizomatous shoots. LEAF MORPHOLOGY Foliage leaves in seedlings of all species tend to be pester | alike, but diverge with age in different species. Leaves of Coray' i Sis visa show the most extreme divergence from the seedling pl Saatly ‘0 of leaves representing the change from seedling to sapling an a he adult is shown for C. australis in Ficure 19. The close resem 472 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 ies: representative leaves from adult plants, ; transverse sections of leaves, taken at the ut C. pumilio. Note that the leaf of C. indi- B oO p a oO n ° Bh QQ ea 79 & Lo) _ Q So & o > a 8 be QQ = & a dR wW Q ~~ = wm, om > i . pumilio cf. se 24, C. kaspar; 25, C. terminalis, adult ustralis in Ficure 19): tivar, from New Guinea populations. ra ves, each representing a different cul 1971] TOMLINSON & FISHER, CORDYLINE, I 473 tween the adult leaves of C. pumilio (Ficure 23) and the seedling leaf of C. australis is obvious. Adult leaves of other species studied, representing quite fully the range throughout the genus as a whole, are shown in Ficures 20 to 25. The largest adult leaves are those of C. indivisa (rep- resented in FIGURE 22 on a smaller scale than the leaves of other species) and may exceed a meter in length. They develop a thickened fleshy base. Leaves of other species differ largely in the width of the blade and in whether they have a conspicuous grooved petiole (e.g. FicuRE 21) or not (e.g. Ficure 24). Leaf shape differs considerably in different cultivars of C. terminalis; the series shown in FicureE 25 is only a small sample but still shows almost the complete range of adult leaf shape for the genus as a whole. Varieties of C. stricta also differ markedly in the morphology of the adult leaf (Ficure 20). This range of variation in both species includes leaves with and without grooved petioles. Leaves of Cordyline are of interest to the comparative morphologist because, within the one genus, they show a range from linear, undifferen- tiated leaves to those in which there is a sharp differentiation between petiole and blade. However, a pronounced sheathing base is never devel- oped so that although the leaf insertion is encircling it is never over- wrapping. FLOWERING A brief description of inflorescence morphology and structure will serve to introduce the topic, although it is largely outside the scope of the present study. Flowering of Cordyline in New Zealand is clearly seasonal, from October to December, and seasonal flowering may be universal in the genus. Inflorescence morphology. Inflorescences are terminal panicles with the same simple principles which underlie the construction of flowering axes in all woody monocotyledons. In the transition to the reproductive stage the main axis of the shoot elongates (FicurE 26) kg ee rapid transition from normal foliage leaves to reduced leaves (bracts) which subtend the first-order branches. These branches repeat o ta ture of the main axis (FicuREsS 27 to 29) often with branches of g apn 1 but rarely of a third order. Bracts are progressively smaller in a dista -order directi i i hes. The first leaf on each first-or oeesertenlermachere ties nserted basally. The prophyll both prophyllar buds may grow out so that app branches may develop; this tends to be a character inflorescence. of the expanded Parent shoot. This shape is represented figures for each species in the habit drawings (F1cu 474 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 In Cordyline australis (FicuRE 18), C. indivisa (Ficure 16), and C. kaspar (FicurE 14) the inflorescence is much branched and congested. In C, banksti (Ficure 12), C. stricta (FicurE 13), and C. terminalis (Ficure 17) the inflorescence is little branched and lax. Cordyline pumilio (Ficure 15) has the most lax inflorescence and the longest peduncle, characters no doubt related to the habit of the species. Inflorescences of all these species are more or less erect initially but become pendulous with age, especially in C. indivisa (F1cuRE 16). The inflorescence remains erect in C. pumilio, The arrangement of first-order branches on the inflorescence continues the leaf arrangement of the vegetative shoot, so that it is spirodistichous in one group of species (banksii, pumilio, stricta, and terminalis) and spiral in the other (australis, indivisa, kaspar). This, in part, accounts for the difference in density of inflorescence branches in the two groups. Floral morphology. Flowers occur singly in the axil of a bract (Fic- URES 30 to 37). Each is subtended by a bracteole which is continuous and 2-keeled around the flower bud (Ficures 30 to 32) but ruptures as the flower expands (e.g., Ficure 36). Brown (1914) described the two split halves of the bracteole as two separate structures and emphasized this as a character diagnostic for Cordyline. Study of flower development in C. : australis shows that this interpretation is incorrect because the bracteole Is initiated as a single adaxial organ. Flowers of four different species are shown in Ficures 33 to 37. The basic construction is uniform (FIGURE 32) but different species show slight quantitative differences. The relative lengths of members of the inner and outer perianth whorls, together with the length of the perianth tube have been accorded major diagnostic significance by earlier authors (e.g., Baker, 1875; Brown, 1914) but these characters seem difficult to apply and often are not very clear cut. Flowers are white or streaked with lilac in most species, but blue in C. stricta. DISCUSSION develop a tap root i in monocotyledons with age. The only *g0 secondary thickening, although tap roots are not ere are older descriptions of rhizomes in Dracaena australis FY 26-29. C Salat if Cordyline species, flower. early stage of . is, young inflorescence: 26, entire terminal inflorescence at an evelopment, dissected from the leafy crown, X 1/2; the con- i at X: 27, one first order mo : i Oganengl ae young flower bud enclosed b , with flower bud removed, X 3; 32, floral diagram, wi i. flower, X 3; 33, longitudina bl ack, bracteole hatched. 33 and 34, C. banksii, fi inal from the side, X 3: 35, C. 476 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 by Goebel (1905) and Lindinger (1908). Comparable examples are found in Dioscorea and some species of Costus (Holttum, 1955). A The sapling stage of development of Cordyline (Ficure 8) also F ers the interesting situation of an axis with a shoot apex at each en ] an unusual structure which has been commented upon by T roll (1937; p. 722). The hormonal balance in this axis is undoubtedly peculiar. In the continuing investigation of Cord yline, structure and developmen of the vascular system in the axis are being examined. A comparative survey of leaf anatomy may help in solving some of the evident pce puzzles in the genus. Cordyline terminalis has been selected as a subjec for detailed study because it offers itself as an object for morphogenetic experiment. The questions to be answered include: under what circum- stances does a bud develop as a rhizome or a leafy shoot? Under what circumstances does a rhizome become transformed into a leafy — What determines the distribution of secondary tissues in these organ Future reports will describe the variety of shoot expression in this species and the results of surgical and hormonal experiments which throw some light on these problems. SUMMARY An introduction to the morphology and development of the genus C or- dyline (Agavaceae) is made with reference to seven species, mostly native to New Zealand, which have been studied as natural populations. Ger- mination of seeds may be epigeous or hypogeous, depending on the species, and is followed by the growth of a positively geotropic rhizome from an axillary bud on the seedling axis. This replaces the first-formed part of the seedling axis. In this way a structure which has the function of a tap root is developed from an organ which is morphologically a stem. The subsequent diversity of habit in the adult stage of different species is de- scribed in relation to this seedling whereupon it is demonstrated that C. pumilio, a dwarf species, may be regarded as a neotenous form of larger, descriptive and experimental studies on the morphology and anatomy of the genus. ACKNOWLEDGMENTS This research is phology of the monocotyledons, su otanical Illustrat irchi ical Garden. For laboratory faciliti _ustrator at Fairchild Tropica to cultivated specimen 1971] TOMLINSON & FISHER, CORDYLINE, I 477 Professor V. J. Chapman and Associate Professor L. H. Millener, Univer- sity of Auckland, New Zealand; the Director, D.S.I.R., Mount Albert, Auckland, New Zealand: the Assistant Director, ‘Dominion Museum, Auck- land, New Zealand; Mr. J. S. Womersley, Chief of Division of Botany, Lae, Territory of New Guinea; Mr. John Parham, Department of Agri- culture, Suva, Fiji. Permission to study Cordyline indivisa in the Ton- gariro National Park, New Zealand was kindly given by the Chief Ranger. APPENDIX SOURCES OF MATERIAL OF CORDYLINE EXAMINED: C. australis Waitakere Hills and Bethells Beach Road, North Island, New Zealand C. banksii Waitakere Hills, North Island, New Zealan C. indivisa Tongariro National Park, North Island, New Zealand C. kaspar Cultivated, Auckland, North Island, New Zealand C. pumilio Whangaparoa Peninsula, North Island, New Zealand C. stricta Cultivated, Auckland, North Island, New Zealand C. terminalis Cultivated, Miami, Florida, U.S.A., together with populations in New Guinea and Fiji LITERATURE CITED ArBER, A. 1925. Monocotyledons —a morphological study. Cambridge Uni- versity go. Cambrid Baker, J. G. 1875. Revision of the genera and species of Asparagaceae. Jour. Linn, ie (Bot.) 14: 508-632 i ces Brown, N. E. 1914, Notes on the genera Cordyline, Dracaena, Pleomele, San- seviera and Taetsia. Kew Bull. 1914: 273-279. ; Sake Fee . CHEADLE, V. I. 1937. Secondary growth by means of a thickening ring in certain monocotyledons. Bot. Gaz. 98: 535-555. Cockayne, L. & H. H. ALtan. 1934. An annotated list of ae of wild hy- brids in the New Zealand flora. Ann. Bot. (Lon GANDICHAUD-BEAupRE, C. 1852. Voyage pri - monde sur la corvette La Bonité. Bot. Atlas (1841-52) livr. 23. pl. 1 GorseL, K. 1905. Morphologische und sare a - ‘ wo Knolle der Dioscoreen und die Wurzeltrager der eer oe welche zwischen Wurzeln und Sprosse stehen. Flora 95: 167-21 ae Hotttum, R. E. 1955. meaner oe Me monocotyledons — varia theme. Phytomorphology 5: 399-4 : Hutcuinson, l. 1959. The families of pase plants. Vol. II. Monocotyle dons. ed. 2. Clarendon Press, Oxford. Pflan- Krausse, K. 1930. Liliaceae in ENGLER, sae & Prantl, K. Bd. 15a. Nat. Pfla zenfam m. ed. 2. Wilhelm Engelmann, Leipzig. LinvINcER, L. 1908. Die sekundiren Adventivwurzeln = ee es morphologische Wert der Stigmarien. Jahrb. Hamburg w Beihefte: 59-88. Oxrven, W. R 8. 1956. The Three Kings Cabbage Tree. Rec. Auckland Inst. Mas. 4: 381, 382. 478 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 RotHert, W. 1909. Uber die anatomischen Differenzen der bape es Dra- caena und Cordyline. Bull. Dept. Agric. Indes néerl. 24. 13. — we C. 1903. Die Stammesbildung der Monokotylen. rie (Jena) 92: 32- eae P. B. 1970. anes aeons towards an understanding of their morpho ology and anatomy. R. D. Preston, ed. Recent Advances in Botany vol. III. 207-292. hate Press, New Yor . ZIMMERMANN. 1966. Anatomy of the palm Rhapis excelsa. III. Juvenile phase. Jour. Arnold Arb. 46: 301-312. & - The “wood” of monocotyledons. Bull. Int. Assoc. hat Anatomists, 1967(2): 4-24. ——. 1969. Vascular anatomy of eye dame with secondary am introduction. Jour. Arnold Arb. 50: 159- ; & P. G. Smpson. 1970. Dichotomous if Pseudodihotomo branching in mnonpcotyledonous trees. Phytomorphology 20: TROLL, W. 1937. S gers Morphologie der hoheren Pflanzen. 1 Bd. Vegetationsorgane. 1 Teil. Gebriider Borntraeger. Berlin. Mege tiore, J. 1907. Hae ah Morphologie der Pflanzen. 2 Teil. Fr. ivnaé, Prague. Womans a 1950. Die Agavaceae Hutchinsons im Lichte ihrer Em- bryologie, ihres ibaa Staubblatt- und Blattbaues. Osterr. Bot. Zeitschr. 97: 437- senor M. H.,, PP. B. TOMLINSON. 1969. The vascular system in the xis of Dracaena fragrans (Agavaceae). 1, a and development Z primary — Jour. Arnold Arb, 50: 970. The vascular system in eo axis 5 of Dracaena be ans (Agavaceae). 2. Origin and distribution of secondary tissues. Ibid. 51: —491, [J. B. F.] (P. B:T FAIRCHILD TROPICAL GARDEN FAIRCHILD TROPICAL GARDEN 10901 OLp CuTLEeR Roap 10901 OLp CuTLER Roap MIAMI, FLormpa 33156 HARVARD UNIVERSITY Casot FounpATION ATHENS, OHIO 45701 PETERSHAM, MassACHUSETTS 01366 1971] WOOD, SAURURACEAE 479 THE SAURURACEAE IN THE SOUTHEASTERN UNITED STATES ? CarROLL E. Woon, Jr. SAURURACEAE E. Meyer, De Houttuynia atque Saurureis 20. 1827, “Saurureae,” nom. cons. (Lizarp’s-TAIL FAMILy) A small, well-marked family of rhizomatous or stoloniferous herbs of wet habitats characterized by their simple, alternate leaves; ethereal oil cells in stems and leaves; racemose to spicate inflorescences terminal and/or opposite the leaves; small, perfect flowers lacking a perianth; [3 or] 6 hypogynous to epigynous stamens; small monocolpate pollen; Five genera (including seven species) of Asia and North America. The least specialized genus, Saururus, includes one species of eastern North America and one of eastern Asia; Gymnotheca Decaisne comprises two species of southern and western China and Indochina; while Circaeocar- pus C. Y. Wu, known from southeastern Yunnan and northern Vietnam, Houttuynia Thunb., distributed from Japan to southern China and the Himalayas, and the greatly modified Anemopsis Hook. & Arn., of the southwestern United States and Mexico, are monotypic. The genera show many close and reticulate relationships, and the six Species demonstrate well evolutionary changes by reduction, connation, and adnation of floral parts. In both species of Saururus, the four carpels are united only at the base; in S. cernuus L. the six stamens are hypogy- nous, while in S. chinensis (Lour.) Baill. they are adnate to the lower Part of the ovaries. In Circaeocarpus saururoides C. Y. Wu the three or four carpels are united, the ovary is one-locular with a single ovule, of the i Harvard University made possible : ee hae Published, follows the format initiated in the fi Arb. 39: 296-346. 1958). The area dealt wit > Georgia, Florida, Tennessee, Alabama, Mississippi, _Arkansas, og eormyr gene Scriptions are based on the species occurring in this region, sh 4 eauk iad mation from extra-regional taxa in brackets. References followed by an those that I have been unable to check. E : [ am grateful to Dr. Shiu-Ying Hu for her translation from Spat dag acer information pertaining to Circaeocarpus. The illustration ebro fo re “sg Clapman from materials collected by Dr. R. B. Channell a ' e ie ; ad Be Rock and by Dr, K. A. Wilson and me. ms “ge” and “h” in the illustrati after D. S. Johnson in Bull. Torrey Bot. Club 27: pl. 23. 1900. 480 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 and the six stamens are hypogynous. In Gymnotheca, the four carpels are united, the ovary is one-locular with four parietal placentae, and the six stamens are epigynous. In Houttuynia cordata Thunb. and Anemop- sis californica (Nutt.) Hook. & Arn. the carpels are reduced to three and the ovary, three-locular above, one-locular below, has three parietal pla- centae. The three stamens of Houttuwynia are perigynous, the six of Anemopsis epigynous. The bract subtending each flower shows elabora- tion and adnation to various degrees. In Saururus the small bract is ad- nate to the pedicel; in Circaeocarpus, Gymnotheca, and Houttuynia, it is free; and in Anemopsis it is adnate to the ovary and appears to be epigynous, In Saururus, Circaeocarpus, and Gymnotheca chinensis De- caisne the bracts are inconspicuous, but those of the lowermost three or four flowers of the open spike of G. bracteata P’ei are petaloid and white. The bracts of the short spikes of H outtuynia and those of the lowermost six to eight flowers of the condensed spikes of Anemopsis are petaloid, white, and appear to form an involucre, the whole inflorescence in Hout- tuynia and Anemopsis simulating a flower of the type of Anemone vir- ginica L. The adnation of stamens and ovary seen in Saururus chinensis, Houttuynia, and Gymnotheca, is climaxed in Anemopsis by the adnation of both stamens and bract to the ovary and by the lateral developmental be represented by a carpel with its own stamen or stamens. (Cf. Raju.) he accumulating evidence suggests that the Saururaceae and Piperaceae a nat ~— related to each other and that they fall within the group gee a amilies with monocolpate pollen (or phylogenetically de- Vv ypes), ethereal oil cells, and trilacunar or multilacunar nodes, most specialized members. Although the irregularities leading to a high de * Sacs, and parthenogenesis have | pears to be polyploid (2n — 96). 1971] WOOD, SAURURACEAE 481 The development of both male and female gametophytes is quite similar in Saururus, Houttuynia, and Anemopsis, and in all the embryo sac development is of the normal (Polygonum) type. The endosperm is of the cellular type. A conspicuous haustorial cell which persists in the ripe seed is formed by the first division of the endosperm nucleus (see Ficure 1, The group is of little economic importance. Saururus cernuus may be weedy, and its dense growth may favor the breeding of anopheline mos- quitoes; Houttuynia can be an obnoxious weed. The strong-flavored shoots of Houttuynia are eaten as a vegetable in China, and both this and S. chinensis are used in Chinese medicine. REFERENCES: Battton, H. Monographie des Piperacées et des Urticacées. Hist. PI. 3: 465- 537. 1871. [Série des Saurures, 465-4 BENTHAM, G., & J. D. Hooker. Piperaceae. Gen. Pl. 3: 125-133. 1880. [Sau- ruraceae as a tribe of Piperaceae. | DC. Prodr, 16(1): 2356-471. 1869. [Tribe CRETE, P, Sanne reap ary 2 de Yembryon chez |'Houttuynia cor- data Thunb. Rend. Acad. Sci. Paris 245: 1331-1334. 1957; La transformation a ae en graine chez |’Houttuynia cordata Thunb. Ibid. 1652-1655. ca J. Gymnotheca genus nov. e Saururearum familia. Ann. Sci. Nat. I. 3: 100-102. pl. 5. 1845. sates A. Saururaceae. Nat. Pflanzenfam. III. 1: 1-3. 1887. Hitt, T. G. On the seedling-structure of certain ‘les “Ann Bot. 20: 160- 1 Houttuynia, Saururus cernuus, 171, 172.] ab igeoeinsivbaatige 5. H. & A. An aes study. Am. Hoim, T. Anemiopsis ee (Nutt.) Jour. Sci. IV. 19: 76-82. 1905. ; Q Iizumr, S. Studies on the Reis of Houttuynia cordata with special aici to the colony formation. Sci. Rep. Téhoku Univ. Biol. 24: age the ie Jounson, D. S. Seed development in the Piperales and its bearin 3 : gti] tionship of the order. Johns Hopkins igs Circ. tore a ] [ 1905, [Includes Anemopsis, Houttuynia, Saururus cern . Maekawa, F. Inflorescence of alse ba olyp ot a "2a prover In Japanese.) Jour. Jap. i nae S Hs Sate onig i: maa ’s Syllabus der Pflanzenfamilien. ed. _ 2: 147-151. 1964. [Saururaceae, “48; ovules of Saururus cernuus inco tro rectly illustrated as anatr Sd of Houttuynia cordata Thunb. (In MrmarA, T. On the reduction mmary.) Bot. Mag. Tokyo 73: Japanese; English su Nozeran, R. Contribution a l'étude de quelques structures Pike 197k. 1988 morphologie florale comparée. ) ae Sci. Nat. Bot A Houttuynia, 129-1 t. Ko wig fe stoi ogenesis bei Houttuynia cordata, Jap. Jour. Gene : * 6: 14-19. 1930. ansis, Decaisne. Hooker's Tc. Pi. 19: pl. 1873. 1889. ve . rae Matas sete mae = mnotheca of the family, Saururaceae. Contr. "Bigl. Lab Sel. Soc. China Bot. 9: 109-112. 1934. 482 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 QuIBELL, C. H. Floral anatomy and morphology of Anemopsis californica. Bot. Gaz. 102: 749-758. 1941, Raju, M. V. S. Morphology and anatomy of the Saururaceae. I. Floral anat- omy and embryology. Ann. Missouri Bot. Gard. 48: 107-124. 1961. [ Pri- marily Saururus cernuus, Houttuynia, Anemopsis. | SuipaTa, K., . Miyake. Ueber Parthenogenesis bei Houttuynia cordata. Bot. Mag. Tokyo 22: 141-144. pl. 6. 1908. STEENIS, C. G. G. J. vAN. Miscellaneous Botanical Notes II. 19. Gymnotheca Decne, a good genus of the Saururaceae. Blumea 6: 244, 245. 1948. [In- cludes key to genera. ] - Saururaceae. Fl. Males. I. 4(2): 47, 48. 1949. [Houttuynia, Sauru- rus chinensis. | Swamy, B. G. 1. Saururus Linnaeus, Sp. Pl. 1: 341. 1753; Gen. Pl. ed. 5. 159, 1754. Emergent aquatic perennial herbs of shallow, standing or slow-moving waters. Rhizomes subterranean, horizontal, white, branching. Roots ture, one ascending, the other transverse ey below it, only the ascending one developing into a seed. Carpels becomi nly the asc g ng wrinkled and warty in fruit, 1971] WOOD, SAURURACEAE 483 indehiscent, separating at maturity, the styles persistent. Seed solitary, with a pale, filmy outer coat and a hard, brown inner coat; perisperm massive, filling most of the seed; endosperm at the micropylar end of seed surrounding the minute, poorly differentiated embryo. (Spathium Lourei- 10, Mattuschkia Gmelin, Saururopsis Turcz., Neobiondia Pamp.) Type SPECIES: S. cernuus L. (Name from Greek, sauros, lizard, and oura, tail, from the appearance of the inflorescence of S. cernuus.) — Lizarp’s TAIL. Two species, Saururus cernuus, 2n = 22, in eastern North America, and S. chinensis (Lour.) Baill. (S. Loureiri Decaisne, Spathium chinense Lour.), a-h, S. cernuus: a, habit, showing erect savers: § — IGURE I. Saururas. dicel, and rea X 55 ¢ aid sPeonial: Tee, x 1/8; OE vule in carpel at “eft, orizontal ovule in secti a and horiz " ovule in right-hand carpel, cending ovule on removed phe - fruit, X 5; e, mature dry fruit, ae ae face, X 8; soaking, X 12; g, ca él and see in diagra € outer seed coat t Cwhite), hard inner seed coat ney perisp en stipples), endosperm . OP - a) srt mute eno, ee 18; ae erminating seed in diagrammatic ver- tical section, pee | 2n 22, in eastern Asia, ranging from Indochina and southern and cen- tral “China to Taiwan, the oe a westward to Texas, and no i} western Quebec, southern Ontario, centra 484 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 and southeasternmost Kansas and Oklahoma. Although mostly restricted to low altitudes, it apparently is tolerant of a wide range of pH and temperatures. Spreading by rhizomes, Saururus cernuus forms dense colonies that compete successfully with other emergent aquatics. In a study in Tennes- see, Hall estimated that the aérial stems from a square foot of ground in a dense colony bore 13 square feet of leaf surface, while the total length of rhizomes was 13 linear feet. Seed production for the same area was estimated at 20,000, but no seedlings were seen anywhere. The fruits and the fragmented rhizomes presumably are distributed by water (al- though about 95 per cent of the fruits sink within three hours). Johnson found that a dormant period apparently independent of temperature is necessary for seed germination. In germination the endosperm bursts through the seed coats at the micropylar end, projecting as a white papil- la. The embryo, still quite rudimentary, then begins to develop, the hypocotyl first pushing through the tip of the endosperm and eventually only the ends of the cotyledons remaining within the endosperm, which serves as a digestive organ to absorb the food reserves of the perisperm filling most of the seed. Baldwin & Speese suggest that Sawrurus cernuus would be excellent for teaching purposes, since the meiotic divisions of the pollen mother cells are especially clear and all stages are found in the anthers of a single flower. Thirteen chromosome counts (made from very young leaves) rep- resenting 12 widely scattered localities all showed 2n — 22. The four carpels of Saururus cernuus are arranged spirally, and the carpellary vascular bundles originate spirally. The first formed carpels usually have two stamens each, while the last (fourth) does not have one associated with it. The genus is characterized by pedunculate racemose inflorescences without enlarged bracts at the base; flowers t free carpels, each with two and wrinkled, indehiscent one-seeded fruits. The flowers of Saururus white, hypogynous staminal filaments much ex- Ansects. ts of both species are fragrant. Baillon placed S. chinensis in a separate section, SparHIuM feu} Baill : : REFERENCES: Under family ref. ‘oad Vice can erences see BAILLon, Hitt, Jounson, Metcuior, RAJU, BALbwin, J. T. ptt Bot foes ys SPEESE, Cytogeography of Saururus cernuus. Bull. . > 213-216, 1949, [Includes distribution map. ] 1971] WOOD, SAURURACEAE 485 Fujita, T. Ueber die Anatomie und das Ergriinen der Panaschierung bei Sau- rurus Loureirii Decne und Actinidia polygama Planch. (In Japanese; German summary.) Bot. Zool, Tokyo 3: 737-748. 1935.* [S. chinensis.] Hatt, T. F. The biology of Saururus cernuus L. Am. Midl, Nat. 24: 253-260. 1940. Horm, T. Saururus cernuus L., a morphological study. Am. Jour. Sci. V. 12: 162-168. 1926. [Mostly anatomy of vegetative parts and comparison with Piperaceae. | Jounson, D. S. On the development of Saururus cernuus L. Bull. Torrey Bot. Club 27: 365-372. 1900. Murty, Y. S. Studies in the order Piperales. VII. A contribution to the study of morphology of Saururus cernuus L. Jour. Indian Bot. Soc, 38: 195-205. 1 Rousseau, D. Contribution 4 l’anatomie comparée des Piperacees. Mém. Acad. Belg. Cl. Lett. Sci. Mor. Polit. Arts 9: 3-45. 1927.* [Reissued as reprint without change in pagination in Arch, Inst. Bot. Liége 7: 45 pp. 1928; Saururus cernuus, 35, 36, pls. 10, 11.] ARNOLD ARBORETUM HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 486 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 A REVISION OF THE BORAGINACEAE OF WEST PAKISTAN AND KASHMIR * S. M. A. Kazmir 3. A. euchroma (Royle) I. M. Johnston, Contr. Gray Herb. 73: 49. 1924; Riedl in Rechinger, FI. Iranica 48: 164. 1967. Lithospermum euchromon Royle, Illustr. Bot. Himal. Mount. 1: 305. 1839. Stenoselenium perenne Schrenk ex Fisch. & C. A. Mey. Enum, Pl. Schrenk 1: 34. 1841, Arnebia perennis (Schrenk) A. DC. in DC. Prodr. 10: 95. 1846; Ledeb. Fl. Rossica 3: 139. 1847. Munbya perennis (Schrenk) Boiss. Diagn. Pl. Orient. 11: 115. 1849. Macrotomia endochroma Hook. f. & Thoms. ex Henderson & Hume in Lahore 111: 152. 1962; Riedl in Rechinger, Fl. Iranica 48: 166. 1967, syn. nov. Type: India: Naho in Kunawur, Royle s.n. (x). Icon.: Paulsen, 1. c. fig. 20. 1928, under Macrotomia euchroma. * Continued from volume 52, p. 363. 1971] KAZMI, BORAGINACEAE 487 eae subcordate at the base. Inflorescence terminal, subglobular, ending to become broader than long, not elongated in fruit, 5-6 cm. in diameter, bracteate, bracts short, leaflike, not exceeding the calyces, Pedicels erect, densely hairy, very short in flower, to 5 mm. long in fruit. Calyx divided to the base, lobes hairy, linear-lanceolate, acute, to 10 mm. long, 1.5 mm. broad in flower, enlarged to 20 mm. long, 3 mm. broad in fruit. Corolla purplish-white, purple or brownish-purple, funnelform, 16- 22 mm. long, tube 10-15 mm. long, limb 5—7 mm. long and 8-16 mm. in diameter, divided nearly one-half its length, lobes ascending, suborbicular to ovate, throat usually glandless, appendages lacking, but in long-styled flowers occasionally developing 5 small, weakly invaginated swellings, nectary not developed. Anthers borne either at the middle or at the sum- mit of the corolla tube. Style reaching the middle of the tube or slightly exserted, usually forked at the apex; stigmas 2, compressed, rounded, broader than long. Nutlets gray, dusky, irregularly and coarsely tuber- culate and more or less rugose, surface dull, covered with microscopic papillae, broad near the middle, 3-4 mm. long, with a well developed keel on the venter, which extends to the beaked apex and continues down the back side, attachment scar flat and frequently bearing 2 ventral traces. Distrisution: Iran, Afghanistan, West Pakistan, Kashmir, northwest India to Nepal, Turkestan, Pamir Alaj. 3a. Var. euchroma. Plants less pungently hirsute, having narrower cauli or roundish at base and less acute at apex in compar! Arnebia euchroma var. grandis. ne leaves, truncate son with those of West Pakistan: CuHrrraL STATE: Haute vallée de Yarkhun, 4000 m., Schmid 2285 (sm, G), upper Yarkhun Valley, Schmid s.n. (MICH); Chitral Gol, left of Chitral, 3300 m., Stainton 2696 (pM); Turikho River, Istar, 3000 m., Stainton 2487 (gm); Birmogh Lasht, 3300 m., Bowes Lyon 735 (BM). Gitcit AGENCY: Baltistan, upper end of Hushe Valley, Chogolisa Glacier, 3500 m., Webster & Nasir 6001 (cu); Chalpani Nala, Duthie 13855 (BM, ©); Turmih Nullah, _ m., Ludlow 299 (a, pM); Naltar Lakes, 3500 m., R. R. Stewart 26328 (sx): Deosai Pass from Minimarg, 3300 m., R. R. Stewart 22171 (cH); Satpura Nul- wart 20289 (cH); Karakorum Expedition isi Glacier, 4000-4500 m., Russell 1494 s of Buri Harar, 4500 m., Russell 1138 (gm); Baifa Glacier, (Bm); north of Hindukush, Giles s.n. (E). Haz beyond Burawai, Burtt B932 (E). KURRAM jean 498, 824 (c). Querra Dist.: Zarghum, Lace 3760 (®). Kashmir: Dras Valley, 3000-3300 m., Duthie 11695 (E); DetwG™. Pindras, 3000 m., Dickason 84 (MICH); southwest of Dras, Mitsahoi-Matayan, 3000 m., 88 road-miles east-northeast of PAE Mitlohed tadak Rosa, #100 meme fe see 1008 ee) oe ee Road, 300 m., R. R. Stewart 100535 (cu); Zanskar, Chumik Marbo, , Koelz 5381 (Gu, MICH, US); Tsanskar, Pader on the north 488 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 kula Pass to Sulle, Schlagintweit 6258 (GH); Kargia, 3300 m., Koelz 5460 (cu, MICH, US), 5553 (MicH, us); Bok, Pasni La, 5000 m., Koelz 5834 (GH, MICH, us); Zoji La, 3000 m., Ludlow & Sherriff 8311 (GH). Miscellaneous: Tibet occ., 14,000-16,000 ped., Herb. Ind. Orient. Hook. f. & Thoms., Thomson s.n. (GH-iso- type of Macrotomia endochroma). 3b. Var. grandis (Bornm.) Kazmi, comb. nov. Macrotomia grandis Bornmiiller, Oesterr. Bot. Zeitschr. 47: 289. 1897. Arnebia euchroma (Royle) I. M. Johnston subsp. grandis Riedl, Oesterr. Bot. Zeitschr, 111: 152. 1964; in Rechinger, Fl. Iranica 48: 166. 1967. Type: Persia: Kerman, Kuh-e-Laheh Zar, 3600 m., Bornmiiller 3948 (w-holotype, GH-isotype). Plants more pungently hirsute, having broader cauline leaves, obviously subcordate at base, abruptly narrowed in the upper half into a long acu- minate tip. DISTRIBUTION OF VARIETY: Iran. Arnebia euchroma is very variable in the size and indumentum of the leaves, in the density and length of trichomes and the size of their basal tubercles, as well as the length of corolla and calyx. The primary basal leaves are small and usually imperfectly developed whereas the secondary basal leaves are well developed and reach 11 cm. in length and to 1.5 cm. in width. Usually the primary leaves disappear on the emergence of the secondary leaves. The cauline leaves are more stable in their shape compared to the basal ones. The latter have broad sessile bases, broadest at the middle of the lower half, then becoming gradually narrowed to acute apices. Only in few specimens from Tibet and Lahul (in northwestern India) were the bases of the cauline leaves found to be slightly subcordate. The trichomes on both surfaces of the leaves vary greatly in their density, the length and the size of their basal tubercles, and whether they are erect or subappressed. I could not estab- lish any correlation between the ratio of corolla and calyx length and the variations in the leaf-size and the type of trichomes borne. Only in a few specimens from Chitral are the cauline leaves narrower compared to those of the specimens from other parts of West Pakistan, Afghanistan and Kashmir, But such plants are also not rare in the collections from Nepal, the easternmost limit of distribution of 4. euchroma. —_ only variation worth recognition is var. grandis, based on Born- miiller’s collection from Iran, which can be distinguished from the typical variety by being pungently hirsute and having cauline leaves much broader at base, evidently subcordate and abruptly narrowed into a long- acuminate apex. 4. A. inconspicua Hemsl. & Lace, Jour. Linn. Soc. Bot. 28: 326. 1891; Riedl In Kgie & Rechinger, Biol. Skr. 13(4): 240. 1963, in Rech- inger, Fl. Iranica 48: 166. 1967 1971] KAZMI, BORAGINACEAE 489 Arnebia argyrea Rech. f. Ann. Naturh. Mus. Wien 58: 59, 1951. Lithospermum inconspicuum (Hemsl. & Lace) I. M. Johnston, Jour. Arnold Arb. 34: 12. 1953. Macrotomia inconspicua (Hemsl. & Lace) Kitamura, Fl. Afghanistan 321. 1960. Type: BALucHISTAN: Zarghum, at 9500 ft., 1. 5. 1888, Lace 3759 (E-holotype). Icon.: Riedl, 1. c. fig. 187-189. 1963. Erect perennial herb with a woody rootstock formed of crowded branches, ensheathed by old leaf-bases, from a deep dye-stained root with easily peelable bark. Stem usually solitary, arising from the center of the leaf- rosette, unbranched, 5—15(—25) cm. tall, covered densely with appressed, silvery trichomes, not tuberculate at bases, to 0.5 mm. long, intermixed profusely with spreading, thick, stiff, pungent, white trichomes, to 2 mm. long. Basal leaves linear to linear-lanceolate, entire, sometimes slightly revolute at margins, acute with a prominent midrib below, covered with trichomes like those of the stem, 3~7 cm. long, 3-5 mm. broad; cauline leaves few, sessile, shorter and narrower. Inflorescence capitate, globular, 3-5 cm. in diameter, slightly elongated in fruit; bracteate, bracts linear; rarely a small cluster of cymes is developed in leaf-axils below the major cluster. Calyx divided nearly to the base, lobes subulate-linear, more or less equalling the corolla tube, 10-16 mm. long, + 1 mm. broad at base, not elongated in fruit. Flowers heterostyled. Corolla whitish to yellow, sometimes purplish-yellow in dried specimens, narrowly tubular, externally hairy, trichomes appressed, glabrous within with few scattered trichomes on the lobes, 12-19 mm. long; tube 9-12 mm. long, throat funnelform, some- what glanduliferous, short, 2 mm. long in short-styled flowers; tube 12- 14 mm. long and 2 mm. thick at base, 1-1.5 mm. above the base, and above very gradually broadening into the throat in long-styled flowers; limb ascending, 2.5—4.5 mm. in diameter, lobes triangular-ovate, acute, 1-2 mm. long. Anthers 2 mm. long; filaments very short, inserted either above the middle of the corolla tube or on the throat. Style forked, branches slightly unequal, reaching the middle or up to the throat of the corolla pie stigmas 2, distinct, compressed, nearly as long as broad. anes - re - let developed, ovate, indistinctly beaked, dorsally and ventrally keeled, smooth, shining, + 5 mm. long. Distripution: Afghanistan, West Pakistan. West Pakistan: Quetta Dist.: Zarghun, 2800 m., fo _ wil per Stocks 866 (x-in part); Quetta to Ziarat, at the top of 7200-2500 m., Rech- 1176 (e); Kazmi 1663b (pes); in montibus S.W. Ziarat, 1304, Elliot s.n. (x). inger 29346 (c, w). Miscellaneous: Without locality, 1891- , site DC. in DC. Prodr. 10: 94. 1846; Jaub. 4: 100. 1852; Boiss. FI. Orient. 4: 213. 5. A. hispidissima (Lehm.) A. Soc. Bot. 18: 524. 1881; in Hook. f. & Spach, Ilustr. Fl. Orient. 1875; C. B. Clarke, Jour. Linn. 490 JOURNAL OF THE ARNOLD ARBORETUM [vox. 52 Fl. Brit. India 4: 176, 1883; Riedl in Rechinger, Fl. Iranica 48: 155. 1967. Anchusa hispidissima Sieber ex Lehm. Icon. 1: 23. 1821, im synon. Anchusa asperrima Delile, Fl. Aegypt. Illustr. 7. 1813, nomen nudum. Anchusa polygama Hamilton ex G. Don, Gen. Syst. 4: 325. 1837, in synon. Echiodes asperrima (Del.) Rothm. Repert. Sp. Nov. 49: 56. 1940. Lithospermum hispidissimum Lehm. Icon. 1: 23, %. 39,2821. Lithospermum vestitum Wall. Cat. No. 941. 1829, nomen nudum; Benth. in Royle, Illustr. Himal. Mount. 305, 1836; G. Don, Gen. Syst. 4: 325. 1837. Strobila hispidissima G. Don, Ibid. 327. 1837. Toxostigma luteum A, Rich, Tent. Fl. Abyss. 2: 86. 1850. [1851 f. Pritzel.] Toxostigma purpurascens A. Rich. Ibid. 87. t. 77, 1850. [1851 f. Pritzel.] Type: Arcyptus: Wadi Gamuhs, Sieber s.n. (BM, GH, W). Icon.: Wight, Icon. Pl. Indiae Orient. 4: t. 1393. 1848; Jaub. & Spach, 1. c. 4: 4. 363. 1852; Hutch. & Dalz. I. c. fig. 246. 1931, under Arnebia asperrima, Erect or decumbent annual herb, with dye-stained tap root. Stems soli- tary to few, compactly branched; the whole plant covered with stout, white, pungent, spreading to rarely subappressed unequal trichomes with large tuberculate bases, the area between the tubercles usually densely villulose externally, villulose in a band near the base within, ca, 1 mm broad, 10-15 mm. long, tube 7-13 mm. long, nectary very narrow, lobu- late, sparingly villulose, throat glabrous, without glands or appendages, limb funnelform, 4-8 mm. in diameter, lobes roundish, spreading to sub- erect, 1-2 mm. long. Flowers heterostyled. Stamens attached at the mid- dle of the upper quarter of the tube or at its summit, anthers 1.2-1.8 mm. long, filaments usually 0,9-1.2 mm. long. Style reaching the middle of the corolla tube or its summit or Sometimes even exserted, forked, more 1971] KAZMI, BORAGINACEAE 491 deeply so on the long styles; stigma reniform or in long-styled flowers flabellate-reniform and almost lobed. Nutlets pyramidal-ovoid, broadest at the base, pink to brown, nearly smooth to tuberculate, 1.5-2 mm. long. DIsTRIBUTION: north and east Africa, Arabia, Iran, West Pakistan and plains of northwest India. West Pakistan: ATTock Dist.: Campbellpore, ca. 400 m., Kazmi 228b (PEs). BANNv Dist.: between Sarai Naurang and Lakki Marwat, 3 miles from Sarai Maurang, Kazmi 119a (pes); Bannu to Miram Shah, near Mir Ali, Kazmi 127a (pes). CHITRAL STATE: Drosh, 1350 m., Stainton 2254 (BM), DERA GAZI KHAN Dist.: 25 miles from Dera Ghazi Khan to Taunsa, ca. 300 m., Kazmi 2566 (PES); 3 miles from Dera Ghazi Khan to Muzzafargarh, ca. 300 m., Kazmi 2551 (PES). Dera Ismart Kuan Dist.: between Moghal Kot and Daraban, Kazmi 1978b (pes); Dera Ismail Khan, along the bank of Indus, Kazmi 2479 (pes); 10 miles to Tank on Dera Ismail Khan-Tank road, Kazmi s.n. (pes); N. Dera Ismail Khan, Rechinger 30024 (c, w). Hazara Dist.: between Haripur and Hasan 2544 (gz), Tasnif 478 (pes); Jemindar ka Landa near Karachi, Stocks s.n. (K); Dabbeji, Koelz 7596 (GH, MICH, US); Stewart 7091 (cx, K); May 1937, H. Choudhri s.n. (cH). Mutan DIst.: - tan, Kazmi s.n. (PES). PESHAWAR DISsT.: near : | ; and Khairpur, pers 486b (PES). RAWALPINDI Dist.: Golra, Kazmi 2522a (PES); Texila, Kazmi 2548 (PES). eis psa 7s : Ind. Or. Hook. f. & Thoms., 1846, Thomson s.n. (GH, 1 Kew Herb. No. 20708 (kK); Scinde, Stocks 417 (x); Herb. Late East India Co., Falconer s.n. (GH). Reported from Baluchistan, without locality, Stocks s.n. in A. griffithii are equal in ors to 9 mm. long an whereas in A. hispidissima these lobes are Us : long, i to 1.5 fie broad. The corollas in both the tain ee oe se but those of A. grifithii have prominent dark purple ‘a re slightly zygo- of the corolla lobes. In A. hispidissima the conlls 2g vn pispidissions has morphic, while in A. grifithii they are anways sie ee ee a band nearly a glabrous corolla throat and the tube is minutely a $8 1 mm. broad within, which is not the case in —e . -. the eastern Arnebia hispidissima has a wide distribution from gyp 492 JOURNAL OF THE ARNOLD ARBORETUM [von. 52 Himalayas [Griffith 5950 (cu) |, and therefore the plants vary greatly in habit, apparently in response to the varying rigor of the habitat. The plants of the coastal and sandy dry areas are usually lower and more compactly branched, whereas those of the inland areas are usually erect with loosely ascending branches. The plants of Egypt, Arabia and those collected from the southern, sandy and dry parts of West Pakistan usually bear both types of trichomes, short ones without tuberculate bases and the very long, stout and pungent type with broad calcareous tuberculate bases. The short trichomes are usually lacking in the plants collected from the northern part of West Pakistan, Punjab, and Utter Pradesh prov- inces of India. In a few cases these plants have broader bracts. 6. A. griffithii Boiss. Diagn. Pl. Orient. 3: 135. 1856; Fl. Orient. 4: 213. 1875; C. B. Clarke in Hook. f. Fl. Brit. India 4: 176. 1883; Riedl in Rechinger, Fl. Iranica 48: 156. 1967. Echioides griffithit (Boiss.) Rothm, Repert. Sp. Nov. 49: 56. 1940. Lithospermum grifithii (Boiss.) I. M. Johnston, Jour. Arnold Arb. 33: 326. 1952, Type: Afghanistan, without locality, Griffith 5953 (x-holotype, GH, w-isotypes). Icon.: Hooker in Curtis’s Bot. Mag. 87: t. 5266. 1861. Annual erect herb with slender roots slightly stained with purple dye. Stems pale brown, single to many, ascendingly branched, to 30 cm. long; usually densely, rarely sparsely covered with stiff, spreading, white tri- chomes, to 1.5 mm. long, usually arising from tuberculate bases. Basal leaves linear to oblong-obovate, entire, roundish, obtuse or rarely subacute at apex, gradually narrowed towards the base, to 7 cm. long and 1 cm. broad, covered on both surfaces and margins with spreading or subap- pressed trichomes like those of the stems, but usually with larger tuber- culate bases; cauline leaves usually shorter and narrower, Jess attenuate towards the base. Inflorescence congested and short in flower, later elon- gated, bearing distant flowers or fruits, to 10 cm. long; bracteate, bracts hairy, linear, usually shorter than the calyx, to 15 mm. long, 1 mm. broad. Flowers sessile, fruits subsessile. Calyx divided to the base, lobes densely hairy, to 9 mm. long, 0.5 mm. broad, erect, enlarged and dilated at base pote a to 15 mm. long, 1 mm. broad. Corolla yellow or orange, 20- 9) mm. long, tube slender, much surpassing the calyx, 15-18 mm. long, limb spreading with five dark purple spots, 10-15(-20) mm. in di- ameter, lobes broadly ovate or roundish, entire at the margins, throat glabrous or scantily glanduliferous, without appendages, nectary a narrow villous collar. Flowers heterostyled. Stamens attached either just above the middle of the corolla tube or at the summit, and almost exserted. Style reaching to above the middle of the tube or to the summit, sometimes oven slightly exserted, bifid; stigma compressed reniform or flabellate, entire or obscurely lobed. Nutlets trigono-pyramidal, brownish, more Of 1971] KAZMI, BORAGINACEAE 493 less lustrous, sparsely and coarsely tuberculate, back somewhat cordate at base, ca. 2 mm. long and broad at base. DistripuTIon: Afghanistan, West Pakistan. West Pakistan: Atrock Dist.: R. R. Stewart 27521 (pM); Campbellpore, R. R. Stewart 15303 (GH), Kazmi 2720 (pes); between Campbellpore and Grand Trunk Road, ca. 350 m., Kazmi 2729 (pes): between Attock bridge and village, 21 (E). KHyBer AGENCY: Jamrud, 400 m., Kazmi 2693 (pes); between Jamrud and Landi Kotal, Kazmi 2708 (pes); near Afghan border, Rodin 5262 (us); Khyber Pass, near Zintara, Burtt & H. H. Naira 580 (e); between Landi Kotal and Torkham, Kazmi 2714 (pes); along road from Jamrud to Warsak, 300 m., Kazmi 2697 (pes); Warsak, ca. 3500 m., Kazmi 2702 (pes). Konat Dist.: be- tween Matanni and Kohat, 300-400 m., Kazmi 2628 (pes); north of Kohat, Kazmi 2636 (pes); between Hangu and Kohat, ca. 300 m., Kazmi 2626 (PEs); between Kohat and Sulaiman Khel, Kazmi s.n. (pes); Khushalgarh, Kazmi & A. J. Ahmad 2645 (pes); between Khushalgarh and Kohat, Kazmi & Baquar Warsi S.m, (PES); near Hangu towards Tal, Kazmi 2641 (pes). Kurram AceEncy: south of Parachinar, Kazmi 2595 (pes). MALAKAND AcENcy: between Dargai and Malakand village, ca. 5 miles from Dargai, ca. 350 m., Kazmi 370a (pes); south of Malakand village, Abid in Kazmi s.n. (PEs); 4 miles before Chakdara from Malakand, Kazmi 2793 (pes). MIANWALI Dist.: Kalabagh, near bridge, Kazmt $n. (PES). PESHAWAR Dist.: 7 miles from Peshawar towards Khyber Pass, near ( Aitchison 1081 (K); between Texila and Rawalpindi, s Islamabad, hills around Government offices, Kazmi 5.n. (PES). WAZIRISTAN AGENCY: south Waziristan, Wana, 23.4.1939, Ludlow 643 (ce). ZHos Dist.: 26 km. from Fort Sandeman on road to Wana, between Zhob and Sambaza, Lamond 1442 (&), Kazmi 1926a (PES). 1840: 7. 1840; Ledeb. Jour. Arnold Arb. 33: 53; Riedl in Rechinger, 7. A. guttata Bunge, Index Sem. Horti Dorpat Fl. Rossica 3: 139. 1847; I. M. Johnston, 331. 1952; M. Pop. Fl. URSS 19: 176. 19 Fl. Iranica 48: 157. 1967. : Arnebia tibetana Kurz, Jour. Asiat. Soc. Bengal 43(2): gs : 2 fe shee e in Hook. Fl. Brit, India 4: 176. 1883; M. Pop. Fl. ae a wae oe ae Lithospermum guttatum (Bunge) I. M. Johnston, Jour. Arn . 1952, excl. A, thomsoni C. B. Clarke (1883), im synon. ba Type: Hab. ad fluv. Ima in Tschiyam influentem [without citation of collectors name] s.n. (LE-holotype, P-isotype). wee nt or procumbent herb, flowering in decumbe Erect, suberect to rarely 1, with dye-stained tap roots. Stems the first year but becoming perennia 494 JOURNAL OF THE ARNOLD ARBORETUM [vou. 52 usually many, simple or bearing short ascending branches, to 25 cm. long; densely to very densely covered with soft or stiff, spreading, white tri- chomes of various lengths, to 3 mm. long, arising from large, prominent, tuberculate bases. Basal leaves well developed, usually disappearing at anthesis, coriaceous, oblanceolate, entire, obtuse to roundish, attenuate towards the base, to 7 cm. long, 10 mm. broad, covered on both surfaces with trichomes, like those of the stem; cauline leaves oblong-ovate, shorter and narrower, usually more or less narrowed towards the base, rarely broad. Inflorescence of terminal cymes, at first glomerate, 1-2 cm. in diameter, elongating in age to 10 cm., bearing unilateral, closely set flowers; bracts usually equalling or slightly exceeding the calyx. Calyx erect, divided to the base, lobes linear, densely covered with trichomes to 1 mm. long, 7 mm. long in flower, elongated to 12 mm. in fruit. Corolla orange or yellow, minutely hairy with spreading or antrorsely subappressed trichomes ex- ternally, 14-19 mm. long, tube 10-15 mm. long with obvious suprabasal nectary of well developed villous collar, limb spreading, sometimes with a dark purple spot at each corolla sinus, 8-12 mm. in diameter. Flowers heterostyled. Stamens attached just above the middle of the corolla tube or near its summit and partially exserted; anthers 1.5—2 mm. long. Style reaching up to the middle of the tube or slightly exserted, usually dis- tinctly forked with the branches once to several times as long as the stigma; stigmas two, reniform, entire or more or less bilobed. Nutlets dull gray or olivaceous, broadest at the base, + tuberculate, attachment sur- face large, rough, usually with a quadrate prolongation upwards and out- wards in the recessive angle beneath the base of the ventral keel of the nutlet, 2-3 mm. long, DISTRIBUTION: Afghanistan, West Pakistan, Kashmir, northwestern India, Altai, Dsungaro-Tarbagatai, Tien Shan, Mongolia. 7a. Var. guttata. Plants usually with many, much branched stems; calyx densely hairy, trichomes thin, soft, spreading; nutlets densely, minutely and usually uni- formly tuberculate, Pakistan: CHITRAL STATE: Rosh Gol, northeast of Tirich Mir, 3000 m., (GH, RAW); Skardu, 7500 m., Siddiqui, Nasir & Zafar 4233 (pm); Skardu, 2400 m., Aug. 4, 1876, C. B. Clarke 30037A (x), 30037 (sm), 2200-2500 m., Duthie 12064 (BM, E), 2500 m., R. R. Stewart 20302 (us); Bar Loongma Nullah, Mount Haramah, 2200 m., Ludlow 261 (sm); Tsumgak, on the northern foot of the Tsorbat La Pass, to Poen, opposite Tsorbat, Schlagintweit 6060 (cu); on the right side of the Shayok, via Kiris to Neru, on right side of Indus, Schlagintweit 5811 (BM, GH); Kris to Parkutta, Indus Valley, 2500 m., R. R. Stewart 20924 (cH); Udmaro, 3000 m., Koelz 9626 (cu); Kardong to Diskit, on the left side of Shayok, Schlagintweit 2273 (Us); Shayok Valley, Doghari to Kuru, 2700 m., os R. Stewart 20853 (cH); Hispar Valley, 2700-3000 m., Russell 1178 (BM), 700 m., Russell 1843 (gm): above Nomal, 1700 m., R. R. Stewart 26325 (BM), 1971] KAZMI, BORAGINACEAE 495 July 17, 1954, R. R. Stewart s.n. (micw); entre Astor et Bunji, 1300-2300 m., Schmid 1809 (em), Schmid s.n, (mic); Karakoram, Sassir Pass, 3500 m., July 1925, Meinertzhagen 53 (BM), s.n. (BM). Miscellaneous: Tibet occ. regio alp., 14000-16000 ped., Herb. Ind. Orient. Hook. f. & Thoms., Thomson, s.n, sub Arnebia No. 7 (GH, BM). Kashmir: Leh, Hemis nullah, 4500 m., Ludlow & Sherriff 8470 (pm); Ranbir- pur, Indus Valley, 3000 m., Ludlow & Sherriff 8482 (Bm, cH); Pitug, Koelz 6317 (GH, MIcH, us); Nima Mud, 3500 m., Koelz 3333a (gz, GH, us); Chakar Talao, 3000 m., Koelz 2459a (cH); Hemis, Koelz 6391 (micH); Kugzil La, Rupshu, 5000 m., Koelz 2325 (GH), 4000 m., Koelz 2323a (cu, us); near Kyber, April 1915, Mrs. Prescott-Decie, s.n. (BM); Indus Valley, near Pituk, 3100 m., Voigt 1060 (us). 7b. Var. thomsonii (C. B. Clarke) Kazmi, comb. nov. Arnebia thomsonii C. B. Clarke in Hook. f. Fl. Brit. India 4; 176. 1883; M. Pop. Fl. URSS 19: 177. 1953. Type: Tibet occ., regio alp., 14,000 ped., Herb. Ind. Or. Hook, {. & T. Thoms., T. Thomson, s.n. as Arnebia No. 6 (K-holotype, GH-isotype). Plants usually with single or poorly branched stems; calyx very densely hairy, trichomes thick, stiff, subappressed; nutlets smooth or sometimes with few scattered tubercles. West Pakistan: Gitcir AGENcy: Baltistan, Piti Valley, Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (GH, K). Kashmir: Reported from Ladak, Leh, Lance s.n. (K?). The widely distributed species, Arnebia guttata, is very variable in the form of its leaves and in the type of trichomes they bear. In agi guttata many spreading or decumbent stems which are usually es subdivaricately branched arise from the root. The cauline stg ats from linear (Stewart 26325) to broadly lanceolate (Mrs. Pr sie in i sn.). The upper part of the stem, inflorescence, and the ca si a densely covered, usually with long, thin spreading arg thom- lets are uniformly tuberculate, at least on their dorsal surface. Var. 1 ] | . ] of var. guttata. The sepals are more or less comp trichomes. The bracts in var. thomsonii are also ¢ e than those of the typical variety. All these characters ae al surface. C. nutlets, smooth or with few scattered tubercles on t F spe a a B. Clarke (J. c.) gave some differences 1n ener Popov (i. c.) dif- of the corolla tube, which I was able to verify. } 4 ge ‘sufficient ferentiated Arnebia tibetica from A. guttata but I “ a difference between the two to justify their separate iden'i'y- 3 4 Misc. 3: 8. A. fimbriopetala Stocks, Hook. Jour. Bot. & pa Fl. 180, 1851; Boiss. Fl. Orient. 4: 205. 1875; Rie ; Tranica 48: 157. 1967. 496 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Arnebia bungei Boiss. Fl. Orient. 4: 215. 1875. Echioides bungei (Boiss.) Rothm. Repert. Sp. Nov. 49: 56. 1940. Lithospermum fimbriopetalum (Stocks) I. M. Johnston, Jour. Arnold Arb. 33: 326. ‘1952. Lithospermum bungei (Boiss.) I. M. Johnston, ibid. Arnebia fimbriopetala var. bungei (Boiss.) Riedl in Rechinger, Fl. Iranica 48: 157. 1967, syn. nov. Type: Upper Baluchistan, Stocks 977 (x-holotype). Icon.: Stocks, 1. c. 3: 180. #. VI. 1851. Erect to suberect annual herb with short dye-stained tap roots. Stems usually solitary or rarely few, slender, sparingly branched near the base, branches spreading, 5-15 cm. long, covered with soft, crisped, white, usu- ally antrorsely appressed to subappressed, trichomes without tuberculate bases. Basal leaves few, oblong-ovate, entire, sometimes irregularly revo- lute, with roundish apex, slightly narrowed towards the base, thick, 20- 25 mm. long, 5 mm. broad; both surfaces covered + uniformly, with very short, very thin, subappressed trichomes to 0.5 mm. long, arising from prominent tuberculate bases, rarely glabrous below; cauline leaves few, sessile, shorter and Narrower, sometimes ciliate on the margins of the bracteate, bracts leaflike, sometimes ciliate at the margins, to 6 mm. long, 1.5 mm. broad at base. Pedicels stout, + 1 mm. long. Calyx divided near- ly to the base, lobes linear, obtuse, bearing minute trichomes on both sur- faces and densely so on margins, 15-17 mm. long, 1.5 mm. broad at base in flower, enlarged at maturity, glabrescent with midrib becoming prom- Inent, erect, to 30 mm. long, with major lobes 3—5 mm. broad near the base, in fruit. Corolla yellow, externally villulose, 25-30 mm. long, tube slender, surpassing the calyx, 20-24 mm. long, bearing a membranous more broad, rounded, with distinctly lobulate or fimbriate margins. Flowers heterostyled, in long-styled flowers the corolla tube cylindrical, abruptly enlarged below the summit, 6-8 mm. in diameter. Anthers 2—2.3 mm. long, attached low in the upper third of the corolla tube; filaments short. Style , lobes 0.5 mm. long; stigmas 2, compressed, as broad as long, obscurely bilobed. In short styled flowers corolla tube gradually ampliate, anthers attached at the summit of the corolla tube with their upper third exserted; style two-thirds as long as the corolla tube. Nutlets olive-green to reddish, beaked above the middle, lower half i of a broad depression, upper half with » venter angulate, attachment surface broadly triangular, basal, gynobase very depressed, 33.5 mm. long, 3 mm. broad at the base. DistTRIBUTION: Iran, Afghanistan, West Pakistan. Pes Pakistan: Makran Disr: Between Pasni and Kappar, road to Gwader, azmi 1011 (PEs); Lamond 456 (E), Rechinger 27871 (c, w). Quetta Dist.: 1971] KAZMI, BORAGINACEAE 497 Doband, Chiltan, Stocks s.n. (kK); Gwal, Munro s.n. (K); Maslakh, R. R. Stew- art 28281 (K, RAW); Chahar Sar, Crookshank 69 (x). Riedl (/. c.) distinguished Arnebia fimbriopetala var. bungei from the typical variety, by attributing to var. bungei the crenate or subentire co- rolla lobes and to var. fimbriopetala deeply fimbriate corolla lobes. Variety bungei is based on Bunge’s collection (Ssertschah, 12 Mart., 1859; c-holo- type, GH-isotype) from Persia. A careful study of this collection and other material from the area showed that the depth of laceration of the corolla lobes varies from plant to plant and this character can not be correlated with any other. Therefore, I came to the conclusion that A. fimbriopetala and A. bungei are conspecific. 9. A. linearifolia A. DC. in DC. Prodr. 10: 95. 1846; Boiss. FI. Orient. 4: 214. 1875; M. Pop. Fl. URSS 19: 182. 1953; Riedl in Rechinger, FI. Iranica 48: 162. 1967. Arnebia flavescens Boiss. Diagn. Pl. Orient. 11; 117. 1849. Echioides linearifolium (A. DC.) Rothm. Repert. Sp. Nov. 49: 56. 1940. Lithospermum aucheri I. M. Johnston, Jour. Arnold Arb. 33: 328. 1952. Type: In deserto Syriaco, Aucher-Eloy 2368 (c-holotype). Erect to suberect annual herb with tap root and lower part of stem dye- Stained. Stems usually many, spreading, sparingly branched, to 15 cm. long; stem and branches rarely totally glabrous, sometimes partly . glabrous, but usually covered sparsely to densely with thin ig an- trorsely subappressed or spreading very short trichomes, to 1 ey ane usually not at all to sometimes with tuberculate bases, intermix ine few scattered, stout spreading trichomes, to 2.5 mm. long, — prominent large tuberculate bases. Basal leaves few, —, ak thesis, oblong, entire, with roundish or subobtuse apex, cpiinage ie the base, sometimes into short petioles, to 5 cm. long, gow ‘ck spe eed per surface densely, lower surface sparsely covered wit rash Pll pressed trichomes, or the lower surface sometimes subglabrous; - sais the lower half of the leaves with long trichomes on = Dee ee tee. leaves few, subsessile, obtuse to subacute with more Jong tric recall culate at the base on both surfaces and margins. te leaflike, pact in flower, later elongated, to 12 cm. long, stent a clos. shorter and narrower. Pedicel very short in sonieghe ; ength, lobes erect, : . divided most 0 : : gated to 3 mm. in fruit. Calyx mee trichomes, intermixed with long 3.5 mm. above the base of the calyx, with very stout and hardened psi an parts, midrib and the margins bearing V —5 mm. broad arising from large mamilliform tubercles, to A cougars Lay a at base. Corolla yellow, subtubular, densely hairy , 498 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 long, tube slender, usually surpassing the calyx by 2-3 mm., nectary a papery lobulate collar, ca. 0.5 mm. high, throat glabrous, without glands or appendages, limb of small ascending lobes, to 4 mm. in diameter. Anthers attached at the summit of the corolla tube, included. Style reaching 1-3 mm. below the anthers, short, forked at apex, bearing 2 compressed, reniform stigmas. Nutlets rubiginous, sparingly punctate and tubercu- late to sometimes rugose, back rounded, 2.5-3 mm. long, ca. 2 mm. broad near the base. DistriBuTION: North Africa, Transcaucasus, Syria, Iraq, Iran, Afghanis- tan, West Pakistan. West Pakistan: KaLat Dist.: Between Kalat and Mastung, 1800 m., Jafri & Akbar 2001 (=). Maxran Dist.: Between Hushab and Panjgur, 20-30 km. from Panjgur, Kazmi 1186b (pes), Lamond 566 (E), Rechinger 28150 (w); about 10 miles from Hushab towards Panjgur, Kazmi 1159b (pes). Quetta Dist.: Pishin, 29.3.1889, Lace, Herb. Watt No. 3566 (E); Quetta below Yaro and Sarnan, 6 miles from Quetta, on Quetta~-Chaman Road, Kazmi 1497 (pes); Quetta to Chaman, Yaro near Bostan, ca. 40 km., from Quetta, 1450 m., Lamond 587 (£), Kazmi 1480b (prs), Rechinger 28927 (c, w). Reported from Qilla Abdullah, Duthie 8681; 45 km. east Mirjaveh, 800 m., Grant 15329; Baluchistan, without locality, Stocks s.n. (?). Riedl (/. c.) differentiated Arnebia linearifolia into two subspecies, viz. subsp. linearifolia and subsp. desertorum Riedl (Oesterr. Bot. Zeitschr. 109: 74. 1962). Subsp. desertorum is distinguished from our subspecies | by having much dilated bases of the calyx lobes at fruiting stage, which are to 10 mm. broad. Subspecies desertorum is confined to Iraq, whereas the typical subspecies is distributed all over the area cited for the species. 10. A. decumbens (Vent.) Coss, & Kral. Bull. Soc. Bot. France 4: 402. 1857; M. Pop. Fl. URSS 19: 180, 1953; Riedl in Rechinger, Fl. Tranica 48: 159, 1967, Lithospermum decumbens Vent. Descr. Pl. Jard. Cels. 37. 1801. Lithospermum micranthum Viv. Fl. Lib. Spec. 10. 1824. ! . 1818. Lithospermum cornutum Ledeb. Fl. Alt. 1: 175. 1829 spree: pees: (Ledeb.) Fisch. & C. A. Mey. Index Sem. Hort. Petrop. Arnebia vivianiit Coss. & Dur. Ann. Sci. Nat. Ser. 4. 1: 240, 1854. Arnebia orientalis Lipsky, Acta Horti Petrop. 26: 513. 1910. Tyre: Inter Baghdad et Hit, Bruguiere & Olivier s.n. (p-holotype). Annual suberect herb with slender, dye-stained tap roots. Stems pale, usually many, branched near the base, to 30 cm. long, covered densely with white, soft, subappressed trichomes, to 0.5 mm. long intermixed with long, stiff, spreading trichomes, to 2-2.5 mm. long, long trichomes with distinctly, and short ones without or with only inconspicuous tuberculate ases. Basal leaves usually disappearing at anthesis, oblong-ovate, entire, 1971] KAZMI, BORAGINACEAE 499 sometimes revolute at the margins, with roundish apex, gradually at- tenuated towards the base, 2—3 cm. long, 2—5 mm. broad, covered on both surfaces with short and long trichomes, to 1 mm. long, usually longer towards the base of the leaves, nerves sunken above, prominent below; cauline leaves sessile, oblong to oblong-lanceolate, usually broadest at the middle of the lower half, entire, slightly revolute in the upper half, obtuse, to 2.5 cm. long, 5 mm. broad, lower surface sometimes subglabrous, but usually bearing shorter trichomes on both surfaces. Inflorescence terminal, simple to geminate, short and congested in flower, later elongated to 9 cm., bearing fruits to 10 mm. distant, bracteate, bracts leaflike, usually bearing some long trichomes on the margins and midrib of their lower half, flowers sessile, fruits subsessile. Calyx divided nearly to the base, lobes linear much shorter than the corolla tube, 4-5 mm. long, ca. 0.5 mm. broad in flower; fruiting calyx 10-15 mm. long, connate 2-3 mm. from the base, lobes linear in the upper half, much dilated below, midrib stout, hardened, subconnate to connate, lateral veins reticulate and very prominent in the connate part, bearing few to many stiff, white trichomes usually with tuber- culate bases, lobes 7-12 mm. long, 1.5 mm. broad at base. Corolla yel- low, tubular-infundibuliform, densely pubescent externally, especially on the lobes, to 12 mm. long, tube slender, widened gradually from the mid- dle upwards, throat ca. 1-5 mm. wide, limb ascending, lobes roundish, ca. 1 mm. long and broad. Anthers attached above the middle or on the throat of the corolla tube, ca. 0.8 mm. long, filaments very short. Style ome the middle of the upper half of the corolla tube, bifid to once again bi . , stigmas 4, capitate. Nutlets brown, beaked, ventrally and chr mr “ keeled, dorsally densely tuberculate, 2-2.5 mm. long, 1-1.5 mm. broad a the base. Distrmution: North Africa, Syria, Caucasus, Iraq, Iran, Afghanistan, West Pakistan, Turkestan, Pamir Alaj, Tien Shan, China. West Pakistan: Katat Dist.: Punjgur to ae ar nd tale 7 Lamond 662 (£), Kazmi 1263 (PES); between Kolp hetween Qilla Abdullah eal R. R. Stewart 29292 (MICH); ou sery, Kazmi 1421b (pes). Reported trom *'™ 20-30 km. S. Punjgur, Rechinger 28131 (w); a sn.; Kharan, Hughes-Buller 23184, 23148; Stocks $.n. : mbens var. macr Arnebia decu guished from the 500 JOURNAL OF THE ARNOLD ARBORETUM [vot. 52 typical subspecies by having fruiting calyx lobes dilated to 1.5 mm., and the lower connate part covered with white tubercles; it is confined to North Africa and southern Iran, whereas the typical subspecies is dis- tributed in our area and is spread over the area of distribution of the whole species. 11. A. grandiflora (Trautv.) M. Pop. in Lapin. Opred, rast. Tasch- kentsk. Oasisa 342, No. 319. 1938, Fl. URSS. 19: 705. 1953; Riedl in Rechinger, Fl. Iranica 48: 161. 1967. Arnebia cornuta (Ledeb.) Fisch. & C. A. Mey., var. grandiflora Trautv. Bull. Soc. Nat. Mosc. 4: 422. 1866, Arnebia orientalis Lipsky, Acta Horti Petrop. 26: 522. 1910. Type: Ad urbem Tashkent pagum Toj-Tjube, inter Tashkent et Angren, Regel s.n. (tE-holotype). Icon.: Ried], 1. c. t. 47. fig. 1967. _ Annual erect or suberect herb with slender, dye-stained tap roots. Stems simple or branched, branches divaricate (at least in specimens from West = its throat, sometimes exserted, slender at apex, bifurcate, branches s stigmas 4, + clavate. Nutlets dark greenish-brown, pyramidal, subcordate at base, abruptly narrowed at the middle into a triangular 1971] KAZMI, BORAGINACEAE 501 upper part, to 1 mm. long, ventrally and supradorsally slightly keeled, unequally tuberculate all over the surface, ca. 2 mm. long and broad at base. DistRipuTION: Iran, Afghanistan, Turkestan, West Pakistan. West Pakistan: CHITRAL STATE: Mastuj, 1950 m., Stainton 2463 (pm). ZHoB Dist.: between Zhob and Sambaza, 15-16 miles on Fort Sandeman-Wana Road, Kazmi 1908 (Gu, PES), Rechinger 29837 (G, w). The material collected from Chitral agrees in all details with the type and other material collected from Turkestan and Afghanistan. The speci- mens (Kazmi 1908, Rechinger 29837) collected from north Baluchistan differ from the type material in being more profusely branched with the inflorescence composed of many closely set flowers or fruiting calyces. The calyx lobes are not at all connate at base in fruit; are more dense- ly pubescent with trichomes slender and long. The flowers are usually longer than in the plants from Turkestan. I could not examine any ma- terial from Iran and, of course, not much even from West Pakistan to establish this variation fully. On further study, with more material in hand, the plants from Baluchistan may prove to be a subspecies or even a distinct species worthy of recognition. 12. A. waziristanica H. Riedl, Oesterr. Bot. Zeitschr. 109: 58. 1962; in Kgie & Rechinger, Biol. Skr. 13(4): 245, 1963, in Rechinger, Fl. Tranica 48: 156. 1967. Type: Pakistan: Waziristan, 21 May 1895, Harsukh 15803 (pp-holo- type). Icon.: Riedl, 1. c. fig. 193. 1963. divided at base, 3-20 cm. long; narrow-linear, similar to the calyx lobes. nearly to the base, densely covered in the spreading trichomes, lobes meen flower, to 15 mm, long, 1 mm. broad in IT ely | the calyx, in dried scecsinil brown to dark purple, | eee pal ex ternally pilose, tube slender, hardly 1 mm. In diameter, a neavted. to 2 mm., annulate within at the base. Flowers distinc a aerige Anthers sessile attached at the base of the limb or . ee corolla tube at the same level. Style in long-styled ey - aa vai stigmas 2, minute, capitate; not seen in short-styled ao eae trally distinctly keeled, densely minutely verruculose an gulose, 2-3 mm. long. 502 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 DISTRIBUTION: West Pakistan. West Pakistan: Waziristan, Harsukh 15803 (pp). I could not examine any plant of this taxon. The description given is according to Riedl (/. c.). From the description Arnebia waziristanica ap- pears to be closely related to A. griffithii and A. hispidissima, from which it is distinguished by its brown to dark purple flowers. 13. A. simulatrix H. Riedl, Oesterr. Bot. Zeitschr. 109: 70. 1962, in Kgie & Rechinger, Biol. Skr. 13(4): 243. 1963. Type: Afghanistan, Neubauer 205 (w-holotype). Icon.: Ried], 1. c. fig. 3. 1962; 1. c. fig. 192. 1963. Annual herb with an elongate, fusiform dye-stained root. Stems simple to branched, to 10 cm. long, covered with stiff trichomes. Basal leaves dis- appearing at anthesis; cauline leaves few, lanceolate, to 6 cm. long and 6 mm. broad, densely covered on upper surface, margins, and midrib, loosely so on the lower surface, with stiff, spreading to subappressed trichomes. Inflorescence short, congested in flower, 1-2 cm. long, later elongated to 4-5 cm.; bracts lanceolate, shorter than the calyx, dilated at base. Pedi- cels hardly developed in flower, to 2 mm. long in fruit. Calyx 8-12 mm. long in flower, 18-20 mm. long in fruit, connate from the base up for 3 mm., scarcely angulate, medial nerve less prominent compared to that in Arne- bia decumbens (Vent.) Coss. & Kral., bearing some long tricomes, aris- ing from minute tuberculate bases, which are gradually reduced in size upward; lobes pilose externally and at the midrib, 1 mm. broad in flower, 2-3 mm. broad in fruit. Corolla yellow, densely pilose externally, infundi- buliform, 15-18 mm. long, tube thick, dilated at the throat, 2—2.5 mm. in diameter, glabrous internally, lobes erect to slightly recurved, margins ciliate, to 1 mm. long. Anthers 5, three attached 3 mm. below, and the other two ca. 5 mm. below the margins of the corolla lobes. Style 10-12 mm. long, bifid to once branched; stigmas 4. Nutlets 4, sessile, horizontal, acuminately beaked, dorsally convex, slightly keeled in the upper part, verruculose, ventrally angulate, obtusely keeled, the two faces minutely verruculose, 3-3.5 mm. long. DISTRIBUTION: Afghanistan, West Pakistan. West Pakistan: Quetta Dist.: 8 km. northwest of Harnai, 1000 m., Rechinger 29551 (w). _ In habit Arnebia simulatrix is very similar to A. lineari folia, from which it is distinguished by having 4 distinct stigmas. It is also closely related to A, decumbens but differs from that species by its less dilated calyx lobes and less prominent midrib on the connate part of the calyx in fruit. 29. Onosma L. Sp. Pl. ed. 2.1: 196. 1762; Gen. Pl. ed. 6. 76, 1764. TYPE SPEctEs: OQ. echioides L. 1971] KAZMI, BORAGINACEAE 503 Perennial or biennial, coarsely hispid, herbs or sometimes undershrubs. Leaves various, usually numerous. Inflorescence of single or of paired terminal cymes, usually not much elongated in fruit, bracteate. Flowers regular and radially symmetrical. Calyx usually accrescent and divided to the base or sometimes with a short tube, lobes 5, equal, linear lanceolate, oblanceolate, or ligulate. Corolla yellow, blue or rarely white or red, tu- bular, obconic-tubular or rarely ellipsoid, usually gradually broadening up- wards from the base; inside the base of the tube a well developed nectarif- erous ring or collar, the throat without faucal appendages, lobes small erect or with tips more or less recurved. Anthers elongate, terminated by well developed appendages, usually coherent at the base or frequently joined laterally by the interlocking of minute marginal trichomes to form a tube, terminal appendages elongate, usually hyaline and chartaceous, tips emarginate, basal appendages usually short and thick, filaments inserted below, near, or above the middle of the corolla tube, short to very long. Stigmas 2. Nutlets 4, erect, venter usually angled, dorsum convex or sometimes flattened below the middle, attachment scar basal, usually triangular, as broad as long, horizontal or somewhat oblique. Gynobase pyramidal or flat or even depressed at the center, the attachment faces usually distinct, frequently concave. Species about 85, with greatest diversity and maximum concentration in Iran, distributed westward to Syria and Turkey, eastward and northeast: ward to Turkestan, Altai, China, Afghanistan, West Pakistan, Kashmir, Tibet, and India to Burma. KEY TO THE SPECIES a. Anthers united only at the base, laterally free. ) b. Corolla ae or at most equalling the calyx; bracts lancoatete aor excessively prolonged tips; leaves densely clothed WO ah ee trichomes on the lower surface. ....--------- ee webs ptt b. Corolla longer than the calyx, obviously surpassing the ca: ae eet lanceolate or subulate with acute tips; lower surface various é va nr c. Corolla 25-40 mm. long; anthers 9-12 mm. long ‘aaa long; nectary glabrous; nutlets 4-5 mm. long. +i slang at ale c. Corolla 8-15 mm. long; anthers 4-6.5 mm. long; es cesta: ng; nectary pubescent; nutlets 3-4 mm. sg ni .U. a. Anthers united laterally as well as basally, ction gem se ceatile orolla tube villous; corolla externally e. Base of filaments and adjacent coroli ee e lower covered with antrorsely appressed trichomes; peer | b tae, surface of the leaves appressed. .--------- °° oe lly cov- e. Base of filaments and adjacent area glabrous, cS face of the leaves spreading. -.---------";7 32 rt d. Anthers included in the corolla or with .. Filaments 2-2.5 mm. long, Oe of filaments. (2.6065 ee 504 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 f, Filaments 5-10 mm. long, base of the anthers 2-6 mm. above the base of the filaments. g. Plants with strong perennial root; stems 10-75 cm. tall; basal ad hispid aa a pet) =] + nm z ~ | 1 ae Oo i=) e. 9 pany ° nd wm j=) ion] o a bare < Oo Qa -t ° ° . co wn wm ct oO =] wn _ .. Cy) « Q = co bss] a and acuminate tips; corolla 30-32 mm. long; filaments 9-10 MPR erect eek, SPS Be 3. O. khyberianum. florescence not slenderly attenuate in the upper part, tips acute; corolla 22-30 mm. long; filaments 6-9 mm. long. ee eres icca i, 6c pcp. J; 4. O. griffithii. = w bah) a: vBe wn ° aa) ee * =] 1. O. chitralicum I. M. Johnston, Jour. Arnold Arb, 32: 209. 1951; Riedl in Rechinger, Fl. Iranica 48: 179. 1967. Type: West Pakistan: Chitral, S. M. Toppin 48 (x-holotype). Perennial herb with a stout woody root. Stems many, erect or ascend- ing, branched above, to 30 cm. long, 7 mm. thick at the woody base; densely covered with minute, thin, spreading trichomes, to 0.2 mm. long, intermixed sparsely to profusely with long, pungent trichomes, to 4 mm. long, arising from prominent, white, tuberculate bases, + retrorsely sub- appressed on the lower parts of the stem, spreading on the upper parts. revolute at the margins, roundish to subobtuse, elongate and gradually gated to 8 cm. in fruit, Calyx hairy, divided to the base, lobes narrowly 1-1.5 mm. broad in flower, enlarged to 10-12 mm. long and 2 mm. broad in fruit. Corolla white, brownish in dried state, Can » externally subglabrous to minutely puberulous, glabrous within, 12-14 mm. long, tiie 2 mm. thick us aie gradually widened upwards, lobes short, at the base with a glabrous nectary, 0.2-0.3 cos high. Anthers united laterally as well as basally, forming a tube, : plas exserted, 6-7 mm. long; filaments linear, inserted below the mid- € of the corolla tube, 2-2.5 mm. long. Style ca. 15 mm. long, slightly exserted from the corolla, stigma not conspicuous. Nutlets 3-4 mm. long, smooth, shining. 1971] KAZMI, BORAGINACEAE 505 DISTRIBUTION: West Pakistan. West Pakistan: CH1TrAL State: Chitral, Toppin 481 (K-type); Chitral village, 10 miles south, 1350 m., Stainton 2216 (BM). Onosma chitralicum is based on the collection of Toppin from Chitral. In the type specimen the basal parts of the plant are missing. I. M. John- ston, in his description, assumed this plant to be an annual. On close examination of another specimen collected from the same locality by Stainton, which is very complete in all respects, the species proves to be a perennial. It has a strong, thick woody root and persistent basal leaves. The description of the basal parts of the plant, therefore, has been com- pleted above. Riedl (1. c. 48: 185. 1967) cited the collection (Stainton 2216) under Onosma limitaneum subsp. limitaneum, and probably his measurements of the leaves of subsp. limitaneum are also based on this plant. Onosma chitralicum is distinguished from all related species in our area by having laterally as well as basally united, very short anthers, filaments 2-2.5 mm. long, with the bases opposite the bases of anthers, and white corollas. 2. Onosma hispidum Wall. ex G. Don, Gen. Syst. 4: 317. 1838; Riedl in Rechinger, Fl. Iranica 48: 177. 1967. Onosma hispida Wall. Cat. No. 938. 1829, nomen nudum. Onosma kashmiricum 1. M. Johnston, Jour. Arnold Arb. 21: 50. 1940. Onosma hispidum var. kashmiricum (1. M. Johnston) I. M. Johnston, Jour. Arnold Arb, 32: 210. 1951. Type: Inpta: Kumaon, Wallich 938 (kK). Perennial herb with a strong, thick, tap root. Stems fistulous, reo many, simple or rarely branched, 10-70 cm. long, 3-6 mm. a . = covered densely with bristly, spreading white trichomes, to AN mage arising from large tuberculate bases, the area between the bases, ae a quite glabrous, densely puberulous, or bearing very short, gate ia chomes. Basal leaves persisting at anthesis, narrowly oblanceolate, , roundish, obtuse or at times subacute, usually midrib inconspicuous above, raised and prominent below, to that on the stems, trichomes on the a omits? estes aii spreading below; middle cauline leaves sessile, linear - — ea. ahiorter, obtusish, 4-10 cm. long, 4-10 mm. broad; upper cauline lea eal narrower to sometimes comparatively broader. ae cea sida geminate cymes, dense at flowering, 3-4 cm. in at ie <9 ; gating to 15 cm. long in fruit; bracts lanceolate, broa long in flower, narrowed towards the acute apex. Pedicels hairy, 1-3 mm. 5-10 mm, in fruit. Calyx divided to the base, pies oe covered with -_ subappressed trichomes like those 0 , mm. long in flower, 15-25 mm. long 1n frui labrous pale yellow, externally sparsely hispidulous, puberulous on _—* 506 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 inside, 18-30 mm. long, 3—3.5 mm. thick at base, expanding gradually up- wards, 8-11 mm. wide just below the sinuses, lobes 1.5 mm. high, re- curved outside. Anthers united basally and laterally to form a tube, 9-11 mm. long, tips 2 mm. long; filaments ligulate 5-9 mm. long, decurrent bases to 5 mm. long, inserted 8-13 mm. above the corolla base. Nectary glabrous, collar + 0.3 mm. high, Style glabrous, 13-23 mm. long. Nut- lets + 6 mm. long, smooth, obscurely roughened, + shining. DISTRIBUTION: Afghanistan, West Pakistan, Kashmir, northwestern dia. West Pakistan: CHrTraL STATE: Kafiristan, Rumbur-Gangalwat Gol, 2850 m., Bowes Lyon 669 (pm); Basin Gol, 3450 m., Bowes Lyon 952 (sm); Chitral vil- lage, 1500 m., Stainton 2182 (pm). Gitcrr AcENcy: Naltar, 3500 MT,» My te Stewart 26341 (Bm); Astore, Gudhai Valley, 3500-3900 m., July 21, 1892, Duthie s.n. (Bm); Kamri Kotal, 3600 m., Giles 140 (kK); expedition to Karako- ram Glacier, 1892, Conway 335 (k). Hazara Dist: Kaghan Valley, Naran, An- war Ali 117 (micH); Naran to Saiful Maluk Sar, 2900 m., Kazmi 2273a (PES); to Garhi Habibullah, Kazmi s.n. (pgs); Abbottabad, Sherwan hills, Kazmi 101a (PES); Batal, Kazmi sn, (pes). Kuyper AGENCY: In jugo Landi Kotal, 700- 1000 m., Rechinger 3034 (w); Kazmi 2705 (pes). Kurram AGENCY: Parachinar, m., Rechinger 28413 (w); inter Bostan et Saran Tangai, 1700-1800 m., Rech- inger 29179 (w); inter Qilla Abdullah et Sheila Bagh, 1600-1900 m., Rechinger 28978 (w); in jugo Khojak, 2300 m., Rechinger 29055 (w). Reported from: between Quetta and Ziarat, Jafri & Akbar 2035 (?). > Below Tragbal, 2100-2400 m., R. R. Stewart 19428 (cH); Tashgam, Treaty Road, Ladak, 3000 m., Ludlow & Sherriff 8340 (Gu); Matayan to Mechu- lor, Sept. 1880, Young s.n. (pM); Gangangir, Sind Valley, Ludlow & Sherriff 7959 (BM), 1960 m., Ludlow & Sherriff 7625 (BM); Bringhi Valley, 1850 m., Lidder Valley, 27 road-miles north of Islamabad, 2700 m., Dickason 74 (MICH); 000 ped., Herb. Ind. Or. Hook. f. & Thoms., April 22, 1848, Timmins 1 (sm); Himal. bor. ( OcCc., regio t é : Ind. Or. Hook. f. & Thoms., Thomson Sn (oni): “aio cians eae deeanaior 1971] KAZMI, BORAGINACEAE 507 Onosma echioides L. (in Hook. FI. Brit. India 4: 178. 1883), the name given to the plants belonging to Onosma hispidum, by C. B. Clarke is a misidentification. 3. O. khyberianum I. M. Johnston, Jour. Arnold Arb. 32: 211, 1951; Riedl in Rechinger, Fl. Iranica 48: 178. 1967. Type: West Pakistan: Khyber Pass, 3700 ft., H. H. Johnston 22 (r- holotype). Perennial herb with stout, somewhat knotted roots. Stems solitary to many, usually simple, sometimes branched above, covered densely with unequal, spreading, white trichomes to 6 mm. long, arising from large tuberculate bases, the area between the bases glabrous to puberulous or covered densely with erect, minute trichomes. Basal leaves persistent at anthesis, oblong to oblong-lanceolate, entire, obtuse to subacute, gradually narrowed towards the base, 4-9 cm. long, 5-9 mm. broad, covered on both surfaces with trichomes like those on the stem, midrib inconspicuous above, raised and prominent below; cauline leaves sometimes slightly revolute at the margins, acute, to 4 cm. long, 5—6 mm. broad. Inflorescence of ter- minal cymes, simple or branched, short and dense in flower, to 5 cm. in diameter, elongated to 18 cm. in fruit, bearing fruiting calyces to 20 mm. apart; bracts lanceolate, roundish at base, broadest in the lower quarter, then gradually narrowed into a long, narrow upper part, acuminate at apex, to 20 mm. long, 2-3 mm. broad in flower, to 30 mm. long, 5-6 mm. broad in fruit. Pedicel 2-4 mm. long in flower, elongated to 13 mm. in fruit. Calyx hairy, divided to the base, lobes lanceolate, attenuate in the upper part, 18-23 mm. long, + 2 mm. broad in flower, enlarged to 30 mm. long, 3 mm. broad in fruit. Corolla white to cream, externally glabrous to puberulous or bearing some thin, long, trichomes on the lobes, saa long, 2~3 mm. thick at base, gradually widened upwards, 10-12 sig ee at the top; lobes 2.5-3 mm. broad, recurved. Nectary glabrous, . “ss 0.2-0.4 mm. high. Anthers joined basally and laterally, forming rr gs 11-13 mm. long, sterile tips 1.5—2 mm. long, attached to the pare es mm. above the anther base; filaments ligulate, dilated towards the hs, and decurrent for 5-6 mm., inserted 13-15 mm. above the base 0 corolla tube. Nutlets 3-4 mm. long, smooth, shining. DISTRIBUTION: West Pakistan. 72 (K, PEU). KHYBER AGEN- ; -type); Landi Kotal, cy: Khyber Pass, rocky mountains, 1110 m., Johnston 22 @ ha ca. 1000 m., 1080 m., Lowndes L.709 (£); between ae Kazmi 389b (pes); Landi Kotal to Torkham, Leak if (PEs); a y J. Ahmad in Kazmi s.n. (PES); Wanak, ee reservoir, Kazmi 2699 (pes). WAzIRISTAN AGENCY: give rs azmak, Kazmi 182b (PEs); without locality, Duthie . chr. Gesammt. Naturw. 45: 127. 967. 1875; 4. Onosma griffithii Vatke, Zeits : Riedl in Rechinger, Fl. Iranica 48: 178. 508 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Onosma barbigerum I. M. Johnston, Jour. Arnold Arb. 32: 212. 1951. Tyre: AFGHANISTAN: without locality, Griffith 5947 (B-holotype, GH, K, W-isotypes). Biennial to perennial (?) herb. Stems usually solitary to sometimes few, simple or occasionally with short floriferous branchlets near the top, dense- ly covered with long, stiff, spreading trichomes, to 4 mm. long, arising from tuberculate bases, the area between the bases densely puberulous or covered with thin, minute, spreading trichomes, to 0.2 mm. long. Basal leaves linear, oblong or oblanceolate, entire, roundish to subobtuse, gradu- ally narrowed towards the base, 6-12 cm. long, 4-17 mm. broad, midrib sunken above, raised and prominent below, indument like that on the stems, long trichomes on the upper surface sometimes subappressed; mid- dle cauline leaves lanceolate, ovate-lanceolate to elliptic-oblong, broadest above the base, base usually roundish or rarely attenuated, 3-5 cm. long, to 15 mm. broad; upper cauline leaves reduced in size, subacute. Inflo- rescence of terminal cymes, simple to sometimes forked, scorpioid, 3—6 mm. in diameter in flower, elongated to 12 cm. in fruit; bracts lanceolate, acute. Pedicels very short, to 2 mm. long in flower, erect, densely hairy, elongated, especially those of the lower fruits, to 6(-10) mm. Calyx divided nearly to the base, lobes linear to linear-lanceolate, 15-20 mm. long, 1—-1.5 mm. broad in flower, acute, enlarged to 18-20 mm. long, 2—4 mm. broad in fruit, covered with long spreading trichomes. Corolla yellow, cylindrical-cam- panulate, glabrous or with few scattered trichomes outside, 22-30 mm. long, 2-3 mm. thick at base, gradually widened upwards, becoming 7-9 mm. in diameter below the sinus, lobes broadly triangular. Nectary narrowly lobulate, glabrous, less than 0.5 mm. high. Anthers united basally and laterally to form a tube, 9-12 mm. long, connectives thickened, muriculate; filaments ligulate, 6-9 mm. long, decurrent base 4-7 mm. long, inserted 10-15 mm. above the base of the corolla tube. Style filiform, shorter than the corolla. Nutlets sparsely verruculose, shining, to 6 mm. long. DistriBuTIoN: Afghanistan, West Pakistan. AM, Johnston (1. c.) described Onosma barbigerum from Baluchistan, on the basis of Duthie’s collection 3682 at the Dehra Dun Herbarium. 1971] KAZMI, BORAGINACEAE 509 He distinguished his species by its broadly elliptic-ovate middle cauline leaves with antrorsely appressed trichomes on the dorsal surface and its shorter corolla, compared to the narrow middle cauline leaves with spread- ing trichomes on the upper surface and the longer corolla of Onosma grif- fithii. Both the form of middle cauline leaves and the length of the corolla are very variable in the specimens cited above and O. barbigerum repre- sents only the form with broadest middle cauline leaves. As there is no correlation between the width of leaves, length of corollas and the ap- pressed or spreading type of trichomes, O. barbigerum hardly deserves specific or even varietal recognition. Onosma griffithii is closely related to O. khyberianum and O, hispidum. It is distinguished from the former by having much broader bracts not attenuated above, and from the latter in having much shorter basal leaves. The plants belonging to O. hispidum have strong perennial roots and stems to 75 mm. long, whereas O. grifithii has biennial and short-lived roots with stems never exceeding 35 cm. in length. 5. O. hypoleucum I. M. Johnston, Jour. Arnold Arb. 32: 220. 1951. Tyre: INpIA: without locality, Herb. Late East India Company, Fal- coner, sm. (GH-holotype). Erect perennial herb with a stout woody root. Stems usually several, simple or branched in the upper part, to 50 cm. long, 1-2.5 mm. thick and woody at base, densely covered with thin, white, antrorsely appressed trichomes, 0.1—0.4 mm. long, not arising from tuberculate bases, intermixed sparingly to profusely with long, stiff, spreading trichomes, with tuber- culate bases, 1-1.3 mm. long. Basal leaves drying and disappearing at anthesis; middle cauline leaves sessile, lanceolate to linear-lanceolate, broadest at base, margins slightly revolute, acute, 1-3 cm. Hane, er. broad, nerve on the upper surface inconspicuous, below raised and prom- inent; upper surface covered uniformly with short and long, antrorsely ap- pressed trichomes, lower surface densely covered with short, antrorsely appressed trichomes, with only few long trichomes tuberculate at scare the margins and midrib; upper cauline leaves shorter and narrower. In- florescence usually terminal, sometimes axillary on peduncles, Spacnee long, dense, ca. 2 cm. in diameter in flower, elongated at mahirny : ‘lage times to 12 cm. long; bracts leaflike, smaller, comparatively sree a base. Pedicels slender, densely hairy, to 5 mm. long in flower, to _ long in fruit. Calyx divided to the base, lobes linear, acute, densely hairy, trichomes like those on the stem, equalling or sometimes slightly Jgsg ty the flowers, to 11 mm. long in flower, not or slightly elongated in .. to 13 mm. long. Corolla blue, campanulate, externally deeey . with minute, antrorsely appressed, trichomes, inside villous oes ie ds above the base, 10-11 mm. long, 2.mm. thick at the pele — “68 if the ened upwards, 4-5 mm. in diameter just below the sie : hhects margins slightly revolute, 1-2.5 mm. broad, 1-1.5 mm. “egnitie tube joined basally and laterally forming a tube exserted from the c 510 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 for more than half its length, 9-10 mm. long, sterile tips 1-1.5 mm. long; filaments gradually widened upwards from the base and then gradually narrowed towards the apex, 6-7 mm. long at base, 0.3-0.5 mm. broad, 0.5-1 mm. above the base, 0.5-0.7 mm. broad, villous for the lower 1-2 mm., inserted 1.5—2 mm. above the corolla base. Nectary lobate, 0.1—0.3 mm. high, villous. Style glabrous, to 15 mm. long. Nutlets shining, minutely rugulose, 2.5—3 mm. long. DistrIBsuTION: West Pakistan, Kashmir, northwestern India. West Pakistan: Hazara Dist.: Thandiani, 2550 m., R. R. Stewart 27788 (micu), Nasir s.n. (E), Kazmi 2746 (pes); Kalabagh, ca. 2700 m., Kazmi 2770 (pes); Nathiagali, near Government House, ca. 2200 m., Kazmi 2763 (PES); between Nathiagali and Dungagali, ca. 2700 m., Kazmi 99b (pes); Changlagali to Dungagali, 2500-2600 m., R. R. Stewart 23521 (GH), 23521a (us); Dungagali, R. R. Stewart 28753 (£); Changlagali, 1950 m., R. R. Stewart 2956 (Kk); Aug. 1880, Young s.n. (BM); between Mansehra and Garhi Habibulla, Kazmi 107a (pes); between Garhi Habibullah and Balakot, Kazmi 391 (pes); Kaghan Valley, climb from Kewai to Shogran Rest House, Kazmi 2210a (prs); hills on east of Kaghan Village, 2000 m., Kazmi 2217 (pes); Siran Valley, near Jabori, Kazmz s.m. (PES), RAWALPINDI Dist.: Murree Hills, Upper Topa, 2500 m., Kazmi 756 (pes); on Lower Topa-Patriata Road, near Patriata, Kazmi s.n. (pes); Murree- Kohala road, Kazmi s.n. (PEs). Kashmir: MuzzAFFaraBabD Dist.: Kishanganga Valley, Nilam, Kazmi s.n. (PES). 6. O. thomsonii C. B. Clarke in Hook. f. Fl. Brit, India 4: 178. 1883; I. M. Johnston, Jour. Arnold Arb. 32: 346. 1951. Type: Banahal, southern Kashmir, 6000 ft., Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (K-holotype, c-isotype, as Onosma No. 3). Perennial herb. Stems many, erect or ascending, arising from a loosely branched fruticose caudex, 2-3 mm. thick towards the base, simple or branched, 15-50 cm. long, densely covered with slender, white crisped, spreading trichomes, 1-3 mm. long, arising from minute tuberculate bases, usually intermixed with very short, thin, spreading trichomes. Basal leaves drying and disappearing at anthesis; lower cauline leaves numerous, lan- ceolate to linear-lanceolate, slightly revolute at margins, acute, nerve on the upper surface inconspicuous, raised and prominent below, 5-10 cm. long, 5-10 mm. broad, upper surface grayish-green, covered uniformly with stout, pungent, subpatent, white trichomes, to 1.5 mm. long arising from large tuberculate bases, area between the bases bearing minute, thin, + subappressed trichomes not exceeding 0.2 mm. in length; lower surface white, totally covered with soft, slender, subpatent trichomes, to 3 mm. ong, with long, pungent, stiff trichomes tuberculate at the base only on the midrib and margins; upper cauline leaves usually linear-lanceolate, 2-5 sn long, 2-9 mm. broad, gradually reduced upwards. Inflorescence of terminal scorpioid cymes, when young 1.5—3 cm. in diameter, at ma- turity loosely racemose, 9-10 cm. long, frequently forked; bracts leaf- like much smaller, lanceolate, acute, broadest at the base. Pedicels 1971] KAZMI, BORAGINACEAE 511 slender, hairy, 1-3 mm. long in flower, 3-10 mm. long in fruit. Calyx di- vided to the base, lobes lanceolate, acute, densely villous or villous hispid, 8-10 mm. long in flower, not at all to slightly elongate in fruit. Corolla blue, narrowly-campanulate, 10.5-12 mm. long, 2—2.5 mm. thick at base, widened into an elongate throat, 3-4.5 mm. in diameter, externally densely strigulose, trichomes minute, retrorsely appressed, glabrous inside, throat longitudinally tightly infolded below each corolla sinus, lobes triangular, 1.5 mm. broad, 1 mm. long, tips recurved. Anthers united laterally, form- ing a tube long exserted from the corolla, 7-8.5 mm. long, sterile tips 1-2 mm. long; filaments subulate, broadest near the base, not decurrent, 6-7 mm. long, arising 3—3.5(—4) mm. above the corolla base. N ectary narrowly lobulate, 0.1 mm. high, villulose. Style 12-17 mm. long, glabrous. Nut- lets shining, obscurely rugulose or nearly smooth, 2.5-3.5 mm. long. DIsTRIBUTION: West Pakistan, Kashmir, northwestern India. West Pakistan: Hazara Dist.: Thandiani, 2700 m., Kazmi 392b (prs); be- tween Thandiani and Kalabagh, 2400 m., Kazmi 439b (pes); from Murree north to Nathiagali via Changlagali, 2500-2700 m., Dickason 71 (micH); between Barian and Changlagali, Dickason 70 (micH); Barian-pipeline, between Murree and Changlagali, Dickason 75 (micH); below Kalabagh, 1650 m., R. R. Stewart 28754 (E); Abbottabad to Nathiagali, above Thai barrier, Burtt B492 (r); Ab- bottabad, hill above Kehal near the top, 1500 m., Kazmi 550 (PEs); Abbottabad to Sherwan, ca. 1500 m., Kazmi s.n. (pes); between Dadar and Jabori, Kazmi 589b (pes); Kaghan Valley, Balakot, Jafri s.n. (e); between Balakot and Kewai, Duthie 19479a (cx); Mahandri to Kaghan Village, Kazmi 642a (PES); Shogran to Paya, 3000 m., Kazmi 2486 (pes); Siran Valley, between Jabori and Shahid Pani Rest House, Kazmi s.n. (pes). MALAKAND AGENCY: between Dargai Post and Malakand village, Kazmi s.n. (PES). RAWALPINDI Dist.: Upper Topa to Kashmiri Bazar, Kazmi 2846 (pes); Jhikagali to Lower Topa, Kazmi 108b (PES); Ghoragali to Murree, Kazmi 2829 (pes); Barian, Murree Hills, R. R. Stewart 7123 (Kk); Murree, 20 May 1938, Madan Gopal s.n. (MICH). Swat State: Saidu Sherif towards Maina, 1000 m., R. R. Stewart 27443 (kK). Mis- cellaneous: Punjab, Bassahi, Drummond 26013 (E). shmir: MuzzarraraBaD Dist.: Dhanni to Titwal, 1000 m., R. R. Stewart 17371 (GH); Dhanni, 1130 m., Schmid 1655 (BM, MICH, RAW); between Garhi Habibullah and Muzzaffarabad, Kazmi 290b (PEs). 7. O. bracteatum Wall. in Roxb. FI. Indica, ed. Carey 2 : 13. 1824; DC. Prodr. 10: 66. 1846; C. B. Clarke in Hook. f. Fl. Brit. India 4: 178. Onosma bracteatum Wall. Cat. No. 936. 1829, nomen nudum. 4: 316 Onosma macrocephala D. Don, Prodr. Fl. Nepal. 99. 1825, Gen. Syst. 4: 316. 1838, TYPE: “Shree-nugur” (Srinigar), Kamroop, Wall. Cat. No. 936 (K). Perennial herb. Stems brownish, usually many, arising from a persistent cluster of basal leaves, simple, leafy, to 50 cm. long, 4-6 mm. chick sa base, covered with intermixed short, thin, crisped trichomes not tu 512 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 late at the bases and long, moderately stiff spreading trichomes with large tuberculate bases; short trichomes 0.2—0.6 mm. long, long ones 1.4— 4 mm. long. Basal leaves persistent at anthesis, lanceolate, entire, acute, gradually narrowed towards the base into a winged petiole, 3-5 mm. wide, including petioles 6-8 mm. long, 15-35 mm. broad, veins evident, green above, greenish-white below, both surfaces covered with short and long types of trichomes, to 1.5 mm. long; middle cauline leaves lanceolate, broadly sessile, entire, acute, rounded at the base, 6-8 cm. long, 1—2.7 cm. broad; upper cauline leaves reduced upwards. Inflorescence of terminal glomerate clusters, very dense, 4-10 cm. broad in flower, later becoming lax, revealing the crowded densely flowered, forked individual cymes; bracts hairy, lanceolate, with acuminate, very prolonged, slender tips. Pedicels 3-5 mm. long in flower, 8-12 mm. long in fruit. Calyx divided nearly to the base, lobes linear-lanceolate, acuminate, 15-18 mm. long in flower, covered densely with spreading trichomes to 1-5 mm. long, elon- gated to 25 mm. in fruit. Corolla blue (fide Duthie), bearing loose, long, crisped, slender trichomes externally above the middle, glabrous inside except on the nectary, usually shorter to sometimes equalling the calyx, 13-17.5 mm. long, 3 mm. thick at the base, gradually a little widened up- wards, 4-5 mm. broad below the sinus; lobes triangular, 2-3 mm. broad, 11.5 mm. high, weakly revolute at the margins. Anthers joined only at base, 6-7 mm. long, sterile tips weakly developed, 0.5—-0.8 mm. long; fila- ments 2—2.5 mm. long, attached 5.5-6.5 mm. above the corolla base, decur- rent base swollen, fusiform, 3.5-4 mm, long, marked by fusiform depres- sions on the outside of the corolla. Nectary a weakly lobed collar, densely villous on inner side. Style glabrous, 9-13 mm. long. Nutlets gray, an- gulate, coarsely rugose and tuberculate, 4-4.5 mm. long. DistTRIBUTION: Kashmir, northwestern India. Kashmir: Srinagar, Kamroop in Wallich 936 (kK). Onosma bracteatum is very distinct from the other species of Onosma in our area in having anthers coherent only at the base, corolla usually much shorter than the calyx, and upper cauline leaves and bracts broad- ly lanceolate and narrowed into long attenuated tips. 8. O. dichroanthum Boiss. Diagn. Pl, Orient. Nov. I. 2(11): 107. 1849; Riedl in Rechinger, Fl. Iranica 48: 180. 1967. Onosma setosum B dichroanthum (Boiss.) Boiss. Fl. Orient. 4: 181. 1875. Type: Iberia prope Baku, C. A, Meyer s.n. (G). Biennial or very short lived perennial herb. Stems usually several, erect or ascending, simple or branched above to 75 cm. long, 3-15 mm. thick at base, branches to 20 cm. long, covered densely with very short, thin trichomes, 0.1—-0.2 mm. long, and stiff spreading, unequal trichomes tu- berculate at base, to 5 mm. long. Basal leaves usually dried at anthesis, clustered, linear to linear-oblanceolate, entire, obtuse, 10-20 cm. long, 1971] KAZMI, BORAGINACEAE 513 4-15 mm. broad, covered on both surfaces with trichomes like those on the stems, usually shorter, sometimes hairy only on the midrib and the margins below; middle cauline leaves oblong-lanceolate, obtuse to sub- acute to 10 cm. long, 15 mm. broad; upper cauline leaves gradually re- duced upwards. Inflorescence of terminal racemes, short, scorpioid in flower, 5—6 cm. in diameter, lax, elongate in fruit, 10-15 cm. long; bracts lanceolate, especially the lowermost. Pedicels densely hairy, ca. 5 mm. long in flower, 10-15 mm. long in fruit. Calyx divided nearly to the base, lobes linear to linear-lanceolate, acute, covered densely with long, + antrorsely subappressed trichomes, 15-25 mm. long in flower, elon- gated to 25-30 mm. in fruit. Corolla yellow, cylindrical-campanulate, ex- ternally glabrous except for tuft of hairs near the tip of the lobes, 25—40 mm, long, 2.5-3 mm. thick at base, gradually expanding and becoming 10 mm. thick just below the sinus; lobes deltoid, 2.5-4.5 mm. broad, re- curving. Anthers joined at the base, 9-12 mm. long, sterile tips 1-1.5 mm. long, connectives swollen, muriculate; filaments linear, 3.5-4.5 mm. long, arising 17-20 mm. above the corolla base, base decurrent for about 10 mm. Nectary a slightly lobed, glabrous collar, 0.3-0.8 mm. high. Style 25-33 mm. long. Nutlets pale, shining, smooth or slightly roughened, 4-5 mm. long. DisTRIBUTION: Turkestan, Afghanistan, West Pakistan. West Pakistan: Hazara Dist.: 5 miles from Dadar, on Abbottabad-Dadar road, Kazmi 60b (PEs); Mansehra, Nasir 1659 (k), Kazmi 568b (pes); between Gahri Habibullah and Balakot, Kazmi 2749 (prs); between Abbottabad and Thandiani, Kazmi 853 (pes). Konat Drst.: on top of the pass, between Peshawar and Kohat, 15 miles from Kohat, Kazmi 2651 (PEs). KurraM AcENcy: Amal kot to Sadda, ca. 15 km. from Parachinar, 1400 m., Lamond 1900 (£); without locality, Afandi 244 (PES); Parachinar, Kazmi 2612 (pes), A. Rehman in R. R. Stewart 25906 (Bm); Shalozan, Kazmi s.n. (PEs). MALAKAND AGENCY: between Dargai Check Post and Malakand village, Kazmi 2586 (pes). Quetta Disr.: Gwal, 1800 m., 27 April 1888, Lace s.n. (E); Zandara (Ziarat), 2400 m., Jafri & Akbar 2196 (£); Ziarat, 2100 m., Santapau 6449 (cH); Kach-Ziarat road, 2100 m., R. R. Stewart 550 (mic); 10 miles before Ziarat, on Quetta-Ziarat road, Kaz- mt 1658 (pes). Swat STATE: Mingora, 900 m., S. Ali 26149a (BM). WAZIRISTAN AGENCY: 2 miles to Miram Shah, on Bannu-Miram Shah road, Kazmi 2116 (PES); between Miram Shah and Razmak, ca. 1350 m., open stony hillside and Nullah bank, Lowndes 644 (z). Reported from: Baluchistan, Kawas Pangi, 1800 m., April 30, 1893, Munro s.n. (pp). CHITRAL STATE: Chitral, 1896, Hamilton Sm, (DD), 1894, Younghusband s.n. (pp); Chitral, 1450 m., June 2, 1895, Har- riss sn. (pp); Guirat (Ziarat?) to Chitral, 1450 m., May 31, 1895, Harriss s.n. (pp); Broz, 2540 m., June 6, 1895, Harriss s.n. (pp); Drosh, Chitral, 1350 m., 1908, Toppin 110 (x); Gilgit, Kaghusi, Gilgit Expedition 1650 m., June 30, 1881, Giles 375 (pp); Hazara, Shinkiyari, June 3, 1899, Jmayat s.n. (Dp); Kurram Agency, Shinnak, Kurram Valley, April 1879, Aitchison 28 (c); Parachinar, April 14, 1894, Harriah s.n, (pp), June 30, 1897, Harriah s.n. (Dp). 9. O. limitaneum I. M. Johnston, Jour. Arnold Arb, 32: 354. 1951; Riedl in Rechinger, Fl. Iranica 48: 185. 1967, pro parte excl. var. majus I. M. Johnston. 514 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Onosma limitaneum var. parviflorum I. M. Johnston, Jour. Arnold Arb. 32: 354. 1951. Onosma johnstonii H. Riedl in Rechinger, Fl. Iranica 48: 183. 1967, syn. nov, Type: BALUCHISTAN: Fort Sandeman, without citation of collector’s name 18927 (pp-holotype). the midrib and margins like those on the stem; cauline leaves sessile, linear-oblong, oblong-lanceolate, or oblong-spathulate, acute to obtuse, usually + dilated at the base, 14-45 mm. long, 2-7 mm. broad, In- florescence usually terminal or sometimes lateral, simple or bifurcate, when densely inside and on the margins, with white, stiff, unequal trichomes spreading at the margins and outer surface, appressed inside. Corolla yellow turning purple-yellow-red in time, tubular, externally puberulous, glabrous within, 7-9 mm. long, 1.5-2 mm. thick at base, slightly dilated in the middle, slightly constricted below the sinus, 2-3 mm. in diameter, lobes triangular, longer than broad slightly recurved. Anthers linear, joined only at their bases, ca. 3.5 mm. long, sterile tips 0.5 mm. long, biapiculate, included; filaments gradually Style glabrous, filiform, not at stigma distinct, globular. Nutle ts usually one developing, brownish, shin- ing, 3 mm. long, ca. 1.5 mm. broad, areola minute, transversely elliptic. DISTRIBUTION: Iran, Afghanistan, West Pakistan. The species is very variable; the following t ieti be dis- esuihe wing two varieties may 9a. Var. limitaneum. 1971] KAZMI, BORAGINACEAE 515 Corolla 7-9 mm. long, externally puberulous, at base 1.5—2 mm. thick, 2-3 mm. in diameter below the sinus, lobes 0.8-1.2 mm. long; anthers 2.5-3.5 mm. long, filaments 1.5-2 mm. long, inserted 3 mm. above the corolla base. DISTRIBUTION OF VARIETY: Iran, Afghanistan, West Pakistan. West Pakistan: MAKRAN Dist.: 25 miles from Hushab on Hushab-Panjgur Road, Kazmi 1184 (pes). Quetta Dist.: Quetta to Sibi, 1500 m., Bolan Pass, Lamond 718 (£), Kazmi 1399b (pes), Rechinger 28421 (w); Murdar and Gwal reserve, 1800 m., April 10-12, 1889, Lace 3794 (pp, ©). ZHoxs Dist.: Fort San- deman to Shinghar, 1450 m., barren rocky ground, Kazmi 1931 (pes), Lamond 1459 (£), 2050-2300 m., Rechinger 29878 (w); 30 km. north Fort Sandeman ver- sus Wana, Rechinger 29857 (w). Reported from: BALUCHISTAN: Kaka in Lora Valley, Lace s.n, (?); Gival, 750-900 m., May 6, 1893, Munro s.n. (pp); Mina Bazar, 1440 m., May 11, 1896, without collector’s name 18928 (pp); Gastoi, May 28, 1897, without collector’s name 20600A (pp); Kharwand, June 7, 1897, without collector’s name 20600B (pp); Sin Kachh, May 12, 1897, without col- lector’s name 20600 (pp). 9b. Var. majus I. M. Johnston, Jour. Arnold Arb. 32: 354. 1951; Riedl in Rechinger, Fl. Iranica 48: 186. 1967. Type: AFGHANISTAN: Herb. Late East India Company, without lo- cality, Griffith 5946 (cu-holotype). Corolla 15-17 mm. long, externally sparsely hispidulous only on the veins, otherwise glabrous, 2—2.5 mm. thick at base, 5~6 mm. in diameter below the sinus, lobes 1.5 mm. long; anthers 5.5—6.5 mm. long, filaments 35-4 mm. long, inserted 6—8 mm. above the corolla base. DISTRIBUTION oF vARIETY: Afghanistan, West Pakistan. _ West Pakistan: Quetta Dist.: Quetta to Chaman, between Qilla Abdullah and Sheila Bagh, 1600-1900 m., Lamond 1046 (), Kazmi 1542a (Es); between Quetta and Bazuzai, ca. 1500 m., Kazmi 1538 (pes); Chiltan mountains, near Quetta, Kazmi s.n. (PES). Miscellaneous: Baluchistan, Stocks 997 (pp, GH). 30. Trichodesma R. Br. Prodr. Fl. Nov. Holl. 496. 1810, nom. cons. Pollichia Medic. Bot. Beobacht. 247. 1783. Borraginoides Moench, Meth. 515. 1794. Leiocarya Hochst. Flora 27: 30. 1844. Streblanthera Steud. Flora 27: 29. 1844. Spiroconus Stev. Bull. Soc. Nat. Moscou 24(1): 576. 1851. Boraginella Siegesb. ex O. Kuntze, Rev. Gen. Pl. 2: 437. 1891. Tyre species: T. zeylanicum (N. L. Burm.) R. Br. typ. cons. Perennial, rarely annual, herb with woody base. Lower leaves usually Opposite, upper leaves alternate. Pedicels axillary, 1-flowered, going off into terminal racemes by the gradual reduction of the leaves. Calyx 5-fid to S-partite, lobes imbricate, ovate to lanceolate, attenuate to cuspidate at 516 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 apex, often dilated, cordate or sagittate at the base, enlarged in fruit. Corolla subrotate, hypocrateriform to infundibuliform, faucal appendages lacking, lobes ovate, cuspidate at apex. Stamens 5, anthers large, oblong to linear-lanceolate, convergent into a cone, apices aristate, awns con- torted, filaments short, inserted on the corolla tube. Ovary 4-locular, conic-ovoid, not divided in flower. Style filiform, stigma small. Nutlets ovoid, triangular or subglobose, smooth, rugose to hirsute, marginate to emarginate. Species 40 to 45, distributed in tropical and subtropical Africa, Asia, and Australia. KEY TO THE SPECIES a. Nutlets emarginate to indistinctly marginate. b. Nutlets glabrous, smooth or rarely slightly rugulose, 4-5 mm. long; calyx sagittate at base in flower, auriculate in fruit, lobes gradually nar- | SR ae pre inside alee aes eae aaa 1, T. indicum. trichomes, arising from tuberculate bases, area between the bases pu- berulous to covered with short, thin trichomes. d. Lower cauline leaves short-petiolate, upper cauline leaves sessile, el- liptic-oblong, obtuse stocksit 1, T. indicum (L.) R. Br. Prodr. Fl. Nov. Holl. 149. 1810; A. DC, in j BB Prodr. 10: 172. 1846; Boiss. Fl. Orient. 4: 280. 1675; CB, Clarke in Hook. f. Fl. Brit. India 4: 153, 1883; Brand, Pflanzenr. IV. 252(Heft 78): 38, 1921; Riedl in Rechinger, Fl. Iranica 48: 226. 1967. Trichodesma perfoliatum Wall. Cat. No. 932. 1828, nomen nudum, tr. Asiat. Soc, Bengal 21: 175, 1853. Type: Habitat in India Orientale, Herb. No. 188.2 (LINN). Icon.: Wight, Illustr. Indian Bot. 2: t. 172, fig. 1-15. 1850. Erect annual herb, Stems profusely branched, branch reading, to 50 cm. long; densely covered A aoe oe Ss with thin, crisped, white trichomes to 0.2 1971] KAZMI, BORAGINACEAE 517 mm. long, intermixed sparsely to profusely with long, stiff, slender, tri- chomes to 0.5 mm. long, arising from tuberculate bases. Basal leaves dry- ing at anthesis; cauline leaves usually opposite, oblong, lanceolate or rarely ovate, entire, acute to subacute, narrowed, dilated, or amplexicaul at base, to 7 cm. long and 25 mm. broad, covered on both surfaces in various den- sities with slender, spreading trichomes to 2 mm. long, arising on the upper surface from prominent tuberculate bases, on the lower surface from inconspicuous ones; middle cauline leaves usually larger than the lower and upper cauline leaves. Inflorescence terminal, lax, few flowered; bracts leaflike. Pedicels slender, pubescent, to 25 mm. long in flower, not at all to very slightly elongated in fruit. Calyx divided nearly to the base, lobes oblong, gradually narrowed from the base towards acute apex, at base sagittate-auriculate, to 12 mm. long and 15 mm. broad at base in flower, slightly enlarged in fruit. Corolla pinkish-blue, infundibuliform, slightly surpassing the calyx, 12-14 mm. long, lobes semiorbicular, spread- ing, abruptly narrowed to filiform apices. Anthers densely villose, aristate, awns equalling the anthers, glabrous, strongly contorted. Nutlets ovoid, glabrous, smooth or rarely slightly rugulose, white to bluish, 4-5 mm. long. DISTRIBUTION oF sPEcIES: Afghanistan, West Pakistan, Kashmir, India, Philippines, and Mauritius. A very variable species in its habit, leaf shape, and density of tri- chomes on the stem and leaves. The following two varieties may be dis- tinguished. la. Var. indicum. Stem and leaves densely covered with trichomes; cauline leaves broad, more or less dilated at base, not amplexicaul. DIsTRIBUTION OF VARIETY: As that of species. : Marpan Dist.: D ew Th tre 2088 es d to Swat, 800 m., Lamond 1648 (£), ] : Islamabad, Para- 400 m., Kazmi s.n. (PES). RAWALPINDI DIsT.: charian Hill, Kazmi 2835 (pes); Rawalpindi to Kahuta by Saon Road, Burtt 518 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 1110 (£); Topi Park, Kazmi s.n,. (pes); Dhamyal near Rawalpindi, 500 m., 30 March, 1940, R. R. Stewart s.n. (GH); near Chaklala Railway Station, S. H. Abid in Kazmi s.n. (ves); environs of Rawalpindi, Schlagintweit 10892 (k). Swat State: Swat Valley, 1100 m., Weatherhead 3, 96 (Bm); Kalangai, Weath- erhead 15 (pm); inter Mingora et Khwazakhiela, 850-950 m., Rechinger 30519 (w); Buttkhela, 900 m., M. A. Siddiqui 27112 (BM, RAW). Miscellaneous: Mera to Marri, 1500-1800 m., Schlagintweit 11509 (GH); Punjab, without locality, Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (GH). Reported from: Las Bela, Hughes-Buller, 18777; Baluchistan, Bsit. Blater & Hallberg s.n. (?). Kashmir: Domel, Jhelum Valley, 800-900 m., Ludlow & Sherriff 8060 (cH); between Baramulla and Kohala, Oct. 19, 1913, Evershed s.n. (BM); Mirpur, Kaz- mi 517 (PES); Muzzaffarabad, Kazmi 305a (pes). Miscellaneous: Western Hima- laya, without locality, Duthie s.n. (A). ib. Var. amplexicaule (Roth) Cooke, Fl. Bombay Presidency 2: 215. 1904. Trichodesma amplexicaule Roth, Nov. Pl. Sp. 104. 1821; C. B. Clarke in Hook. f. Fl. Brit. India 4: 153. 1883. T. indicum var. amplexicaule (Roth) Sedgwick, Rec. Bot. Surv. India 6(8): 350. ¢. XZ. 1919, Type: Not seen. Stem sparsely covered with trichomes; leaves with the upper surface uniformly covered with trichomes with large tuberculate bases; the lower surface with similar trichomes confined to midrib and veins; cauline leaves broad at the base, + subcordate, subamplexicaul. DISTRIBUTION OF vaRIETY: South India, Ceylon, West Pakistan. West Pakistan: KARACHI Dist.: Malir Cantonment, Burtt 131560 (£); Mango Pir, A. Rehman 25713 (Bm); Malir, Jafri 2534 (E). 2. T. incanum (Bunge) A. DC. in DC. Prodr. 10: 174. 1846; Bunge, Mém. Acad. Pétersb. Sav. Etr. 7: 417. 1854; Boiss. Fl. Orient. 4: 282. 1875; Brand, Pflanzenr. IV. 252(Heft 78): 36. 1921, excl. var. griffithii & {. aitchisoni; M. Pop. F). URSS 19: 691. 1953; Riedl in Rechinger, Fl. Iranica 48: 223. 1967. Friedrichsthalia incana Bunge, Del. Sem. Hort. Dorpat. 7. 1843. T; richodesma molle A. DC. in DC. Prodr. 10: 174. 1846. Spiroconus glaucus Stev. Bull. Soc. Nat. Moscou 24(1): 577. 1851. Trichodesma strictum Aitch, & Hemsl. Jour. Linn. Soc, Bot. 19: 178. 1882. Boraginella incana O. Kuntze, Rev. Gen. Pl. 2: 436. 1891. Boraginella stricta (Aitch. & Hemsl.) O. Kuntze, ibid. Borasinlla mollis (DC.) O. Kuntze, ibid. richodesma molle DC. var. stri i V. mS? Ce ee ictum (Aitch. & Hemsl.) Brand, Pflanzenr. I Tyre: In collibus prope Samarkand, Lehmann s.n. (LE, P). Perennial herb or undershrub. Stems erect, ascendent to sometimes pro- 1971] KAZMI, BORAGINACEAE 519 cumbent, profusely branched, branches to 100 cm. long, covered densely with thin, slender, crisped, retrorsely subappressed, white trichomes 0.2- 0.5 mm. long, intermixed sparsely with thick, pungent, spreading trichomes to 1 mm. long, arising from tuberculate bases. Basal leaves drying at an- thesis, lower cauline leaves sessile, ovate, elliptic to oblong, entire, acute, dilate-rotundate to sometimes subcordate at base, 2.5-10 cm. long, (10-)15-35(—60) mm. broad, sparsely covered above with white, slender, spreading to subappressed trichomes, 3-5 mm. long, more frequently in- termixed with longer trichomes with tuberculate bases, lower surface densely covered with short spreading trichomes, less frequently inter- mixed with longer ones; upper cauline leaves shorter and narrower. In- florescence terminal, simple or bifurcate, lax, few-flowered; bracts minute, linear. Pedicels slender, hairy, 2.5-3 cm. long in flower, not elongated in fruit. Calyx divided 2/3 its length, lobes ovate, long-attenuate towards the acute apex, roundish at base, densely pubescent, 14-20 mm. long in flower, slightly enlarged and dilated at base in fruit. Corolla blue, tube —8 mm. long, limb patent, 20-25 mm. in diameter, lobes ovate, abruptly narrowed to the linear apex. Anthers villous, aristate, awns glabrous, + equalling the anthers. Nutlets emarginate to inconspicuously marginate, irregularly strongly rugose and minutely tuberculate on the dorsal surface, 5-6 mm. long. DistripuTion: Iraq, Iran, Afghanistan, West Pakistan, Turkestan. West Pakistan: Kurram AcEeNncy: Kurram Valley, Ali Khel, 2250 m., 1879, Aitchison 864 (cH). Very common in Afghanistan and Turkestan, Trichodesma incanum is known from West Pakistan only from the collection cited above. It is distinguished from the other species of our area by its densely pubescent, sessile, ovate, acuminate leaves and its rough, rugose, and tuberculate nutlets. 3. T. africanum (L.) R. Br. Prodr. Fl. Nov. Holl. 496. 1810; Lehm. Asperif, 195. 1818; DC. Prodr. 10: 173. 1846; C. B. Clarke in Hook. f. Fl. Brit. India 4: 154. 1883; Boiss. Fl. Orient. 4: 1875; Brand, Pflanzenr. IV. 252(Heft 78): 29. 1921; Riedl in Rechinger, Fl. Tranica 48: 220. 1967. Borago africana L. Sp. Pl. 138. 1753. oe Borrago verrucosa Forssk. Fl. Aegypt.-Arab. 41. 1775. Pollichia africana Medic. Bot. Beobacht. 247. 1783. Boraginoides africana Hiern, Cat. Afr. Pl. Welw. 3: 721. in Trichodesma africanum var. homotrichum Brand, Pflanzenr. 30. 1921. 8. 1. 252 (Heft 78): Type: Habitat in Aethiopia, Herb. No. 188.4 (LINN). 520 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Annual or biennial herb with woody base. Stems fistulous, usually many, erect to ascending, usually branched, branches to 100 cm. long and 1 cm. thick at the base, densely covered with prickly, stiff, thick trichomes, 0.5— 2 mm. long, with large calcareous tuberculate bases, area between the bases glabrous, trichomes falling with age. Basal leaves petiolate, lamina ovate, lanceolate-ovate to rarely oblong, acutish, including petioles 2.5- cm, long, 12-30 mm. broad, midrib sunken above, prominent below, densely covered on the upper surface, sparsely so on the lower surface (ex- cept the nerves) with trichomes more or less like those on the stem; cauline leaves short petiolate, subopposite; upper cauline leaves sessile, gradually reduced in size upwards. Inflorescence terminal or axillary, lax. Pedicels slender, densely covered with trichomes, very short to 5 mm. in flower, elongated to 15 mm. in fruit. Calyx divided nearly to the base, lobes broadly lanceolate, broadest at the truncate or subcordate base, gradually attenuate towards the prolonged acute apex, 5~6 mm. long and 2.5—-3 mm. broad, with a prominent midrib, enlarged to 10 mm. long, 5-6 mm. broad, with conspicuous cordate-auriculate base and shorter tips in fruit; external- ly densely covered on the midrib and the margins with stiff, spreading tri- chomes, and with soft, thin, antrorsely appressed trichomes within. Co- rolla bluish-white to light blue, small, tube shorter than the calyx, lobes triangular ovate, + twice as long as the tube, attenuate, acute, subpatent at apex. Anthers villous, awns glabrous. Style shorter than the calyx. N utlets brownish, triangular ovate, carinate, dorsally tuberculate, margined, margins plane, denticulate, 2-4 mm. long. Distripution: Africa, Arabia, Iran, Afghanistan, West Pakistan, India. West Pakistan: Dapu Dist.: Karachi to Dadu, Thano Bula Khan to Kotri, Lamond 1818 (£), Rechinger 28653 (w), Kazmi 1406b (PES), open slopes, La- mond 817 (g). Katat Disr.: Sibi to Quetta, below Mach, rocky roadside slopes, Lamond 876 (£); infra Mach versus Ab-i-Gum, Rechinger 28482 (w). MAKRAN Dist.: 20 km. w. Hushab, Rechinger 28026 (w), Kazmi 946b (pes). Miscel- laneous : Sind, Stocks 375 (GH). Reported from: Sind, Stocks 575; Thano Bula Khan, Woodrow; Laki, Woodrow (Cooke, Fl. Bombay Pres. 2: 216. 1904). 4. T. stocksii Boiss, Diagn. Pl. Orient. Nov. II. 3(3): 140. 1856; FI. Orient. 4: 282. 1875. Burkill, Fl, Pl. Baluchistan 51.1909: Riedl in Rechinger, Fl. Iranica 48: 222. 1967. a Pakistan: “In regione Beloutschistan,’ Stocks s.n. (c-holotype, Perennial herb. Stems decumbent, or more long; orescence subcorymbose, 3—5-flowered: bracts minute, lin- ear-lanceolate. Pedicels long in fruit. Calyx 10-12 mm. long, lobes cordate- 1971] KAZMI, BORAGINACEAE 521 rotundate at the base, enlarged in fruit. Corolla campanulate, hardly longer than the calyx, lobes short ovate, abruptly acuminate and recurved at apex. Anthers densely lanate, aristate, awns slightly contorted. Nut- lets pitcherform, subglobose, margins membranaceous, strongly inflexed. DISTRIBUTION: South Iran, West Pakistan. West Pakistan: Gandava, lower Baluchistan, Stocks s.n. (G, K). I could not examine the type or any other specimen of Trichodesma stocksit from Pakistan or Iran. The description given above is according to Boissier (/. c.). 5. T. longipedicellatum Rech. f. & H. Riedl in Rechinger, Fl. Iranica 48: °222, 1967, Type: Pakistan: Inter Awaran et Turbat, in alveo lapidoso, 100-200 m., 11. IV. 1965, K. H. Rechinger 27760 (w-holotype). Icon.: Riedl, 1. c. t. 38, 39. 1967. Annual (?) or perennial herb. Stems flaccid, angulate, herbaceous, branched, branches axillary, spreading, to 60 cm. long, covered with thin, crisped, retrorsely subappressed, white trichomes, to 0.2 mm. long, not tuberculate at base, sparsely intermixed with thick, prickly, spreading trichomes to 0.5 mm. long, arising from tuberculate bases, falling with age. Leaves petiolate, lanceolate to ovate-lanceolate, entire, obtuse to subacute, attenuate at base, including the petioles 2-9 cm. long, 7-40 mm. broad, densely covered on both surfaces with thin, short trichomes, intermixed profusely on the upper surface and sparsely on the lower surface with thick trichomes similar to those on the stem; upper cauline leaves short- petiolate to sessile, narrower and shorter, usually acute. Inflorescence terminal or axillary, lax, few-flowered; bracts minute. Pedicels flexuous, densely hairy, 12-15 mm. long in flower, elongated to 20 mm. in fruit. Calyx divided to the base, lobes broadly lanceolate, gradually attenuate towards the prolonged acute apex, roundish at base, 8.5-9.5 mm. long and 3.5—4.5 mm. broad above the base in flower, densely covered with short trichomes, intermixed usually at the midrib and margins with thick trichomes tuberculate at the base, triangular ovate, to 12 mm. long, 9 mm. broad, slightly connate at base in fruit. Corolla mauve, white and almost tan in the center, ca. 18 mm. in diameter in dried specimens, tube shorter than the calyx, lobes spreading obtuse, ca. 7 mm. long. sg densely white-villous, aristate, awns contorted, glabrous. Style in after anthesis 10-15 mm. long, stigma very small. Nutlets pitcherform, brown, 7 mm. long, 6-6.5 mm. broad, margins strongly inflexed, goles a central aperture 2 mm. in diameter and radially slightly plicate, iio ovate, ca. 5 mm. long, dorsal middle area inside the pitcher smoo tuberculate, DistrRIBuTION: West Pakistan. 522 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 West Pakistan: MAKRAN Dist.: Inter Awaran et Turbat, 100-200 m., Rech- inger 27760 (w); ca. 20 km. west of Hushab, 100-200 m., rocky slope, sandy soil, Lamond 391 (£); 17 miles from Hushab on Hushab-Turbat Road, Kazmi 1147 (PEs). Rechinger & Riedl] (/. c.) distinguished Trichodesma longipedicellatum from the other closely related species by these characters: “Species habitu elongato, caule flaccido, indumento brevi canescente, foliorum forma et petioli longitudine valde insignis.” From a comparison of the descriptions of T. longipedicellatum and T. stocksii there appears little difference be- tween the two species except that T. longipedicellatum is tall, robust with flaccid stems, and leaves with longer petioles, lanceolate to lanceolate- ovate, with acute apex. These characters are very variable, and even in the three collections cited above can hardly be relied upon. It appears that T. longipedicellatum is only a tall form of T. stocksii, growing under bet- ter conditions. On examination of the type of T. stocksii and other col- lections, T. longipedicellatum may prove to be only a variation of T. k . stocksii. | To be concluded | : Journal of the Arnold Arboretum ~ Published quarterly in January, April, July, and October by the Arnold _ Arboretum, Harvard University, Rs Subscription Price $16.00 per year, beginning January, 1972. Subscriptions and remittances should be sent to Miss Dulcie A. Po Arnold Arboretum, 22 Divinity Avenue, Cambridge, Massachusetts 02138, Volumes I-XLV, reprinted, are available from the Kraus Reprint Cor- oe poration, 16 East 46th Street, New York, New York 10017, U.S.A. JOURNAL OF THE ARNOLD ARBORETUM VoL. 52 OcTOBER 1971 NUMBER 4 COMPARATIVE ANATOMY OF ULMACEAE ! Epwarp M. SWEITZER ULMACEAE Is A FAMILY of trees and shrubs of temperate and tropical regions. It is usually divided into two distinct subfamilies or tribes, the Ulmoideae or Ulmeae and the Celtidoideae or Celteae (as Celtideae). Leaves are always simple, often oblique at the base, and mostly alternate except in Lozanella where they are opposite. Stipules are caducous. Flow- ers are bisexual in Ulmus or functionally unisexual in most other genera. They are solitary, cymose, or in axillary fascicles. The perianth is made up of 4 to 8 connate or distinct sepals which are imbricate, campanulate, and persistent. A hypanthium is present in most genera of Ulmeae, but it is usually not apparent in Celteae. Stamens are differentiated into a bilocular anther supported by a filament; they are erect in the bud and inserted at the bases of the calyx lobes or opposite the sepals. The ovary is bicarpellate with 1 or 2 locules, and 2 styles stigmatose along their inner faces. The solitary ovule is anatropous and pendulous from near the apex of the ovary. The fruit is a winged samara with a straight em- bryo or a drupe with a curved embryo. Seeds lack endosperm. A list of genera, their taxonomic characteristics, and geographical ranges is given in Taste 1. Distribution maps for genera and species can be found in Bernard (1906). Also, a list of genera and the number of species considered by Hutchinson, by Engler and Prantl, and by Willis is given in Taste 2. Hutchinson’s (1967) listing and arrangement of gen- era will be used because it is the most complete modern listing for the family. (For Taste 1, see p. 572.) Palynologically the tribes Celteae and Ulmeae are more or less well de- fined (Erdtman, 1966). The pollen grains in Celteae are mostly tas 3-porate, suboblate to oblate-spheroidal, with an arcoid streaked sexine (some tegillate). Pollen grains of Ulmeae are 4- to 6-porate, colpate, “5 rupate, oblate to suboblate, with an arcoid streaked sexine (some tegillate). Seeds of the Ulmeae are flattened with a straight embryo; fruits ah usually samaras. Seeds of the Celteae are globose with the rag oat surrounded by the bony endocarp of a drupe (McClure, 1957). * Contribution number 4367 of the Maryland Agricultural Experiment Station, Department of Botany, University of Maryland. 524 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 genera studied by Martin (1946), Planera, Ulmus, and Zelkova anes in- vesting embryos, Celtis has a folded embryo, and Trema a bent hi oa Sax (1933) gives the haploid chromosome number for C eltis, Holop es Ulmus, and Zelkova as 14. However, Darlington and Wylie (1955) lis the haploid chromosome number for Celtis as 10, 11, 14. nae . TABLE 2. Genera and number of species in Ulmaceae GENUS HvuTCHINSON * ENGLER & PRANTL” WILLIs Ampelocera 9 2 : Aphananthe 8 3-4 ; Barbeya 1 41 Celtis 150 60 ; Chaetachme 1 ‘ Chaetoptelea 5 Gironniera 30 7-8 ; Hemiptelea 1 9 Holoptelea 2 1 ; Lozanella 3 Mirandaceltis 1 , Paraspo 7 2 ; Phyllostylon 3 1 i Plagioceltis ‘ Planera 1 1 : Pteroceltis 1 1 Trema 55 30 14 Ulmus about 45 16 31 Zelkova 11 4 4 * 1967. » 1893, * 1966. In the fossil record, the fa Cretaceous (Berry, been identified fro Hemiptelea, mily is reported in strata identified from the and that genera of Ulmaceae have been ast , Spitzbergen Island, northern Europe, Asia, an America. The present known distribution for each genus of Ulmaceae is given in TABLE 1. However, the genera may be arranged according to their presence in major geographical regions as follows: North America — Celtis, Planera, Trema, and Ulmus; Central and South America — Ampelo- cera, Celtis, Chaetoptelea, L celtis, and Trema: Euro » f arasponia, Pteroceltis, Trema, and Zelkova. A recent paper by Elias (1970) discusses the taxonomy of Ulmus, Celtis, Planera, and Trema of the southeastern United States. *For tables 1 and 4a, 4b, and 4c, see pages 572-584. 1971] SWEITZER, ULMACEAE 525 ECONOMIC USES AND IMPORTANCE Commerci i : edi ee a Ulmus and Celtis are valuable timber trees. ee ict tc et al. (1964), the wood is used for boxes and crates sige si “gate because of its strength, excellent bending qualities, . pe 1 ek to withstand shock. “The murderous tomahawks of the fea, ca. ort a were largely made of Elm wood” (Hutchinson aa, Phollost yl = sets er — used in the making of small cat: ee sis has been used for weaver’s shuttles, scales [ea ge arn 1948). Chaetoptelea is used in Central Amattcn ago . sa ies and for frames and wheels of vehicles; Trema wood & Hess eon poy aire : eer fine gunpowder charcoal (Record ’ . nn ; : Asia se Oth : eae are employed for timbers in Saar — of the Ulmaceae include the following: Trema has an sepetaae % shade coffee and cocoa trees, and the fibers from bark have . or cordage (Record & Hess, 1943). Fleshy drupaceous fruits Nile axe san Celtis are used for food; seeds of Holoptelea integri- foros ea . in India. Ulmus rubra (slippery elm) and Ulmus glabra a. . z e m) bark has been used in folk medicine decoctions because >. ra mucilage content (Kraemer, 1902; Hiscox, 1955); for example, "aes el par to be soothing in case of inflammation of the mucous eeouicl sna prepared by pouring one cup boiling water over two tea- . S ees powdered or whole slippery elm bark, cooling, straining, any a lemon juice, and adding sugar to taste (Hiscox, 1955). The he of Ho optelea integrifolia has been used as a remedy for rheumatism ndia (Capoor, 1937). Elms have also been used locally as sacred symbols in France (Chevalier, 1942). ay species of Ulmus have been used in the United States and else- where as street and park trees. Some species of Celtis and Zelkova are also used as ornamentals (Rehder, 1940). With the outbreak of Dutch elm disease in 1931 and its subsequent spread, many elm trees have died or are now dying. The organism Ceratocystis ulmi causes the disease on all native elms in the United States as well as on all European species; however, susceptibility to the disease varies. Symptoms include wilting and browning of leaves and subsequent defoliation; secondary xylem darkens with gummy accumulations. Some other diseases of Ulmus include black ea which infects young leaf spot caused by the fungus Gnomonia ulm d drop off; and dieback caused ts wounds causing canker for- Morsus ulmi usually results d by the fungus Verticillium leaves: wetwood caused by the bacterium Erwinia nimipressuralis may result in wilting and defoliation, among other symptoms. Celtis occidentalis ( hackberry) may have broom- like growths on branches, a condition called witches’-broom. The fungus Sphaerotheca phytophylla and species of the gall mite Eriophyes are 526 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 usually found in the buds, and the resulting branches that develop are dwarfed and clustered (Linder, 1931; Pirone, et al., 1960; J. C. Walker, 1969; Nichols, 1970). TAXONOMY The family name, Ulmaceae, can be traced back to Mirbel (1815), at which time the family included only Ulmus and Celtis. Dulac (1867) used Samaraceae as the family name. Link (1831) proposed splitting ’ Ulmaceae into two families: Ulmaceae, composed of Ulmus and related genera, and Celtaceae (as Celtidaceae), comprising Celtis and related genera. For the most part, the family Ulmaceae has been considered a natural taxon. There has been much uniformity both in the generic content and in its placement with relationship to other families. Planchon (1873) recognized the order (family) Ulmaceae which he divided into two sub- orders (subfamilies): Ulmeae containing Ulmus, Holoptelea, Hemiptelea, Zelkova, and Planera and Celteae (as Celtideae) with Celtis, Parasponia, Sponia, Gironniera, Aphananthe, and Chaetachme. Endlicher (1837) listed the orders (families) Ulmaceae, with Ulmus and Planera and Celteae (as Celtideae) containing Celtis, Sponia, and Mertensia, in his class Juliflorae. Bentham and Hooker (1862) arranged the tribes Ulmeae, Celtideae, Cannabineae, Moreae, Artocarpeae, Conocephaleae, Urticeae, and Thelygoneae in the family Urticaceae. Lindley’s (1853) alliance (our present day order) Urticales included Urticaceae, Cannabaceae (as Cannabinaceae), Moraceae, Artocarpaceae, etc.; however, he placed the Ulmaceae, including Celtis and Ulmus, in the alliance Rhamnales along with Aquilariaceae, Rhamnaceae, Hippocrateaceae, Celastraceae, Sapota- ceae, and other families. , Moreae, Artocarpeae, and Cannabineae. Eichler (1878) assigned the Ulmaceae (Ulmus, Celtis) to the taxon Ur- ticinae which included Urticaceae, Moraceae, Artocarpaceae, Platanaceae, and Ceratophyllaceae. divided Ulmaceae into two subfamilies: Ulmoideae, containing Phyllo- , Zelkova, Trema, Parasponia, A phananthe, Giron- ales was placed in the subclass Ar- SHC eee petalae). Engler and Prantl had the Meine § in ependently from the hypothetical Protangiospermae; eichior, in the 1964 “Syllabus,” followed the same pattern. 1971] SWEITZER, ULMACEAE 527 Hallier (1912) derived the Terebinthales from the Ranales and had the Terebinthaceae giving rise to the Urticaceae (in the broad sense of Bentham and Hooker, 1862). Bessey (1915) placed the Ulmaceae (Ulmus, Celtis, Zelkova, Planera) along with the Moraceae and Urticaceae in the Malvales. He derived the Malvales from the Ranales. Wettstein (1935) placed the Ulmaceae, Moraceae, Cannabaceae, Eucommiaceae, Rhoiptelea- ceae, and Urticaceae in the Urticales. He traced this order from the same gymnospermous line which produced Fagales and Juglandales and thought that the Hamamelidales arose close to the Urticales. Rendle (1938) placed the Ulmaceae (Ulmeae and Celteae (as Celtideae), Urticaceae, and Moraceae in the order Urticiflorae, but the overall position of the order is similar to that proposed by Engler and Prantl. Gundersen (1950) put Ulmaceae, Eucommiaceae, Moraceae, and Urti- caceae in the Urticales and placed the order in his “Ulmus group” which included Proteales, Santalales, Balanopsidales, Fagales, Leitneriales, and Casuarinales. He derived this group from his “Rosa group” (Rosales, Hamamelidales, Thymelaeales, and Myrtales) which he in turn traced from the complex containing Magnoliales, Ranales, and Piperales. Gun- dersen considered his ‘“Ulmus group” to be, probably, an unnatural taxon. Hutchinson (1967) divided the Ulmaceae into two tribes: Ulmeae, con- taining Phyllostylon, Holoptelea, Planera, and Ulmus, and Celteae (as Celtideae), with Celtis, Pteroceltis, Ampelocera, Zelkova, Hemiptelea, Trema, Parasponia, Aphananthe, Mirandaceltis, Gironniera, Chaetachme, and Lozanella. He included the family in the Urticales along with Canna- baceae (as Cannabiaceae), Moraceae, Urticaceae, Barbeyaceae, and Eu- commiaceae. Hutchinson (1969) indicated that “. . . the above families ... have a quite lengthy pedigree, maybe as follows: Magnoliales os Dilleniales > Rosales > Hamamelidales > Fagales > Urticales climax family Urticaceae (mostly herbaceous).” a ; Thorne (1968) placed the Ulmaceae, with subfamilies Ulmoideae and Celtidoideae, in Urticales under the superorder Malviiflorae_(Malvales, Rhamnales, Euphorbiales, Solanales, Campanulales). Cronquist (1968), likewise, put Ulmaceae in the order Urticales along with Barbeyaceae, Moraceae, Cannabaceae, and Urticaceae. He placed the order in span class Hamamelidae which included Trochodendrales, Hamamelidales, Eu- commiales, Leitneriales, Juglandales, Myricales, Fagales, and Casuarinales. The Hamamelidae have arisen from the Magnoliidae (Magnoliales, Piper- ales, Aristolochiales, Nymphaeales, Ranunculales, P eo jan (1969) placed Ulmaceae, Moraceae, Cannabaceae, and mergers oe the Urticales which he has connected to the Hamamelidales and he Is Perhaps derived from it. The Hamamelidales are derived directly from the Magnoliales through the Trochodendrales. - eagles has viet, a renewed suggestion a sp gent Mai split into two families, as formerly suggested by eke Skaya (1967) states: In its main diagnostic characters the to the family Moraceae than to Ulmoideae; for e subfamily Celtidoideae is more similar this reason it should be ex- 528 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 cluded from the family Ulmaceae. The distinct features isolating Celtidoideae from the closely allied families justifies the classification of this subfamily as a separate family Celtidaceae Link. The genera Zelkova and Hemiptelea generic composition of the two families is as follows — Ulmaceae: Ulmus, Holoptelea, Hemiptelea, Zelkova, Planera, Phyllostylon; Celtidaceae: Celtis, Pteroceltis, Trema, Parasponia, Lozanella, Aphananthe, Gironniera, Chaetach- me, Ampelocera. Since there is general acceptance that the Ranalian complex is primitive in the dicotyledons, there has been little disagreement as to the position of the Ulmaceae in the order Urticales. Cronquist (1968) asserts that “The naturalness of the order Urticales is not in dispute.” The only major point of disagreement among the four leading contemporary plant phy- logenists, Cronquist, Hutchinson, Takhtajan, and Thorne, relates to the position of Urticales. In Thorne’s treatment it is included with Malvales and related taxa, whereas the others place it close to the Hamamelidales and related taxa. In a recent study, Dickison and g Barbeya in a separate family, and most taxonomists to the genus from paleobotanical in obtaining anatomical material for study. ANATOMICAL REVIEW 1971] SWEITZER, ULMACEAE 529 cera glabra was investigated by Milanez (1937); Record and Garratt (1925) and Desch (1948) described the secondary xylem anatomy of Phyllostylon brasiliensis. Record and Mell (1924) and Record and Hess ( 1943) described some features of the wood anatomy of several genera, in- cluding Ampelocera, Celtis, Chaetoptelea, Phyllostylon, Planera, Trema, Ulmus. In a work entitled “Phylogénie des Ulmacées,” Houlbert (1899) discussed the need of using “echantillons d’age avancé” in phylogenetic studies based on wood anatomy He studied a complete range of Ulmus woods of various ages and attempted to correlate yearly differences in Ulmus with mature wood characters of related genera to show evolutionary development. Major summary works dealing with secondary xylem anat- omy and general anatomy of some genera of Ulmaceae include those of DeBary (1884), Solereder (1908), and of Metcalfe and Chalk (1950). Works pertaining to aspects of anatomy other than wood structure vary in their completeness. A study of the leaf anatomy of 14 genera and 80 Species of the Ulmaceae was carried out by Priemer (1893). Leaf ontog- eny of Zelkova serrata was investigated by Soma (1965). Gardiner (1965) attempted to correlate leaf anatomy and geographic distribution in Trema. Petiolar anatomy of Celtis and several other urticalean genera was studied by Sarabhai and Saxena (1961). Bechtel (1921) published work on the floral anatomy of the Urticales, including Ulmaceae (five species of Ulmus, one species of Celtis). Sev- eral studies have been made of megasporogenesis, microsporogenesis, and embryology; for example that of Shattuck (1905) on Ulmus americana; of Capoor (1937) on Holoptelea integrifolia; of Eckardt (1937) on Ulmus montana, Celtis tournefortii and genera of related pseudomonomerous families; and of R. I. Walker (1938, 1949) on various species of Ulmus. General summary works dealing with embryology which include some genera of the Ulmaceae are listed in the work of Maheshwari (1950), of Eames (1961), and of Davis (1966). The embryo sac of Celtis is of the Polygonum type and that of Ulmus of the Adoxa or Drusa type. The secondary phloem anatomy of Ulmus, Celtis, Gironniera, Trema, and Zelkova was investigated by Zahur (1959). Sieve tube ontogeny in Ulmus americana has been investigated with the light microscope (Evert et al., 1969) and with the electron microscope (Evert & Deshpande, 969). SECRETIONS AND CHEMOTAXONOMY Various intracellular substances, such as silica, calcium carbonate, and occur in different organs, tissues, and Holoptelea, Phyllostylon, Trema, Ulmus, dence of silica formation. Priemer (1893) reported and hairs of the species of Zelkova, A phananthe, Ulmus, that the leaf epidermis Celtis, and Trema 530 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 i i ilica he investigated were silicified. Various techniques for prene se 2 are contained in Satake (1931), Uber (1940), Amos in Prete et al. (1958), and Jones and Handreck (1967). A eee baie be physiology can be found in Lewin and Reimann (1969). : peerenier i analyses for the presence of silica were beyond the scope o - rted we presence of calcium carbonate in old vessels of diese pt in Haberlandt (1928) for Ulmus glabra and Celtis nicl a as a ee xylem deposits of calcium carbonate were noted in severa are ie Celtis, and Phyllostylon by Record (1927). sere nigh coe 8 the presence of calcium carbonate in the form of cysto ‘ ee of several ulmaceous genera. Satake (1931) found calciu San cystoliths in Celtis and Trema leaves. Pireyre (1961) peipasas wipes from morphologic, histologic, and physiologic bases in two _ (4950) listed as well as other non-ulmaceous genera. Metcalfe and es pie nals: various genera reported to have calcium carbonate deposi ere = for the determination of calcium carbonate can be found in ? seo pa Scott (1946), and Pireyre (1961). Santi analyses of c onate were beyond the scope of this stu y. ae be ; Mucilage — been identified in the past by its ACO MNOS A, sectioned tissues, the chemical nature of this substance has paervon investigated and the microscopist has had to rely on a sner observations rather than specific histochemical tests. The on pape replete with reports on mucilaginous and gummy materials in signe bark, and roots of several ulmaceous genera and species, especia ‘6 set rubra and U. glabra bark which has a high mucilage content, La 1950), and Hutchinson (1969). one 59) discussed plant gum and mucilage c eo | implications of their work are yet to i apparent ubiquitous occurrence of dit : tances in Ulmaceae, a brief consideration of gum and mucilag : point. Mucilages are pearls ve ism and may serve as food reserves. There is no clea operties. Taste 3 shows Smith and ei 1on of gums and mucilages. The =e . fulva) mucilage are D-galactose, 3-0- methyl-D-galactose, L-rhamnose, and D-galacturonic acid (Robinson, 1963). In the only comparative chemotaxonomi (1965) discussed the relationships within noids. He specifically analy are noted in the Discussi1o c study encountered, Lebreton the Urticales based on flavo- zed species of Celtis and Ulmus and his results N. 1971] SWEITZER, ULMACEAE 531 TABLE 3. Classification of gums and mucilages * Group I Group II Acidic gums and mucilages Neutral gums and mucilages containing: containing: Acid components Neutral components L-guluronic acid Hexoses Hexoses D-glucuronic acid Pentoses 6-Deoxyhexoses D-galacturonic acid 6-Deoxyhexoses Pentoses Sulfate groups Sugar alcohols Sugar alcohols Phosphate groups Ethers of all of the Ethers of all of the Ethers of all of the ve above above “From Smith & Montgomery (1959). PROBLEMS AND OBJECTIVES A study of the wood and leaf anatomy should provide additional infor- mation for phylogenetic studies of Ulmaceae and at the same time help to clarify ordinal relationships and intrafamilial systematic problems. The latter include the taxonomic status of the two major subdivisions of Ulmaceae and of the genera Chaetoptelea, Mirandaceltis, and Giron- niera, Anatomical data will be correlated with other pertinent information to provide a firmer basis for an interpretive discussion of the affinities and composition of Ulmaceae. eee In addition, leaf and wood anatomy can be used in the identification of species from local floras, which might be useful in the identification of commercially important ulmaceous timbers, possible drug plants, and po- tential ornamentals. A knowledge of wood anatomy is imperative in some aspects of forensic botany. Ulmus species and varieties vary 1n their sus- ceptibility to Dutch elm disease and elm phloem necrosis. These are, at least in part, diseases of the vascular system, and sistance may be related partially to differences 1n the vascular tissue, which a thorough anatomical study might reveal, making possible a more effective breeding program for disease resistance. MATERIALS AND METHODS For the most part, the terminology of xylem anatomy wee here en agreement with that suggested by the Committee on Nomenc ature a : International Association of Wood Anatomists ( SER A Ettin s- describe overall venation patterns in leaves are taken from a ss hausen (1861). Additional terms potas % iy. re et from such usage is explained where employed In the **"- : Standard texoniguin were used in preparing specimens - le Dried wood specimens were hydrated in igh deeb ot ana rved in ethyl alcohol, and embedded in celloidin. Wood specimens prese 532 JOURNAL OF THE ARNOLD ARBORETUM [von. 52 formalin-acetic acid-alcohol (FAA) were dehydrated and embedded in celloidin. Transverse, radial, and tangential sections were cut on a sliding microtome, stained with Heidenhain’s iron-alum haematoxylin, and counterstained with safranin. Wood macerations were prepared using Jefirey’s fluid, washed in water, stained with safranin, and dehydrated with tertiary-butyl alcohol. Sections and macerations were mounted in Canada balsam. Leaves preserved in FAA and dried leaves from herbarium specimens were used in this study. All dried leaves were first hydrated in boiling wa- ter; clearings of leaves were made following Arnott’s (1959) technique using 5 percent sodium hydroxide, followed by a water wash, and treat- ment with a saturated aqueous solution of chloral hydrate. Leaves were then washed with water, stained with aqueous safranin, dehydrated in ethyl alcohol, and mounted in Canada balsam. Transverse and paradermal sections of leaves were embedded in paraffin and cut on a rotary micro- tome. Sectioned leaves were stained with safranin and fast green, cleared in clove oil and toluene, and mounted in Canada balsam. _ Ab croscope slides of wood specimens employed in this study were Prepared in the laboratory of Dr. Oswald Tippo. Prepa- cept that nomenclature of that of species indigenous t Microchemical tests wer in cells and tissues of Ulmaceae. When these terms are us eh greement with reports from the literature; silica has a glass-like appearance, mucilage is viscous to frothy in texture, and cal- cium carbonate stains to varying degrees with safranin and fast green. (For Tastes 4a, 4b, and 4c, see pages 574-584.) RESULTS The sequence of genera for the anatomical descriptions of leaf and wood 1971] SWEITZER, ULMACEAE on , ni 4 jlack on the Ficure 1 Zelkova sinica. Cross section of leaf ¥ with a ‘mineral pla “sence GBC i ie af with a upper epidern mis, X 320. 2, Celtis eS Coot ES 4 ex . cystolith in the upper epidermis and a compact si pli mes zed Een upper , S Ulmus parvifolia. Cross section of eat ga a pegless C , ur pa ermal cell below a mineral plack, X 320. 4 haetachme microcae a. ection of — with an masioe saper Pientld cell con . lege x3 we 534 JOURNAL OF THE ARNOLD ARBORETUM [von. 52 follows Hutchinson (1967), except for Chaetoptelea which he included in Ulmus. In the generic descriptions which follow, the number of species, in parentheses, is followed by the number of specimens if there were more specimens than species examined. LEAF Since several descriptive terms refer to structures and substances which siliceous refers to any material which is glass-like in appearance; the term mineral plack refers to the silica and/or calcium carbonate covering an epidermal area composed of several cells, i.e., both on the surface and between the cells (FicureE 1), : Cystolith refers to an anisotropic crystalline concretion of both calcium carbonate and silica (Pireyre, 1961) of varying shapes and colors which may or may not have a peg. A peg is a cellulosic extension of the cell wall on which the crystalline material is deposited. Pegged cystoliths usually become stained (safranin and fast green), whereas the pegless cystoliths remain glassy to white-opaque even after staining (FIGURES 25 Mucilage refers to cellular contents which appear viscous to frothy in texture and vary in color from gray-green to pink when stained with Safranin and fast green (Ficure 4). The terms loose and compact, as applied to the spongy mesophyll, refer to the relative amount of intercellular space. Also, a loose spongy meso- phyll has relatively fewer cells per unit volume than a compact spongy mesophyll (Ficurrs 5 2) The following leaf family with i bi Ower epidermis is papillate. A parenchymatous bund . Sheath surrounds both major and minor veins; however, surrounding — of the larger veins, there are also gelatinous thick-walled parenchyma cells. Unicellular trichomes have a bulbous base (Ficure 7) and occur on th the upper and lower epidermis. The walls are usually impregnated irs usually found on both leaf surfaces are = : the first found in some species of Phyllostylon, lanera, Ulmus, Chaetoptelea, Celtis, Ampelocera, Zelkova, Hemiptelea, Trema, Aphananthe, and Lozanella comprises a uniseriate linear row of on P73 os) wn 1971] S SWEITZER, ULMACEAE Ficure 5. U/ X 320. 6. i Slabra. Cross section of leaf with a loose spongy mesophyll, roceltis tataranowii. Paradermal section of leaf with an anomocytic 0. 7, Ulmus glabra. 0. 8, U Cross = section ; : of leaf with a bulbous-based unicellular hair, X lular trichomes, X 430. 536 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 cells (3-8) with the terminal cell and sometimes stalk cells being glandular, that is, containing dark staining material (FicuRE 8) and in addition some of the hairs are biseriate throughout their length in Trema. The second type has a uniseriate linear stalk of 3 to 8 cells but a multicellular glandular head of 2 to 4 cells and is found in some species of Ulmus, Celtis, Ampelocera, Parasponia, Gironniera, and Lozanella (FicurE 9). The third type has a multicellular stalk with a single enlarged terminal cell in some species of Ulmus and Trema (Ficure 10). Trichomes are usu- ally associated with veins Phyllostylon (2). Multicellular hairs found only on the lower epi- dermis; overall venation pattern brochidodromous, and terminal tracheids in areoles are surrounded by a parenchyma sheath. Leaf margin is ser- rate in Phyllostylon rhamnoides and entire in P. brasiliensis; palisade mesophyll is 1 to 4 cells deep, the spongy mesophyll is compact; bundle sheath extensions not well developed in P. brasiliensis but conspicuous in P. rhamnoides. Cystoliths present in enlarged cells of both upper and lower epidermis in P. rhamnoides. There are about 60 per mm.” in the upper epidermis and 40 per mm. in the lower epidermis. In P. brasiliensis they are only in the lower epidermis and number about 100 per mm.2 Druses are found in the mesophyll and phloem of both species; in addition, P. rhamnoides has prismatic crystals in the bundle sheath and phloem and styloids in all living tissues. Holoptelea (1). Leaf margins are entire, the overall venation pattern is brochidodromous. In the areole, the veins terminate as individual tracheary elements without a sheath (Ficure 11). A uniseriate palisade layer and a medium compact spongy layer are present. Chlorenchymatous bundle sheath present around both major and minor veins; chlorenchy- matous bundle sheath extensions on both sides of major veins. Cystoliths present in enlarged cells of the lower epidermis and in bases of unicellular hairs. Cystoliths approximately 30 per mm.”; druses 0C- curring in cells of upper and lower epidermis, mesophyll, and phloem. Planera (1). Multicellular hairs on the lower epidermis only, leaf margins serrate, the overall venation pattern is “simply” craspedodromous; veins terminating in areoles as individual tracheary elements without 4 sheath; mesophyll composed of a palisade layer 1 or 2 cells thick and a loose spongy layer; a bundle sheath extension present on both sides of major veins; mineral placks present on upper epidermis; druses present in the mesophyll; styloids in cells of both mesophyll and epidermis. Ulmus (15, 27). Multicellular trichom U. americana, U. crassifolia, U. glabra, U. and U. thomasii. Leaves oblique at base, es found in all species except minor, U. laevis, U. parvifolia, margins varying from simply to wn w “I 197 971] SWEITZER, ULMACEAE ss section of gr showing a short nner x 400, 10, Ulmus rubra. Cross te RE 9, Ulmus carpinifolia. Cro rminal al: x Fic st alked trichom ne with a mu ulticellular head, a ; how 12. Ulmus americana. ae with a bundle sheat th, 538 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 triply serrate; venation pattern is “simply” craspedodromous; terminal vein endings enclosed by a conspicuous parenchyma sheath within the areole (FIGURE 12); palisade mesophyll 1 to 4-seriate; the spongy meso- phyll loose; bundle sheath extensions present on major veins. Enlarged mucilage-containing cells occur in both upper and lower epidermal layers; mineral placks present in the upper epidermis. In several species, e.g., U. parvifolia and U. glabra, some upper epidermal cells below placks contain pegless cystoliths. Druses mostly in mesophyll and bundle sheath cells and to a lesser extent in cells of phloem and bundle sheath extensions; prismatic crystals commonly in the bundle sheath and phloem cells and to a lesser extent in cells of bundle sheath extensions; styloids present in cells of mesophyll and epidermis; mucilage-filled lacu- nae, resulting from cell disintegration found below midrib and other large veins in most species (FicurE 13); these lacunae may be traumatic in origin. Chaetoptelea (1). Leaf margin serrate, overall venation pattern is “simply” craspedodromous; in the areole, terminal tracheary elements are surrounded by a bundle sheath; mesophyll comprising a palisade lay- er of 1 or 2 cells and a loose spongy layer; bundle sheath extensions present only on major veins. Mineral placks found covering clusters of cells in upper and lower epidermis; druses found in cells of mesophyll, bundle sheath, and bundle sheath extensions; prismatic crystals present in phloem, bundle sheath, and bundle sheath extension cells; styloids occurring in mesophyll and epidermal cells. Celtis (17, 28). Multicellular trichomes occur only in Celtis iguanaea, c. pubescens, C. reticulata, and C. sinensis. The leat margins vary from entire in C. occidentalis and C. schippii to serrate in C. triflora and C. pumila. Species with serrate margins have teeth mostly from the middle of the leaf to the tip. Celtis laevigata, C. jessoensis, and C. reticulata have oblique leaf bases. Overall venation pattern brochidodromous except in C. philippinensis where it is “perfect” actinodromous. Most species lack epidermis thickened tourne fortii. Pegged cystoliths present in enlarged upper and lower epidermal cells 1971] SWEITZER, ULMACEAE 539 and in unicellular hairs, these present only in the upper epidermis of C. laevigata, C. occidentalis, C. philippinensis, C. pubescens, C. pumila, C. sinensis, C. tenuifolia, and C. tournefortii, and in the upper and lower epidermis of all other species. Pegless cystoliths found in enlarged upper and lower epidermal cells of C. schippii; the number of cystoliths per mm.? ranging from 40 to 148 in the upper epidermis and from 0 to 48 in the low- er epidermis. Prismatic crystals present in bundle sheaths of most specimens and in phloem of C. caudata, C. iguanaea, and C. spinosa; no prismatic crystals present in C, pallida, C. schippii, and C. trinervia; styloids found in living cells of all tissues of C. pallida; mucilage not found in C. iguanaea, C. Spinosa, or C. trinervia. Pteroceltis (1, 2). Leaf margin serrate and overall venation pattern brochidodromous; terminal tracheary elements surrounded by a sheath of round to oval cells; mesophyll divided into a palisade layer (1 or 2- seriate) and a loose spongy layer; major veins with bundle sheath ex- tensions. Pegged and unpegged cystoliths occurring in enlarged cells and hairs of the upper and lower epidermis; cystoliths in distinctive clusters of 20 (+) cells, usually surrounding a trichome; druses found in cells of the mesophyll, bundle sheath, bundle sheath extension, and phloem; Styloids present in the mesophyll. Ampelocera (3). Multicellular hairs occurring on the lower ae of Ampelocera cubensis: short stalked hairs with a glandular, multicellu 7 capitate head occurring in indentations of upper and lower cpedem 0 A. hottlei (Ficure 14). No multicellular hairs found on A. edentu “i : Leaf margin entire and leaf base slightly oblique; overall ee ni pattern brochidodromous; terminal tracheary elements having a $ hyll in A. hottlei, but not in A. cubensis or in A. edentula. Palisade geasde with 1 or 2 layers, the spongy mesophyll varying at ante h in A to compact in A. cubensis; bundle sheath extensions found only : cubensis. f th Pegged cystoliths in enlarged cells of the uppé! j (1 S40 sl mm.) and ie the lower epidermis of 4. a eae mm.2); druses in the mesophyll, bundle sheath, and pores hottlei and prismatic crystals in bundle sheaths of 4. egewiue ane #- also in phloem of the latter; styloids in bensis. r epidermis of all species 1 or 2 palisade layers and a loose spongy Sheath extensions. ; Se is isu Pegless cystoliths in upper and lower epidermal cells; mineral p er 5 We. sane * Row = *; ~ * rf e = ee FicuRE 13, Ulmus rubra. filled cavities. «x Cross section of midrib of a 5 ee élocera hottlei multicellular hair in an indentatio lamarckiana. Cross l 16, L n of the lower epidermis, 450. LS, Treme section of leaf illustrating the uneven lower epidern ‘Imus americana. Transverse wood section illustrating the tan (wavy) pore arrangement, < 120. 11S Lid, y gential 1 1971] SWEITZER, ULMACEAE 541 on the upper epidermis; prismatic crystals in the phloem, bundle sheath, and bundle sheath extensions; styloids in all living tissues of Zelkova carpinifolia and Z. sinica. Hemiptelea (1, 2). Multicellular hairs on the lower epidermis; leaf margins serrate, the overall venation pattern “simply” craspedodromous; terminal tracheary elements in the areole surrounded by a sheath: two or three palisade layers and a loose to medium compact spongy layer; bundle sheath extensions on upper and lower sides of major veins. Pegged cystoliths in upper and lower epidermis; in upper epidermis cystoliths arranged in groups with a unicellular hair in the center (32 to 120 cystoliths per mm.?); cystolith groups abundant and easily visible to the unaided eye, appearing as dark spots on leaf surface; on the lower epidermis 1 to 20 cystoliths per mm.2 Druses occurring in cells of phloem, mesophyll, bundle sheath, and bundle extension; styloids present in living cells of all tissues. Trema (5). Multicellular trichomes comprising a linear row of cells on both leaf surfaces of all species except in Trema amboinensis and T. canescens occurring only on the lower epidermis. Multicellular stalked trichomes with an enlarged terminal cell on both surfaces of T. /amarck- iana, T. micrantha, and T. orientalis. ‘ P Leaves with an entire margin, overall venation pattern “perfect” acro- dromous in 7. canescens, and “imperfect” acrodromous for all ofher species; terminal tracheary elements in the areole lacking a bundle sheath; mesophyll is divided into 1 to 4 palisade layers; a loose spongy ele Present (except compact in T. amboinensis and variable in T. lamarck- jana; bundle sheath extensions present on major veins. ee Lower epidermis quite uneven in most species due to Stuart er and valleys formed by veins (FicuRE 15); Trema orientalis wit ee ‘ oe ridges on both epidermal layers; pegged cystoliths present in sil : cells and in unicellular hairs of the upper and lower epidermis 9 i lamarckiana and T. micrantha but only in the upper epidermis of te HG boinensis, T. canescens, and T. orientalis. Cystoliths per aed on from 27 to 48 for the upper epidermis and from 13 to . wet ar druses found in mesophyll, bundle sheath, bundle sheat cae iu Phloem cells; prismatic crystals in phloem and bundle sheath; in living cells of all tissues in T. micrantha. d overall venation pattern areole dle sheath, in the lower ); druses present in the cells of the mesophyll, bun 542 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 and phloem; mineral placks covering several cells in both upper and lower epidermis. Aphananthe (2). Unicellular hairs barbed; leaf margins serrate, over- all venation pattern “simply” craspedodromous in Aphananthe aspera and brochidodromous in A. philippinensis; terminal tracheary elements in the areole without a sheath; mesophyll with 1 or 2 palisade layers and a loose spongy layer; major veins have bundle sheath extensions. Pegless cystoliths found in enlarged cells of both upper and lower epi- dermis; druses in cells of the mesophyll, bundle sheath, and phloem; Aphananthe philippinensis with prismatic crystals in the bundle sheath, bundle sheath extension, and phloem; mineral placks in both upper and lower epidermis. Mirandaceltis (1, 2). Some unicellular hairs barbed; leaves with en- tire margin and overall venation pattern brochidodromous; terminal tracheary elements lacking bundle sheath in the areole; mesophyll divided into 1 or 2 palisade layers and a tight spongy layer; bundle sheath ex- tensions either absent or weakly developed. Periclinal walls of the upper epidermis, and to a lesser extent of the lower, thickened; pegless cystoliths present in enlarged cells or hairs on upper epidermis; druses present in cells of mesophyll, bundle sheath, and phloem; prismatic crystals found in the two latter tissues; styloids found in the mesophyll. Gironniera (5). Gironniera parvifolia and G. nervosa with some barbed unicellular hairs; multicellular trichomes on lower epidermis only in G. nervosa. Leaf margins entire and overall venation pattern brochidodromous; terminal trachery elements in the areoles surrounded by round to oval parenchyma cells; mesophyll divided into 1 or 2 palisade layers and a oose spongy layer; major veins with a bundle sheath extension. Pegless cystoliths in enlarged cells of lower epidermis; druses found in cells of mesophyll, bundle sheath, bundle sheath extension, and phloem; prismatic crystals found in cells of the bundle sheath, bundle sheath ex- tension, and phloem of G. cuspidata, G. nervosa, and G. nitida; styloids in mesophyll and epidermal cells of G. nervosa. Chaetachme (1). chidodromous; termin Enlarged cells containing mucila : : ge in upper epidermis; druse and prismatic crystals found in bundle sheath heaths. Lozanella (1). Leaf Margin serrate, overall venation pattern brochido- 1971] SWEITZER, ULMACEAE 543 dromous; terminal tracheary elements surrounded by a parenchyma sheath in the areole; mesophyll composed of 1 or 2 palisade layers and a loose spongy layer; major veins with bundle sheath extensions. Cuticular ridges present on both upper and lower epidermis; pegless cystoliths in enlarged cells and in hairs of both upper and lower epider- mis; druses in cells of the mesophyll, bundle sheath, and phloem. Woop Secondary xylem of Ulmaceae is variable, but the following features are common to all members. Vessel elements ligulate to eligulate; inter- vascular pitting alternate, pits elongate to circular in outline; fibers with simple to slightly bordered slit-like pits. Unilaterally compound pitting common between both axial and ray parenchyma and vessel elements. Pores angular to circular; vascular rays uniseriate, biseriate, and multi- Seriate; ray cells thick walled in all species. Phyllostylon (1). Wood diffuse porous; pore distribution 79 percent solitary, 19 percent radial multiples of 2 to 4 cells, and 2 percent clusters of 3 cells; average pore diameter 53 » with a range of 35 to 70 p; vessel element wall thickness from 1.5 to 4 p; perforations simple; end wall angles from 57 and 90°; inter-vascular pits ranging in diameter from 6 to 7; average length of vessel elements 187 » with a range of 135 to 240 »; fiber wall thickness from 4.5 to 5.5 pw; average fiber length 903 » with a range of 690 to 1215 ph ; All a, rays homocellular and comprised of ape reens: uniseriate rays ranging in height from 3 to 11 cells and from 45 to 2 se biseriate and multiseriate rays ranging in height from 7 to 65 rs . ae from 100 to 900 p, and ranging in width from 2 ve 4 cells and from to 35 4; prismatic crystals observed in some ray cells. : Axial ie wba abies and confluent; vessel-ray “7 — parenchyma pitting alternate, pits ranging in diameter from 2.5-3 Holoptelea (2, 3). All woods diffuse porous; pore ‘apiece le to 80 percent solitary, 18 to 35 percent radial multiples of he oe 2 to 4 percent clusters of 3 to 11 cells; average pore pa ae ; wall from 92 to 110 y, the total range of diameters from 50 to SEB thickness of vessel elements ranging from poke . "0 a 90°: ty- simple; end wall angles of vessel elements ranging ain t length rang- loses in vessel elements of H. grandis; average vessel e pret 0 370 p. ing from 273 to 284 p, the total range of lengths from th from 1169 to Fiber wall thickness from 2 to 3.5 m; average oe 1428 » with a total range of lengths from 925 to ; a rays com- Homocellular uniseriate, biseriate, and multiseriate became with the Prised of procumbent cells; uniseriate rays few ey neing in height Storied biseriate and multiseriate rays. Uniseriate ray Fe Pasais rays from 2 to 12 cells and from 60 to 225 ; biseriate and m 544 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 ranging in height from 6 to 57 cells and from 95 to 750 »; width of rays from 2 to 5 cells and from 10 to 55 pn. Axial parenchyma vasicentric, aliform, and confluent in all specimens; pitting of vessel-ray and vessel-axial parenchyma alternate, pit diameters ranging from 2.5 to 6 up. Planera (1, 3). Two specimens of Planera aquatica (USw21789 and Yw6699) diffuse porous; Yw11169 ring porous. Pore distribution 32 to 40 percent solitary, 18 to 35 percent radial multiples of 2 to 4 cells, and 39 to 46 percent clusters of 3 to 9 cells. Average pore diameter 38 » for the diffuse porous specimens, with a total range of 20 to 55 ». Average pore diameter of 34 » for the ring porous specimen and a range from 20 to 55 » for the early portion of the ring, an average pore diameter of 25 » and a range from 20 to 35 » for the last formed portion of the ring. Vessel wall thickness from 0.5 to 3 »; perforations simple in all three specimens; however, vestigial bars also occurring in the ring porous specimen. End wall angles of vessel elements varying from 24 to 90°; intervascular pit diameters from 5 to 7 a; average vessel element length from 180 to 450 ,; all vessel elements usually with spiral thickenings. Imperforate tracheary elements comprised of fiber-tracheids with slightly to prominently bordered pits and vascular tracheids; many vas- cular tracheids have spiral thickenings. Average lengths of fiber-tracheids from 817 to 885 ,; total range of lengths from 375 to 1485 ». Some storying of fiber-tracheids occurs in Yw11169. Vascular rays homocellular; uniseriate rays ranging in height from 2 to 19 cells and from 50 to 400 »; biseriate and multiseriate rays ranging in height from 5 to 63 cells and from 85 to 1135 p are from 2 to 6 cells and from 15 to 75 » in width; prismatic crystals observed in some ray cells, which are sometimes enlarged. Axial parenchyma diffuse, banded, terminal, and vasicentric; crystals and dark staining contents observed in some axial parenchyma cells; ves- sel-ray pitting and _ vessel-axial parenchyma pitting alternate, pits rang- ing in diameter from 2 to 3 pb. Ulmus (33; 55); Wood specimens examined ring porous; pores often oriented in wavy tangential lines (Ficure 16); pore clusters and solitary pores most numerous and radial distribution; decrease in occur; in U. americana radial multi pore distribution. Number the number per cluster from 3 to 28 1971} SWEITZER, ULMACEAE 545 Average pore diameter ranges from 53 p» in U. pumila, to 211 » in U. glabra in the first formed portion of the ring, and from 19 pw in U, rubra to 53 in U. major in the last formed portion; vessel wall thickness ranges from 1.5 to 8 p. Mostly simple perforations in all species; Ulmus laevis, U. major, U. rubra, U. serotina, and U. stricta having scalariform perforation plates with one bar (Ficure 17); Ulmus alata, U. americana, U. crassifolia, U. glabra, U. hollandica var. major, U. laevis, U. macrocarpa, U. major, U. procera, U. pumila, U. rubra, U. scabra, U. serotina, U. stricta, U. thomasii, and U. wallichii having vestigial bars on perforation plates. Only vessel elements of small diameters with vestigial bars or scalariform perforations (FicurE 18); vessel element end wall angles from 18 to 90°; most Species with at least some vessels with 90° end walls; in general, the greater the diameter of the vessel element, the closer the end wall angle to 90°. Ulmus alata, U. carpinifolia var. suberosa, U. glabra, U. laciniata, U. macrocarpa, U. nitens, U. rubra, U. scabra, and U. stricta all with some Opposite intervascular pitting; diameters of the intervascular pits from §,5 to 10 BP. Average vessel element lengths range from 170 p in U. thomasii to 381 p in U. major; the range of all vessel element lengths varying from 111 to 481 »; tyloses found only in U. crassifolia and U. scabra. Imperforate tracheary elements include vascular tracheids and fibers; wall thickness of fibers ranges from 1.5 to 5.5 ; average lengths of fibers range from 779 » in U. carpinifolia to 1782 » in U. montana; total range of lengths for all specimens is 660 to 2146 p. Vascular rays of all species homocellular, comprised of procumbent cells; unicellular rays ranging in height from 2 to 21 cells and from 20 to 0». Biseriate and multiseriate rays ranging in hei and from 50 to 1375 m; these rays ranging in widt and from 10 to 200 p; vessel-ray pitting mostly alternate; however, Species with some opposite pitting are: U. alata, U. carpinifolia var. 7 berosa, U. glabra, U. hollandica var. major, U. macrocarpa, U. major, U. nitens, U. rubra, U. scabra, and U. stricta. Diameter of vessel-ray pits ght from 3 to 76 cells h from 2 to 19 cells . thomasii, most species showing termina axial parenchyma pitting mostly alternate with some 0 Pee U. U~. carpinifolia var. suberosa, U. glabra, U. laciniata, U. gents 4 major, U. nitens, U. plotii, U. rubra, U. scabra, U. serotina, U. stricta, U. suberosa, and U. thomasii ‘ ; : . bra, Storying of vessel elements and some fibers occurs 1n U : gediag Age ‘ - Scabra, and U. serotina. Prismatic crystals present 1n oat in oe americana, U. major, U. montana, U. plotit, U. rubra, U. pote acetal boldii, U. thomasii, and U. uyematsii; prismatic crystals i. = pe ip Parenchyma cells of U. laciniata, U. major, and U. scaora. 546 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 i orm FIGURE 17, Ulmus mtseerg Radial fags section illustrating a scalarifc bericht plate with one b 30 sh yn ar, X 730. Imus serotina. Radial wood sectic owing a perforation i ee V i 30. h crystal stran ds, : U Imus rubra. Tangential spiral tiichediugs on walls oe the vessel elements, X 330 1971] SWEITZER, ULMACEAE 547 parenchyma containing prismatic crystals (FicuRE 19) in all species ex- cept U. laciniata, U. montana, U. procera, U. scabra, and U. stricta. Dark Staining deposits in many vascular ray cells and axial wood parenchyma cells. Spiral thickenings on the walls of small vessels and vascular tracheids of all species (Ficure 20). Chaetoptelea (1, 2). The names Chaetoptelea mexicana Liebm. and Ulmus mexicana (Liebm.) Planch. are synonymous; some botanists rec- ognizing Chaetoptelea as a distinct genus and others not. For this rea- son, specimens bearing both binomials are treated in the same section here under Chaetoptelea. Wood ring porous; pore distribution 80 percent solitary, 12 percent radial multiples of 2 to 4 cells, and 8 percent clusters of 3 or 4 cells. Average pore diameters for the first formed portion of the growth ring 150 » with a range of 25 to 215 ». Average pore diameters for the last formed portion of growth ring 90 », with a range of 40 to 200 y; vessel wall thickness from 1 to 7 pw; perforations simple; end wall angles range from 44 to 90°; intervascular pit diameters from 3.5 to 6.5 p. Average vessel element length is 300 » with a total range of 148 to 407 m; tyloses found in C. mexicana. : Imperforate tracheary elements include vascular tracheids and fibers. Fiber wall thickness from 1 to 3 p; average fiber length is 1660 p» with a total range of 1221 to 2072 up. yi Vascular rays homocellular, comprised of procumbent cells. ibarecaae rays ranging in height from 2 to 7 cells and 25 to 120 ». Biseriate an multiseriate rays ranging in height from 4 to 48 cells and 60 to 725 Ms in width from 2 to 8 cells and 15 to 75 ». Dark staining materials foun in some ray cells. Axial parenchyma vasicentric, aliform, conflu : ange ray and vessel-axial parenchyma pitting alternate; pits ranging 1n Dc from 3.5 to 6.5 »; prismatic crystals and dark staining deposits 1n s Tay cells; crystalliferous strands present. ent, and terminal; vessel- ter Porous species having 42 to Clusters of 3 to 30 cells, and 0 to 12 percent ra ae cells (Ficure 21). Diffuse porous species have 59 to “UP dial multiples of 2 to 4 ent solitary 548 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 pores, 9 to 49 percent radial multiples of 2 to 10 cells, and 0 to 6 percent clusters of 3 to 8 cells (FicurE 22). Pore diameters for the early part of the growth ring in ring porous species range from 58 to 191 » with a total range of 20 to 265 y»; for the last formed portion of the ring diameters range from 20 to 40 » with a total range of 15 to 100 ». Average pore diameters for diffuse porous species range from 49 to 169 » with a total range of 15 to 225 pu; vessel element wall thickness ranges from 1.5 to 6 y; all diffuse porous species having only simple perforations; all ring porous species, except Celtis lindheimeri with exclusively simple perforations, having at least some perforations with vestigial bars in addition to simple perforations; Celtis jessoensis, C. laevigata, and C. pumila also having scalariform plates with one bar in addition to the vestigial bars; end wall angles varying from 10 to 90°; all species having at least some vessels with 90° end walls. Celtis bungeana and C. reticulata with some opposite intervascular pit- ting; pitting crowded in C. iguanaea, C. kraussiana, C. pacifica, C. philip- pinensis, and C. wightii; intervascular pit diameters ranging from 2 to 9 »; average vessel element lengths ranging from 170 to 455 ; total range of lengths 111 to 666 »; tyloses found in C. iguanaea, C. laevigata, C. lind- heimeri, C. reticulata, and C. ritiensis. Vascular tracheids present in ring porous species, absent in diffuse porous species; fiber wall thickness ranging from 1 to 5.5 p; average fiber length range is 717 to 1709 » with a total range of 518 to 2220 uw. It is difficult to determine whether the uniseriate rays are homocellular, heterocellular, or both; in tangential section some of the component cells appear upright, but in radial section all cells are procumbent, radial di- mensions exceeding the axial dimensions. There is a problem in correlat- ing the appearance of a ray in radial section with its appearance in tan- gential section; however, if these so-called “upright” cells in tangential view were truly upright, they would not have been so scarce in radial sec- tion. Therefore, the uniseriate rays will be designated homocellular, even though some may be technically heterocellular. Whenever a similar con- dition is encountered in other genera, for the sake of uniformity in this paper such a uniseriate ray type is referred to as the Celtis type. Some multiseriate rays classed as heterocellular are conceivably homocellular, as is oreaie sete analysis of the cellular composition of rays from mk = orienta is beyond the scope of this paper. fia" and mu liseriate rays heterocellular with procumbent cells in € body of the ray and upright cells on the margins; height of unicellular 1971} >W Z SWEITZER, ULMACEAE 549 é Ficurr ‘ E Re FiO Weled paice , ._ Celtis reticulata. Transverse wood section illustrating ring poros sI- ihustrating diffuse ty, P.4 120. 22. ( Iti Porosity. \~ 290. oe is pallida, Transverse wood section 23 and 24, Celtis pacifico. ‘al wood sections illu strating oe snantiopkylla. Radi al — ed variable. a ee idl pitting, x 660. , Lozan a > ‘ druse crystal in a ray a s 330: 2 e same. The s ystal as S in Ficure 25, under polarized light, x 330. 550 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 the pits having a circular to elongate outline; some species also had pit arrangements and shapes not easily described; FicurEs 23 and 24, used to illustrate these different patterns, for which the term “variable’’ will be used here and wherever such arrangements occur in other genera. Celtis cinnamomea, C. laevigata var. smallii, and C. lindheimeri with some opposite arrangement in addition to alternate; Celtis pacifica and C. ritiensis with mostly transitional to opposite arrangements; total range in pit diameters for vessel-ray parenchyma pitting and vessel-axial paren- chyma pitting is 1.5 to 7 p. Axial parenchyma is vasicentric, aliform, and confluent; most species also with terminal parenchyma; Celtis pumila with diffuse in addition to vasicentric parenchyma; Celtis iguanaea, C. occidentalis, C. pacifica, C. paniculata, C. paniculata var, multiseriata, C. pumila, C. swartzii, C. tri- nervia, and C. ugandensis with banded parenchyma in addition to the paratracheal types. Spiral thickenings occur on small vessel elements of all ring porous species; prismatic crystals found in ray and axial parenchyma of all species except C. adolphi-friderici and C. ugandensis. Pteroceltis (1, 2). Wood diffuse porous; pore distribution 75 percent solitary, 24 percent radial multiples of 2 to 5 cells and 1 percent clusters of 3 to 4 cells; average pore diameter is 55 » with a total range of 27 to 90 »; vessel element wall thickness ranging from 2.5 to 5 »; perforations simple; end wall angles ranging from 19 to 90°; intervascular pitting crowded, range of pit diameters from 1.5 to 3 ,; average vessel element length 250 » with a total range of 75 to 390 #; fiber wall thickness rang- ing from 2 to 4.5 ; average fiber length 870 » with a total range of 510 to 1260 p. Uniseriate vascular rays homocellular (Celtis type), biseriate and multi- seriate rays heterocellular; height of unicellular rays 2 to 12 cells and 85 to 390 ; biseriate and multiseriate rays ranging in height from 4 to 41 cells and 75 to 760 »; width varying from 2 to 5 cells and 10 to 45 pw; dark staining deposits and prismatic crystals in some ray cells; cells containing crystals usually enlarged. Axial parenchyma vasicentric and confluent; vessel-ray and_vessel- axial parenchyma alternate, pits ranging in diameter from 1.5 to 3 »; pris- matic crystals in some axial parenchyma cells. Ampelocera (4, 5). Wood diffuse porous; pore distribution is 74 to 90 percent solitary, 9 to 22 percent radial multiples of 2 to 5 cells, and to 5 percent clusters of 3 to 4 cells; average pore diameters ranging from 73 to 124 » with a total range of 30 to 170 p»; vessel wall thickness 3 to 5.5 p; perforations simple; end wall angles ranging from 30 to 90°; intervascular pit diameters from 6 to 10.5 #4; average vessel element length ranging from 222 to 277 » with a total range from 148 to 370 p. Some fibers septate; fiber wall thickness ranging from 3 to 4.5 p; 3 1 average fiber i Be ccs length ranging from 1088 to 1310 » with a total range from Uniseri iseri : oe ee an and multiseriate rays are homocellular, comprised 13 cel a a beers rays not abundant, ranging in height Ba: frorn 5 to 50 to 225 yw; biseriate and multiseriate rays ranging 4 10 to-40 .: cells and 80 to 810 »; width ranging from 2 to 5 Be parench ae naires crystals occurring in some ray cells. lin 4 ibaa ea in A. cubensis, diffuse in A. edentula, and en, and aaah ula and A. glabra; all species have vasicentric, ali- iting is alt parenchyma ; vessel-ray and vessel-axial parenchyma ernate, the pit diameters ranging from 2.5 to 5 p. Zelkov 4 Kent Bs ood ring porous; pore distribution is 26 to 38 per- radial multiples , percent clusters of 3 to 25 cells, and 0 to 8 percent ition of the or . 3 cells; average pore diameters in the first formed ey 15 to 5s g nde ring ranging from 73 to 150 p» with a total range growth ring Ste average pore diameter for the last formed part of the vessel wall aaa aie 23 to 41 » with a total range of 15 to 55 p; bar), simple wi me ae SOEES from 1.5 to 5 »; perforations scalariform (1 15 to 90°: inter vestigial bars, and simple; end wall angles varying from sel element len Apacer pit diameters ranging from 5 to 9 pw; average Ves- ee found & ac 185 to 207 p with a total range from 111 to 259 p; in small ail elkova serrata and Z, sinica; spiral thickenings found DS eiorate ane storying of vessels in Z. keakii and Z. serrata. with bordered oy eary elements include vascular tracheids and fibers serrata: a. : Some fiber-tracheids are septate in Z. keakii and Z. fiber-tracheid Seo: wall thickness varying from 2 to 3.5 p. Average 888 to 1998 ,. gth ranges from 1024 to 1687 » with a total range from not abundant; biseriate and bent cells in Z. heakit homocell esas ular uniseriate rays ranging in height from 2 to 11 cells and height from 6 to 60 to 2 sae 90 ant Aspage and multiseriate rays ranging in Prismatic ¢ 10 to 1165 p, in width from 2 to 10 cells and 10 to 115 p; containi rystals and dark contents found in some ray cells; some crystal Ona cells enlarged. 1 . Ditting | alagaiaaaen vasicentric; vessel-ray 4 and dark oe with pit diameters from 7 in Z, aia occurring in some cells; cryst nd _vessel-axial parenchyma es prismatic crystals alliferous strands found Hemi iptelea (1, 2). Wood ring porous, pore distribution 42 percent Solit ’ of aida aia clusters of 3 to 15 cells, and 8 percent ra 0» with a S$; average pore diameter for the ear of the rin eranee from 15 to 165 p; average for g is 22 » with a range from 10 to 30 p; Pe 552 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 (1-2 bars), simple with vestigial bars, and simple; end wall angles vary- ing from 14 to 90°; intervascular pits ranging in diameter from 3.5 to 6.5 ; average vessel element length 220 » with a range of 148 to 296 yn; small vessel elements with spiral thickenings; some storying of vessels occurring. Imperforate tracheary elements include vascular tracheids and fibers; some fibers septate; vascular tracheids with spirals; fiber wall thickness ranging from 2 to 5 p; average fiber length 1100 » with a range from 888 to 1332 pu. Uniseriate rays homocellular (Celtis type) with procumbent cells; ranging in height from 3 to 12 cells and 70 to 300 pw; biseriate and multi- seriate rays heterocellular: height ranging from 6 to 45 cells and 95 to 675 », width ranging from 2 to 28 cells and 25 to 280 #3; prismatic crystals and Cark staining materials found in some ray cells. Axial parenchyma vasicentric and terminal; vessel-ray and vessel- axial parenchyma p‘tting alternate with pit diameters from 1.5 to 2.5 p. Trema (9, 11). Wood diffuse porous; pore distribution 53 to 88 per- cent solitary, 6 to 42 percent radial multiples of 2 to 4 cells, and 0 to 6 percent clusters of 3 to 6 cells; pore diameter averaging from 80 to 117 pu with a total range of 15 to 165 »; perforations simple; end wall angles varying from 10 to 90°; intervascular pitting crowded in Trema guineénse, T. micrantha, and T. vieillardii ; pit diameters ranging from 2.5 to 6.5 ; thickness ranging from 1 to 5 #; average fiber lengths ranging from 777 to 1295 ye with a total range from 518 to 1702 pL. Uniseriate rays homocellular (Celtis type), ranging in height from 2 to 21 cells and 55 to 785 ; biseriate and multiseriate rays heterocellular, ranging in height from 4 to 7 cells and 100 to 1875 » and in width from 2 to7 cells and 10 to 60 »; dark staining material occurring in many ray cells; prismatic crystals found in ray cells of Trema vieillardii. _Ax al parenchyma vasicentric; vessel-ray and vessel-axial parenchyma pitting ‘“‘variable” to alternate; pit diameters ranging from 2.5 to 6.5 p. solitary, 20 percent radia of 3 cells: average pore di . ; pw; average vessel ele- ment length is 381 » with a range from 240 to 450 yw: fiber thickness vary- average fiber length 890 » with a range from 456 to Uniseriate rays homocel ee lular, compri ‘ ing in height from 2 to 1 3 prised of procumbent cells and rang cells and 75 to 390 »; biseriate and multiseriate 1971] SWEITZER, ULMACEAE 553 rays heterocellular, ranging in height from 5 to 38 cells and 160 to 860 By in width from 2 to 4 cells and 20 to 75 p; dark staining deposits and prismatic crystals found in some ray cells. Axial parenchyma vasicentric and terminal; dark staining deposits and prismatic crystals found in some cells; vessel-ray and vessel-axial parenchyma pitting alternate with a pit diameter range from 5 to 7. pe Aphananthe (2). Wood diffuse porous; pore distribution 66 percent solitary, 30 percent radial multiples of 2 to 4 cells, and 4 percent clusters of 3 to 5 cells; average pore diameter 85 » with a total range of 50 to 150 »; vessel wall thickness ranging from 2.5 to 5.5 1; perforations sim- ple; end wall angles of vessel elements ranging from 15 to 90°; inter- vascular pit diameters ranging from 3.5 to 6.5 m; average vessel element length 290 » with a total range from 222 to 518 p; tyloses found in ves- sels of A. philippinensis; fiber wall thickness ranging from 2.5 to 4.5 4; average fiber length 1190 » with a total range from 1036 to 1443 p. Uniseriate rays homocellular (Celtis type) comprised of procumbent cells ranging in height from 3 to 10 cells and 65 to 235 »; biseriate and multiseriate rays heterocellular, ranging in height from 3 to 43 cells and 65 to 525 », in width 2 to 5 cells and 10 to 40 »; dark staining deposits found in some ray cells. Axial parenchyma is vasicentric, confluent, and terminal; A phananthe philippinensis also exhibits some banding; vessel-ray and vessel-axial parenchyma pitting alternate to “variable,” pit diameters ranging from 2.5 to 6.5 ph. Mirandaceltis (1). Wood diffuse porous; pore distribution 73 percent solitary, 20 percent radial multiples of 2 to 3 cells, and 7 percent — of 3 to 6 cells; average pore diameter 57 » with a range from 35 to Bs vessel wall thickness 2.5 to 5 »; perforations simple; end wall angles rang ing from 35 to 90°: intervascular pitting crowded with a pit gree from 3 to 4 p; average vessel element length 307 », ranging from 148 to 407 »; tyloses present in vessels. : Fiber ail iickacs 2 to 3 p; average fiber length 828 p with a range from 740 to 962 p. iad nging in height from 2 Uniseriate rays homocellular (Celtis type), rang a heterocellular, in width from 2 to tals and dark to 15 cells and 65 to 275 »; biseriate and multiseriat ranging in height from 6 to 30 cells and 100 to 510 p, 4 cells and 25 to 40 uw; some ray cells with prismatic crys Staining deposits. Axial parenchyma is_vasicentric, confluent, and terminal; prisma , 5 istribution from 51 Gironniera (5, 9). Wood diffuse porous; pore sage to 7 cells, and to 92 percent solitary, 6 to 46 percent radial multiples 554 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 0 to 6 percent clusters of 3 to 8 cells; pore diameters ranging from 88 to 186 » with a total range from 40 to 255 ; vessel wall thickness ranging from 2.5 to 5 »; perforations simple; end wall angles varying from 30 to 90°; intervascular pit diameters ranging from 2.5 to 12.5 w; average vessel element lengths ranging from 333 to 407 pw with a total range of 148 to 1147 ,; tyloses present in Gironniera celtidifolia. Some fibers septate in Gironniera celtidifolia; fiber wall thickness vary- ing from 1.5 to 6.5 w; average fiber length ranging from 1188 to 1795 p with a total range from 814 to 2590 pe Uniseriate rays homocellular (Celtis type); uniseriate rays of Giron- niera cuspidata and G. nitida ranging in height from 2 to 13 cells and 60 to 275 yw; biseriate and multiseriate rays ranging in height from 5 to 62 cells and 85 to 1115 » and in width from 2 to 5 cells and 15 to 45 p5 uniseriate rays of G. celtidifolia, G. nervosa, and G. subaequalis ranging in height from 2 to 18 cells and 90 to 1335 »; biseriate and multiseriate rays ranging in height from 5 to 89 cells and 230 to 2475 », in width from 2 to 14 cells and 20 to 205 -; prismatic crystals present in ray cells of G. cuspidata and G. nitida. Axial parenchyma vasicentric in G. nervosa; vasicentric, confluent, and banded in Gironniera celtidifolia, G. cuspidata, and G. nitida; vasicentric, aliform, and confluent in G. subaequalis; prismatic crystals present in some cells of G. cuspidata and G. nitida; dark deposits present in some a and axial parenchyma cells: vessel-ray and vessel-axial parenchyma ae are “variable” to alternate; pit diameters ranging from 2.5 to p. solitary, 34 percent ‘ ; perforations simple; vessel end wall angles varying from 15 to 90°; intervascular pit diameters from 3.5 to 6.5 w; average vessel element length 450 » with a range of 296 to 666 »; tyloses present in C. aristata rays heterocellular, ranging in height from 7 to 96 cells and 65 to 1920 y, in width from 2 to 5 cells and 20 to 35 p; dark Axial parenchyma vasicentric, confluent, and banded; vessel-ray and vessel-axial parenchyma Pitting “variable? anes variable” meters from 2.5 to 6.5 p. 8 lable” to alternate; pit dia Lozanella (1). Wood diffuse porous; pore distribution 72 percent solitary, 28 Percent radial multiples of 2 to 4 cells; average pore diameter ; vessel wall thickness 2 to 3 p; perfora- m 36 to 90°; intervascular pit diameters 1971] SWEITZER, ULMACEAE 555 from 7 to 9 »; average vessel length 468 y», ranging from 300 to 630 p; tyloses present. Some fibers septate; wall thickness from 3.5 to 5 ; average fiber length 930 » with a range from 630 to 1350 p. Uniseriate rays homocellular (Celtis type) comprised of procumbent cells, range in height from 2 to 10 cells and 125 to 510 »; biseriate and multiseriate rays heterocellular, ranging in height from 7 to 91 cells and 200 to 2500 », in width from 2 to 7 cells and 40 to 85 »; druses and dark staining contents present in some ray cells (FicurEs 25, 26). Axial parenchyma vasicentric, confluent, and banded; some cells with druses; vessel-ray and vessel-axial parenchyma pitting “variable” to alter- nate; pit diameters from 3.5 to 6.5 DISCUSSION The validity of using secondary xylem characters for systematic purposes has been established through the studies of Bailey and Tupper (1918), Frost (1930a, 1930b, 1931), Chalk and Chattaway (1935), Kribs (1935, 1937), Barghoorn (1940, 1941a, 1941b), and many others. Fol- lowing these basic investigations, application of concepts has provided valuable information on the relationships of many disputed taxa, ¢.g., Tippo (1938), Heimsch (1942), Stern (1952), and Carlquist (1960). The techniques of systematic wood anatomy and its application to prob- lems in plant phylogeny was summarized by Tippo (1946). Bailey, mn 1957, wrote: “The chief trends of evolutionary specialization in the cam- bium and xylem of dicotyledons are now so reliably established (except in the case of certain patterns of wood parenchyma distribution) that they can be utilized to advantage in studying salient problems of phylogeny and classification.” Use of leaf anatomical characters for systematic problems was summarized by Carlquist (1961), who noted, “The leaf Is pies anatomically the most varied organ of angiosperms, and its ae variations often occur closely with generic and specific — occasionally ser: familial lines.” : nN any comparative systematic study using only anatomical rina would invite misconceptions. Anatomical characters are most pate when employed in conjunction with data from other fields of botanical en iain, Important correlative data might come from fields of ye beige ah cytotaxonomy, paleobotany, embryology, morphogenesis, phytogeost ; A etc. It is important to consider and evaluate studies from these miter = tive fields and to apply pertinent findings to those from plant pred Consequently, a complex of characters must be examined in an ane to avoid mistaking convergent evolution for valid genotypic sim! which have resulted from derivation of taxa from common ancestry. ANATOMY The anatomical observations made in thi ment with previous works. Scalariform perforations s study are in basic agree- are reported here 556 JOURNAL OF THE ARNOLD ARBORETUM [vou. 52 for the first time in the secondary xylem of some species of Celtis, Hemip- telea, Planera, Ulmus, and Zelkova. The Ulmaceae as a family are diverse anatomically. However, there are many anatomical characters of leaf and wood which pervade the entire group. Among the common features in leaves are the anomocytic stomatal apparatus, bulbous-based unicellular trichomes, similar multicellular hairs, parenchymatous bundle sheaths and extensions, presence of silica, calci- um carbonate, and mucilage, and all venation patterns basically pinnate. Common features of the xylem include relatively short vessel elements (less than 350 y» long); alternate intervascular pitting; mostly sim- ple perforations; relatively short fibers (less than 900 p» long); simple to slightly bordered fiber pits; uniseriate, biseriate, and multiseriate vascular rays; and some homocellular vascular rays in all genera. The two tribes can be separated to a greater or lesser extent on the fol- lowing anatomical grounds: ULMEAE CELTEAE “Simply” craspedodromous Brochidodromous venation pattern venation pattern Pegged cystoliths absent Pegged cystoliths present Woop Rays homocellular only Rays homocellular and heterocellular Vessel-ray and vessel- Vessel-ray and vessel-axial axial parenchyma pitting parenchyma pitting ‘‘variable” to alternate alternate The ulmaceous genera are more or less distinct anatomically. Celtis can be divided into two groups on the basis of ring versus diffuse porosity. The ring porous species are similar to Ulmus (which is all ring porous) in pore distribution, presence of spirals in small vessel elements, and types 0 perforation plates (scalariform, simple with vestigial bars, and simple). However, Celtis is easily distinguishable by the presence of heterocellular biseriate and multiseriate rays in the secondary xylem, by the brochi- romous venation pattern in the leaf. | ince of pegged cystoliths, , loose spongy mesophyll, and la On TOPTELEA iti Urmus (§ Madocarpus), MIcRropTELEA, and TRICHOPTELFA. Anatomical evidence neither Sustains nor rejects any of these groupings. 1971] SWEITZER, ULMACEAE 557 In the other Ulmeae, Phyllostylon and Holoptelea both have cystoliths in the leaves as well as many other leaf characters in common. However, they can be separated on the basis of wood anatomy: Holoptelea has only few uniseriate rays, biseriate and multiseriate rays are storied, and ray cells lack crystals; Phyllostylon has many uniseriate rays, rays are not storied, and prismatic crystals occur in ray cells. Planera is distinguished from Holoptelea and Phyllostylon by the “simply” craspedodromous venation pattern in the leaf, whereas, Holoptelea and Phyllostylon show a brochidodromous pattern. In wood anatomy Planera is distinguished from Holoptelea and Phyllostylon by the pore distribution: Planera has mostly solitary pores and clusters, Holoptelea and Phyllostylon have mostly solitary pores and radial multiples. Planera has many xylem anatomical features in common with Ul/mus. However, Planera has spiral thickenings in all vessel elements and Ulmus only has spirals in small vessel elements. An apparent anomaly occurs among the studied specimens of Planera aquatica, two of these are diffuse porous and the third is ring porous. ‘taspedodromous leaf venation pattern, presence © tei and storying of vessels. Trema has the acrodromous leaf eaeari a uneven leaf epidermis, and septate fibers in the alee tees ines Aphananthe both have a “simply” craspedodromous leat vena 558 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 but in other anatomical features are quite similar to Celtis. Lozanella is similar to Celtis in all respects except that it has druses in the vascular ray cells. It was not possible in either Celteae or Ulmeae to separate most species within a genus on the basis of wood and leaf anatomy. PROBLEM GENERA The question whether Chaetoptelea should be a separate genus or part of Ulmus cannot be answered on the basis of a few characters. Dr. Leo Hickey of the Smithsonian Institution (personal communication) is able to distinguish fossil Chaetoptelea leaves from fossil Ulmus leaves on the basis of venation. Standley (1922) stated that “All the species of the latter genus [Ulmus], however, have a broadly winged fruit, while in Chaetoptelea there is no vestige of a wing, and this probably is a sufficient basis for the maintenance of Liebmann’s genus.” Leaf anatomy of Chaetoptelea mexicana is quite similar to Ulmus. In wood anatomy both genera are ring porous, but two distinct differences in the secondary xylem occur: Chaetoptelea has mostly solitary and radial pore multiples, whereas, all Ulmus species have mostly solitary and clustered pores; Chaetoptelea lacks spiral thickenings in vessel elements while all Ulmus species have spiral wall thickenings in small vessel elements, Metcalfe and Chalk (1950) state, “The most valuable taxonomic vessel characters are the distribution and pattern of the pores as seen in transverse sections.” Thus, it would appear that Chaetoptelea as a separate genus may be sup- ported by wood anatomy. Sharp (1958) created a separate genus, Mirandaceltis, for the single species C eltis monoica Hemsl. on the following basis: “This genus differs from Celtis in the venation of the leaves which resembles that of Ulmus Resets yo drupaceous fruit is also larger and less spherical than that usu- ally le si on Celtis - - -” Some differences in leaf anatomy were noted eg some unicellular hairs are barbed in Mirandaceltis and not in ye ef ) cystoliths are of the pegless variety in Mirandaceltis, whereas, all Species, except C. schippii, have pegged cystoliths. It must be The genus Gironniera contains two tostigma Planch.) 2 with G. celtidifolia * According to ICBN, Art. 2 subgenus containing the type sp subgenera, GrRONNIERA (Nema- G. nervosa, and G. subaequalis and 2, 1966, the name Gironniera must be used for the ecies, 1971] SWEITZER, ULMACEAE 559 GALUMPITA Blume with G. cuspidata, G. nitida, and G. reticulata (Ben- tham & Hooker, 1862; Engler & Prantl, 1893). Anatomically the two sub- genera are most readily separated by the much larger secondary xylem rays of the former group. Subg. Gatumpita is reported to have pos- sible affinities with Chaetachme (Record, as a footnote to Janssonius, 1932). On the basis of wood and leaf anatomy there are no features which would negate a relationship between subg. GaLumpita and Chaetachme. On the basis of leaf anatomy it is unlikely that subg. G1- RONNIERA is related to Trema, however, since there are several differences: overall venation pattern is acrodromous in Trema and brochidodromous in subg. GIRONNIERA, pegged cystoliths are lacking in subg. GIRONNIERA but present in Trema, and the occurrence of a multiple epidermis in Trema, absent in Gironniera subg. GIRONNIERA. Trema and_ subg. GIRONNIERA are similar in overall wood anatomy. Subgenus Grronniera is similar to Parasponia in wood anatomy, but differs in leaf anatomy, Parasponia having neither pegged cystoliths nor barbed hairs, subg. GrronNrerA having both. Further detailed study of the wood and leaf anatomy is necessary to determine species variation before this problem can be solved. FAMILY RELATIONSHIPS The question has been raised whether Ulmaceae should be one family with two subgroups or two separate families. Link (1831) proposed two families, Ulmaceae and Celtaceae (as Celtidaceae). Grudzinskaya (1967) supported this contention and suggested that the Celteae (as Celtideae) has more characters in common with the Moraceae than with the Ulan but not enough to incorporate it into the Moraceae. Groisinestse ur- ther suggests transferring Hemiptelea and Zelkova from Celteae to U meae. It should be mentioned that her conclusions were drawn without prior study of the secondary xylem of the Ulmaceae. The sesliea ms split this family in two fails 10 aig ae genera with transitional characters in each tribe, e.g-, Am? ere a lostylon, Zelkova, Planera, and Hemiptelea. Baillon (1880) state a a the “Celtideae, on the one hand, could not be disjoined 4209 Ess et the fruit of Planera being intermediate between the drupes of Ce c m8 the samarae of Ulmus.” Lindley (1853) stressed that he Cee nT deae) and Ulmeae should be considered “mere forms meee she: been Maintenance of two subgroups within the Ulmaceae has, accordingly, followed by most taxonomists. : illus- Wood toa leaf anatomy shed additional light on this ese ah — trating further transitional characters. P hyllostylon a A r of the rays, a character of the Ulmeae, and pegged cystoliths, eae with Celteae; Zelkova has many secondary xylem characters ae biseriate oth groups: the venation pattern is like the Ulmeae an ba oS and multiseriate wood rays are heterocellular (a apse eae in many and homocellular (as in Ulmeae). Hemiptelea 1s like 26 560 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 respects and in addition it has pegged cystoliths (as in Celteae). Am- pelocera, which is placed in, and is like the Celteae in most leaf and wood characteristics, has homocellular biseriate and multiseriate rays (a charac- ter of Ulmeae). Even when the secondary rays are considered in other Celteae, the homocellular type is not totally absent since it was noted that homocellular types of transitional nature are found. In the Anatomy portion of the Discussion of this paper is a complete listing of anatomical features held in common by Ulmeae and Celteae. As far as relating the Ulmaceae to the Moraceae is concerned, most present day systematists acknowledge the naturalness of the Urticales, among them Cronquist ( 1968), Hutchinson (1969), Taktahjan (1969). A notable exception is Thorne (1963) who thinks that the Urticales in the sense of the taxonomists noted above, particularly of Cronquist, com- prise convergent groups. Nair and Sharma (1965), on the basis of pol- len Studies, state that the Urticales are a homogeneous group. If Urticales 1s composed of genetically closely related elements, as is conceded by most authorities today, it would not be at all irregular for genera in re- lated families to possess certain common characteristics. That is exactly what occurs between the Moraceae and some members of the Celteae. From the point of view of many morphological and anatomical features, Ulmaceae should be considered a natural taxon with two, more or less distinct subgroups having transitional forms. When a comparison is made between the relatively small family, Ul- maceae, and the relatively large family, Moraceae (53 genera and 1,400 species) , Statistical probability of similarities through parallel evolution is very high. The dangers of comparing a small taxon with a large fam- ily have been pointed out by Stern and Brizicky (1958). GENERAL ANATOMICAL CONSIDERATIONS The use of the terms ] phyll and the possible each genus. Consequently, this variation Imus and Celtis is probably genetic and 1971] SWEITZER, ULMACEAE 561 not related to environment. However, the use of this character for taxo- nomic purposes is too uncertain at present and further intensive study is required to establish its value. Cystoliths were classified as pegged and unpegged. It is obvious from the descriptions of the two that they differ in composition and structure. Pegged cystoliths have a cellulose peg upon which calcium carbonate and silica are deposited, whereas unpegged cystoliths are deposited without a peg and appear to be composed mainly of silica. The taxonomic signifi- cance of the presence and distribution of pegged and unpegged cystoliths, as well as of mineral placks, crystals, and mucilage is not clear. All mem- bers of Ulmaceae secrete and deposit silica, calcium carbonate, calcium oxalate, and mucilage in varying degrees. The major differences are in the position and form of the deposits. Further study of the secretions and deposits in the leaf and wood of the Ulmaceae is needed. The over- all venation pattern and presence or absence of bundle sheaths on terminal tracheary elements seems to be constant within a genus; however, the ef- fects of environment and species variation are not known. The use of such characters in taxonomic treatments may prove valuable later. One difficulty arising from the investigation of any family which has both ring and diffuse porous species growing in both temperate and tropical regions, is the need for caution in interpreting comparative data between the two groups. Gilbert (1940) has suggested that the ring po- rous condition is a response to a temperate (seasonal) environment and that resulting evolution has given impetus to development of other so- called “advanced” characters, such as spirals in vessel elements. Others have addressed themselves to this problem and have arrived at basically the same conclusion (Tippo, 1938; Metcalfe & Chalk, 1950). The sees study tends to support this contention. Although ring porous species © Celtis, Zelkova, Hemiptelea, and Ulmus possess certain associated aoe ters, such as spirals in the vessel elements and relatively shorter fi and vessels which appear to make them advanced anatomically ones have scalariform perforation plates and simple plates with vestigia ai both considered to be primitive features. Diffuse porous species of sien ceae and other woody Urticales invariably have simple perforation pia ce exclusively. Thus, ring porous species, although possessing some associa advanced characters also have retained certain primitive ones ; species of tropical regions are Js while species of temperate _ Celtis species from ‘eg . This conditi pecies of other ulmaceous genera ae os Palace or ioated species of Quercus i ter- (Fagaceae), which also have ring and diffuse porous members, to dete 562 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 mine if there is any correlation in this genus between porosity and sea- sonality due to geographical location. Wood specimens of Quercus from the Smithsonian Institution wood collection and Q. virginiana from Key Largo, Florida, collected by William L. Stern were checked for porosity by means of a hand magnifier. As can be seen from Taste 5 the temperate TABLE 5. Wood porosity of Quercus species arranged in a north-south geographical axis R-RING POROUS SPECIES D-DIFFUSE POROUS COLLECTION SITE QUERCUS coccinea Muenchh. R Brooklyn, N.Y. alba L. aera R Philadelphia, Pa. laurifolia Michx. R Dismal Swamp, Va phellos L. R Dismal Swamp, Va falcata Michx. R Gainesville, Fla. virginiana Mill. R Del Rio, Texas virginiana D Key Largo, Fla. candicans Née R-D Huehuetenango, Guatemala (6,000 ft.) oleoides Mull. D Augustine, British Honduras ; i, copeyensis Mull. D Cartago, Costa Rica , (8,000 ft.) baruensis Mull. R-D Cerro de Punta, Panama (6,000 ft.) zone species are all ring porous. However, the tropical and subtropical _ are diffuse porous or weakly ring porous. Quercus virginiana, like “era aquatica, exhibits both ring and diffuse wood porosity. The only sai Cit difference in the expression of the porosity appears to depend on t e degree of seasonality. The abruptness of seasonal change may be more significant than latitudinal position. Quercus species from the latitu- such. aa Play ae . an porous wood occurs in certain species, ‘ca and Quercus virgini i ion de- ana, but its expression de pends on a seasonal environ ent - ; The solution of the problem 1971] SWEITZER, ULMACEAE 563 within species, even though certain characters appear to be correlated with environment (ring porosity and spirals in temperate regions), other characters remain unaffected and constant and tend to show the relation- ship between the tropical and temperate elements of these taxa. Celtis exemplifies this since it includes species showing ring and diffuse porous wood but having many other anatomical characters in common, such as brochidodromous venation pattern, tight spongy mesophyll cystoliths in leaves, and heterocellular with transitions to homocellulat rays in the secondary xylem. The Ulmaceae as a whole is a good example of this structural-environmental phenomenon in that, even though there are species of temperate and tropical zones (ring porous and diffuse porous) many anatomical characters are held in common by both groups. Con- sequently, ring porosity and spirals in vessels should not be considered as an advanced anatomical condition in comparisons between taxa of Saat and tropical regions. Instead, ring porosity and spirals should e considered as specializations in Ulmaceae which are correlated with a seasonal temperate environment. ORDINAL PLACEMENT AND PHYLOGENETIC CONSIDERATIONS all agree that the Urticales ed to and including the o disagree is Thorne who Cronquist, Hutchinson, and Taktahjan should be assigned to a larger category relat Hamamelidales. The only modern phylogenist t places the Urticales in a group having as its progenitor the Malvales. On purely morphological grounds, it is difficult to justify placing the Urticales in a group which has anything but reduced anemophilous flowers; the Malvales complex does not have reduced flowers; the Hamamelidales complex does. TABLE 6 summarizes some external morphological features of both groups. Looking at wood anatomy to help resolve this problem, “Thus far, the evolutionary we consider the two largest orders on either side of the Urticales (Thorne, lidales and Juglandales) Malvales and Rhamnales; Cronquist, Hamame we find that Urticales are more closely allied in wood anatomy to the lat- Rhamnaceae have sim- 564 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 TABLE 6, Comparison of morphological features of the largest orders on either side of Urticales Cronguist (1968) HAMAMELIDALES URTICALES JUGLANDALES Growth form trees-shrubs trees, trees shrubs, herbs Leaf arrangement alternate alternate alternate (rarely opposite) Calyx parts 4-5 2-8 2-6 Corolla parts 4-5 or absent absent absent Stamens 2-8 228 3-40 Ca 2 1-(2) 2-3 Ovary position inferior-superior superior superior-inferior Ovules 1 1 2(1 Ovule position pendulous ect- erect pendulous THORNE (1968) MALVALES URTICALES RHAMNALES Growth form trees-shrubs trees- trees, shrubs shrubs, herbs Leaf arrangement alternate alternate alternate-opposite Calyx par —5 2-8 _ Corolla parts 3-5 absent 4-5 or absent Stamens numerous 2-8 4-5 arpels os 2-10+- 1(2) 2-4 Ovary position superior superior superior a —man 1 1-2 Ovule position pendulous erect- rect- pendulous pendulous The current anatomical study supports this view. » IM respect to all angiosperms, should be considered a alized to specialized in wood anatomy because of the mostly simple perforation plates; short vessel elements; fibers short to medium in length; at least some vessel-end-wall angles of 90°; alternate intervascular pitting; : moderately speci 1971] SWEITZER, ULMACEAE 565 SUMMARY AND CONCLUSIONS The Ulmaceae is considered to be a natural taxon because of the many leaf and wood characters held in common by its constituent elements. These include the anomocytic stomatal apparatus; bulbous-based uni- cellular trichomes; similar multicellular hairs; parenchymatous bundle sheaths; presence of silica, calcium carbonate, and mucilage; pinnate venation patterns; relatively short vessel elements, alternate intervascular pitting; mostly simple perforations; relatively short fibers; vascular rays uniseriate, biseriate, and multiseriate, at least some of which are homo- cellular; and fibers with simple to slightly bordered pits. The tribes of Hutchinson (1967) are more or less distinct anatomically. This study supports the primitive position of Ulmaceae within the Urti- cales. Wood anatomy negates the placement of the Urticales near the Mal- vales and Rhamnales by Thorne (1968) and supports Cronquist (1968) and others in placing the order near the Hamamelidales and Juglandales. In relation to other angiosperm families, Ulmaceae is considered moderate- ly advanced to advanced because of its many specialized wood features, such as simple perforations; short vessel elements; short fibers; vessel- element end-walls of 90°; alternate intervascular pitting; homogeneous Tays or a tendency to produce homogeneous rays, storied vessels, fibers, and rays; and apotracheal banded, aliform, and confluent axial parenchy- ma Gironniera may be divided into two subgenera characters. Subg. GrroNnNrERA with large vascular rays and subg. Ga- LUMPITA with smaller vascular rays. Wood and leaf anatomy are in- Sufficient bases on which to separate the subg. GALUMPITA from Chae- tachme. Subg. GironNiERA is reportedly related to Trema or Parasponta. Trema differs in having pegged cystoliths, a multiple epidermis, and acro- dromous venation. Parasponia differs in leaf anatomy from Gironniera in having pegged cystoliths and lacking barbed unicellular hairs. Trema, Parasponia, and Gironniera subg. GIRONNIERA are similar in wood anatomy. Further detailed study of wood and leaf anatomy is warranted to determine variation within species before these Gironniera problems can be resolved. he use of “loose” versus “compact” ‘ad and distribution of cystoliths, cryst eserve further study since they might be use a studies. Ulmus ‘eg Pa Ne pect mesophyll while Celtis exhibits a compact” spongy mesophyll. However, other genera 1n the Ulmaceae are less consistent. Here too, intensive study on variation within species ah effects of environment on production and distribution of mineral deposits and mucilage is needed. ae Ring pious woods and associated characters, such as spirals in vessel spongy mesophyll and the occur- als, mineral placks, and mucilage fully employed in taxonomic 566 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 elements, are considered to be structural modifications correlated with seasonal environmental conditions, since these conditions occur only in temperate zone species of Ulmus, Celtis, Zelkova, and Hemiptelea. Planera aquatica can be either ring or diffuse porous. Planera, when ring porous, has smaller pores in the early wood than do other ring porous ulmaceous genera; and differs from these genera also in the occurrence of spirals in all vessel elements. The less pronounced ring porosity is probably related to the less marked seasonal nature of the environment of Planera aquatica in the southeastern United States. Investigations of porosity in several species of Quercus from temperate and tropical regions tend to support this conclusion. Scalariform perforations are reported for the first time in the secondary xylem of species of Celtis, Hemiptelea, Zelkova, Planera, and Ulmus. ACKNOWLEDGEMENTS My sincerest thanks and appreciation to Dr. William L. Stern whose guidance, patience, encouragement, and constructive criticism made this study possible. I am grateful to the directors and curators of collections at the Smithsonian Institution, Yale University, and Harvard University for dried wood specimens and to the curators of the United States National Herbarium for dried leaf material. My special thanks to Dr. Richard Howard, Dr. Donald Wyman, and Mr. Henry Goodell of the Arnold Arboretum for their generous help in obtaining living specimens from their collections. I would also like to acknowledge the cooperation of Dr. A. J. Sharp In providing me with dried material of Mirandaceltis. For his trans- lations from the Russian language, I gratefully thank Mr. Donald Bissing. My sincere appreciation to Dr. Oswald Tippo for suggesting this problem and for providing me with his slides of elm wood. LITERATURE CITED as ae Silica in timbers. Techn. Pap. Div. Pl. Industr. C.S.1.R.O. a =, H. N. 1961. Studies in paleobotany. John Wiley & Sons, Inc. New ork. ph, H. J. 1959. Leaf clearings. Turtox News 37: 192-194. Ment, © 1937. Ulmaceae. [Plagioceltis, gen, nov. Mildbraed ex Baehni] in: ae - Macsripe, Flora of Peru. Field Mus. Nat. Hist., Bot. Ser. 13: 272, - 1941b. The ontogenetic development and phylogenetic specialization of rays in the xylem of dicotyledons. III. The elimination of rays. Bull. Tor- tey Club 68: 317-325 Battey, I. W. 1957. The study of the phyl Arb. 38: 243-254, & W. Tupper. 1918. Size variations in tracheary cells. I. A com- parison between the secondary xylem of vascular cryptogams, gymnosperms and angiosperms, Proc. Am. Acad. 54: 149-204. potentialities and limitations of wood anatomy in the ogeny and classification of angiosperms. Jour. Arnold 1971] SWEITZER, ULMACEAE 567 BatLey, L. H. 1949. Manual of cultivated plants. ed. 2. Macmillan Co., New York. BaILton, H. 1880. The natural history of plants (Transl. by M. M. Harros). Vol. 6. L. Reeve. London BARGHOORN, E. S. 1940. The ontogenetic development and phylogenetic spe- cialization of rays in the xylem of dicotyledons. I. The primitive ray struc- ture. Am. Jour. Bot. 27: 918-928. . 1941a. The ontogenetic development and phylogenetic specialization of rays in the xylem of ponents 18h Se of the multiseriate and uniseriate rays. Am. Jour. Bot. 28: 273-282. BECHTEL, A. R. 1921. The floral pp: of a Urticales. Am. Jour. Bot. 8: 386-410. BENTHAM, G., & J. D. Hooker. 1862. Genera plantarum. Vol. I. L. Reeve. London. BERNARD, C. so Sur la distribution geographique des Ulmacées. Bull. Herb. Boiss, 6 Si. Berry, E. L. ees Tree ancestors. Willliams & Wilkins. Baltimore. BESSEY, C. E. 1915. The Serb tics taxonomy of flowering plants. Ann. Mis- souri Bot. Gard. 2: —164. Capoor, S. W. 1937. liek a history of Holoptelea integrifolia Planch. Beih. Bot. Centralbl. 57A: 233-249. i Carouist, S. 1960. Wood anatomy of Astereae (Compositae). Trop. Woods . 1961, Comparative plant anatomy. Holt, Rinehart & Winston. New York. : : ey Cuatk, L., & M. M, Carraway. 1935. Factors affecting dimensional variations of vessel members. Trop. Woods 41: 17-3 Cuevatier, A. 1942. Les ormes de France. Rev. Int, Bot. Appl. Agric. Trop. 22: 254-256, 429-459. : ComMITTEE on NOMENCLA TURE, INTERNATIONAL emer e has poo OMISTs. 1957. International glossary of terms used in wood anatomy Woods 107: 1-36. five Cox, M. J. 1941. The comparative anatomy of the wean xylem of fv American species of Celtis. A idl. Nat. 25: 348-35 Jants. Cronguisr, A. 1968. The evolution and classification of flowering p n. Dartincton, C. D., & A. P. Wyzie. 1955. Chromosome atlas of flowering plants. George ‘Allen & Unwin. London hn Wiley Davis, G. L. 1966, Systematic embryology ‘of the angiosperms. John & Sons, Inc. New York. live — of the ta De Bary, A. 1884. Comparative anatomy of the vege ei phanerogams and ferns (Transl. by F. O. BowER a endon Press. Oxford. Descu, H. E. 1948. The boxwoods. Wood 13(5): pe cies and relationships Dick1son, W. C., & E. M. Swetrzer. 1970. The re ogy of Barbeya oleoides, Am. Jour. Bot. 57: 468 -pyrénées. F. Savy. ae aris. ' Tournay. Pari —. B. C. 1829. Analyse des familles des Plant rain Co,, New at). Clar- ork, icklun sgeschichte Eoxuanor, T. 1937. Untersuchungen iiber Morphologie, — 568 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 und systematische Bedeutung des pseudomonomeren Gynoeceums. Nova |. eto ellie construirt und erldutert. Wilhelm Ra oe of Ulmaceae in the southeastern United States. ee wiles plantarum. Part IV. Fr. Beck. Vindobonae. ENGLER, ‘., & K. PrantL. 1893. Die natiirlichen Pflanzenfamilien. Vol. 3. Born- Erptman, G. 1966. Pollen morphology and plant taxonomy. Hafner Publish- ing Co. New York. ; ETTINGSHAUSEN, C. von. 1861. Die blatt-skelete der Dikotyledonen, Kais. On. Hof- und Staatsdruckerei. Wien. Licht Evert, R. F., C. M. Tucker, J. D. Davis, & B. P. DESHPANDE. 1969. e err microscope investigation of sieve-element ontogeny and structure in americana. Am. Jour. Bot. 56: 999-1017. ; rhage € sol ——— & B. P. DesHpanne, 1969. Electron microscope investigation a jars element ontogeny and structure in Ulmus americana. Protoplasma 68: 432 Frost, F. H. 1930a, Specialization in secondary xylem of dicotyledons. I. Origin of vessel. Bot. Gaz. 89 67-94, : 1 ee ~———. _1930b. Specialization in secondary xylem of dicotyledons. II. lution of end wall of vessel segment. Jbid. 90: 198-212. ena - 1931. Specialization in secondary xylem of dicotyledons, III. Specializ tion of lateral wall of vessel segment. Jbid. 91: 88-06 GarpDINER, R, C. 1965. Studies in the of western hemisphere Trema. M Storrs. [Unpublished]. Gitzert, S, G. 1940. Evolution leaf anatomy and geographic ere aster’s Thesis. University of Connecticut. ary significance of ring porosity in woody angio- sperms. Bot. Gaz. 102: 105-120 as GrubzINsKaya, I. A. 1967. The Ulmaceae and reasons for distinguishing ee Celtidoideae as 4 Separate family Celtidaceae Link. Bot. Zurn. 52(13): 1723-1749. (In Russian), GRUMBLEs, T. 11961: “The comparative anatomy of the secondary xylem of four oriental species of Celtis. Lloydia 4: 145-152. GUNDERSEN, A. 1950, Families of dicotyledons. Chronica Botanica. Waltham. ERLANDT, G, 1928. Physiological plant anatomy (Transl. by M. Drum- mond). Macmillan. London Hauer, H. 1912. L’origine et le systéme phylétique des angiospermes ns Posés a l’aide de leur arbre généalogique. Arch. Néerl. Sci. Exact. Nat. III. 1: 146-234. HErMscy, C. 1942, Comparative anatomy of the secondary xylem of ae Gruinales” and “Terebinthales” of Wettstein with reference to taxonom illoa —198, . 955. Henley’s twentieth century book of formulas, processes d trade secrets, Henley. New Vork Houser » C. 1899, Phylogénie des U] CH 967 macées, Rev. Gén. Bot. 11: 106-119. Th € genera of flowering plants. Vol. 2. Clarendon - 1969. Evolution and phylogeny of flowering plants. Academic Press. London, 1971] SWEITZER, ULMACEAE 569 JANSSONIUS, ~e gs 1932. Note on the wood of the genus Gironniera. Trop. Woods 2' in Jones, L. P., ee A. HANDRECK. 1967. Silica in soils, plants and animals. Jn: Advances in agronomy. ed. A. G. Norman. Vol. 19. Academic Press. New Pisa. H. 1902. A course in botany and pharmacognosy, Edward Stern Co. Philadelphia. Kriss, D. A. 1935. Salient lines of structural specialization in the wood rays of Hatyledona. Bot. Gaz. 96: 547-556. . 1937. Salient lines of structural Ne icmtpy in the wood parenchyma of dicotyledons. Bull. Torrey Club 64: 187. Nyouw, J., & F. A. STAFLEv. 1964. er Spree Regnum Vegetabile oaa2 51, Lannine, F. C. B., W. X. Ponnarya, & C. F. CrumprTon. 1958. The chemical nature of silica i in plants. Pl. Physiol. 33: 339-3 Lawrence, G. H. M. 1951. Taxonomy of vascular plants. Macmillan Co. New Yo rk. LEBRETON, P. 1965. Eléments de chimiotaxinomie botanique. 2. Cas de flavo- noides chez des Urticales; conclusions générales. Bull. Soc. Bot. Fr. 111: 80- 93. Leroy, J. F. 1946. Le genre Aphananthe (Ulmacées) révision a. : distribution géographique des espéces. Bull. Mus. Hist. Nat. Paris I. basque L. 1878. Contributions to the fossil flora of the western territories Par The Tertiary flora. Jn: Report of the United States Geological og of the Territories 7: 187-191. Hal . 1883. Contributions to the fossil flora of the resi territories. Par 3, The i gee and Tertiary floras. bid. 8: 160-1 ae a Lewin, J., & B F. ReErmMANN. 1969. Silicon and ee, growth. Jn: Rev. Pl. fe ed. L. Macutis. 20: 289-304. LInper, D. H. 1931. Some diseases of elm trees. Missouri Bot. Gard. Bull. 19: 61-69. bury & Evans. London. barsten und am haufig- Linptey, J. 1853. The vegetable kingdom. ed. 3. Brad Link, D. H. F. 1831, Handbuch zur Erkennung der nutz Team). sten_ prieenumen en Gewiachse. Haude & Spenerschen (S. J. Joseep Ber ae Lire, E. L. 1953. Check list of native and naturalized trees of the Unite States. Agriculture Handbook 41. Washington, Manesuwart, P, 1950. An introduction to the embryology of angiosperms. McGraw-Hill. New York. Midl. Martin, A.C, 1946, The comparative internal morphology of seeds. Am. Mi at 2: ae characters of selected families. Iowa St. Coll. eid oe toe Nideices In: A. Engler’s Syllabus der Pflanzenfamilien. pecs C R. Ne Ce opte Anatomy of the dicotyledons. 2 vols. aoe oe Pa genes do lenho de Ampelocera glabra. Arq. Inst. Biol. Veg. 3: 211-215. a ee saniiiene. Mireez, C. F, B. 1815. Eléments de physiologi Magimel. Paris. 570 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Narr, P. K. K., & M. Sarma. 1965. Pollen morphological studies in Indian Urticales. Bot. Not. 118(2): 177-186. Nicuots, L. P. 1970. Tree diseases description and control. Pennsylvania State — Extens. Serv. Special Circ. 85: 1-29. PansHIy, A. J., C. D. ZeEuw, & H. P. Brown. 1964. Textbook of wood tech- nology. ed. 2. McGraw-Hill. New York. Prreyre, N. 1961. Contribution a |’étude morphologique, oe et phys- iologique des cystoliths. Rev. Cytol. Biol. Veg. 23: 93-32 Prrone, P. P., B. O. Dopce, & H. W. RicKETT. 1960. ae and pests of ornamental —. Ronald Press, New York. PLANCHON, J. E. 1873. Ulmaceae. In: A. P. DE CANDOLLE, Prodromus syste- matis naturalis regni vegetabilis. Vol. 17. Truettel & Wiirtz, Paris. PRIEMER, F, 1893. Die anatomischen Verhiiltnisse der Laubblatter der Ulma- ceen (einschl. — und die Beziehung zu ihrer Systematik. Bot. Jahresb. 17: 4 REcorp, S. J. ek “occurence of calcium carbonate deposits in woods. Trop. Woods 12: 22-2 G. A. Gana, 1925. Boxwoods. Yale Univ. School of Forestry. Bull. 14. New Have & R. W. Hess. Timbers of the New World. Yale Univ. Press, New Haven. & C. D. MELL. 1924. Timbers of tropical America. Yale Univ. Press, aven na ew REHDER, A. 1940, Manual of cultivated trees and shrubs hardy in North Ampaice: ed. 2. Macmillan w York. RENDLE, A. B. 1938. The HaUGcatidese of flowering plants. 2 vols. University Press, Cambridge. RosInson, T. 1963. The organic constituents of higher plants. Burgess Co. SARABHAI, R. P., & A. K, SAXENA. 1961. Petiolar anatomy of some Urticales. Bull. Bot. Six Univ. Sangar 13: 102-107. SATAKE, Y. 1931. Systematic and anatomical studies on some Ee cange ane i: Sax, K. 1933, — numbers in Ulmus and related genera. Jour. Arnold Arb. 14: 8 Scorr, F. M., on Cystoliths oa plasmodesmata in Beloperone, Ficus, and Boekmeria. Bot. Gaz. 107: —378. Suarp, A. J. 1958. Memeacdice a new genus from Mexico. Bol. Soc. Bot. México 23: 38-42. SHATTUCK, C. H. 1905. 40: 209-223. Stnnortt, E. W., & I. W. Barzey. 1 1914. Investigations on the phylogeny of the angiosperms, 3. Nodal anatomy and the morphology of stipules. Am. Jour. Bot. 1: 441-4 —453. SMITH, F., & R. Mon NTGOMERY. 1959. The chemistry of plant gums and mu- cilages, Reinhold Co. New York. SOLEREDER, H. 1908. Systematic anatomy of the dicotyledons (Transl. L. A. Bonus & F. E. Frirscu; revised, D. H. Scorr). 2 vols. Clarendon Press. A morphological study of Ulmus americana. Bot. Gaz. Soma, S. 1965, Developmental studies on the orientation and the dorsiventrality 1971] SWEITZER, ULMACEAE 571 of the leaf of Zelkova serrata Makino. Jour. Fac. Sci. Univ. Tokyo. Sect. 3. Bot. 9: 1-17 STANDLEY, P. oF 1922. Trees and shrubs of Mexico. Contr. U.S, Natl. Herb. STERN, W. L. 1952. The me anatomy of ae Bf and the phylog- eny of the Julianiaceae. Am. Jour. Bot. 39: 220- 2 Index xylariorum. Reetin Vegetabile Pe 1-36. . K. Brizicky. 1958. The peered anatomy and taxonomy of ec Bull. Torrey Club 85: 111- TAKHTAJAN, A. 1969. Flowering plants: sage and dispersal (Transl. C. JEr- THoRNE, R. F. 1963. Some eps and guiding principles of angiosperm phylogeny. Am. Nat. 97: ; =~. 1968. Synopsis of a spate phylogenetic classification of the flow- ering plants. Aliso 6: 57— Tiero, O. 1938. Comparative anatomy of the Moraceae and their presumed allies. Bot. Gaz. 100: 1-99. : os ~The role of wood anatomy in phylogeny. Am. Midl. Nat. 36: 362-3 User, F. ‘i 1940. Microincineration and ash analysis. Bot. Rev. 6: 204-226. WALKER, J. C. 1969. Plant pathology. ed. 3. McGraw-Hill Co. New York. WALKER, R. I. 1938. Macrosporogenesis and embryo development in Ulmus fulva. Bot. Gaz. 99: 592-598. Ul aaa age Megasporogenesis and development of megagametophyte in U/- :; 37: 47-52. Wertsrem, i is Haas der systematischen Botanik. ed. 4. Franz ”"\'« State College of Agric. 160 pp. JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 572 Jeosesepryy “WILY rt d s $ n a V $ “YUL auyovjany) suolsol jeodos} pur ayeioduiay, a S-p S-b a S-a Vv SL "1 91189 IBISEBEpPLY ‘eIpeaysny vAvpeyy ‘eIsy Wioyseq a Sb S—b a S Vv SL ‘Your ayzuounydp SAPUT SIA4 ‘woueWy jyestdory a ol-+ o-> n‘a a Vv Sa YPszZ}O/ yy vsa20j/adup aVILTID equ suaydsiwap{ Wey WON S 8-+ 8-+ a $ V L “I Suu) SaIBIS payug wiayseayynog S Sr Sh a Ss Vv i ‘PUD “Af Pseuntg vunuasay pure '} YOOH 3 “EyUeg xe [Izb1g] 0} SOIpUT ISAAA S$ 8-s 8-S a s V L vursuede) wopcysoncyd BLY ‘vIsy S 8+ 8-+ a a v L ‘Your Dajazgojoyy vureueg 0} OorKapy S S ¢ a S Vv - “Wigar'y Dajazgojzany) aVaNTQ eqUy val Zouy 2 qu ¢ & =z 2 4 sana uoHNgUystp [eotydvssoax : vs ; J : : S : . Fy 5 34 ee eee SRG F 4 3. EO gi 8 e, v an ' & F "28 5 2 o > : E : g Z yng SIIMOL A uBIvy qusWAsUEy 1qeH yeay » SLIMWUITL) JO BIVUID “] ATIVE, 573 vueind) ysnig 98T9ZMS 1) a DInjuapa [izeig TLZ78EMA qH “WYNY yynyuapa eqn) s09Z9T MA 4 68 Joyong “"qasliry sisuaqna vaaooTaa Ny NISMO, NOLLOATIOD q MAHONOA AOLIATIOD sa1oadS aoo Wworavaday » PPUTWEXS deIIEUIT-) JO susUNIZeds poom pag ‘ep ATAV ‘pzs ‘ezg “d uo | atavy, 03 saduaiajai ag, SWEITZER, ULMACEAE 1971] uedef pure evury) 0} UBIUBLIDPIY W19}seq 1a s- s+ a $ Vv L yreds v204jaZ suoal [eotdoayqns pur [ertdosy, ad s- S—-+ an $s Vv SL “NOT OMi94] BISY }SP9Y}ION Ta s g a Ss V 1 “WIXEIY 5177990497 tuysrg xa (9961 ‘SHIEA) Nae a S-b t a Vv V s “AQP 8191220130] g Spur[st 9yoed BOUINE) MON ‘eALTeHY da s g q s Vv SL bry vmuodsosg BLAWY [BIQUaD ‘OoIKa ad s g n s Vv SL cdieys sisjarvpuns py MI 0} OOKa ad s s a S) O L UBUIUIIIL) O//IUDSOT uly) N t S—$ q s V i qourjd vajasdima yy Spuryst 2yloeg “eury) wiayynos ‘eALpeyy ‘Uo]AaD a $ S a S-a V SL yoIpney vsaruuosy JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 574 OITXITT O6TPEMSQ Sa "US SUTT[OYY ®W SUIBSIAA DIDIIAAD] OdTXOTA vSSZZ7MS/) A ‘W'S UOSTIM B a[BON ‘SyLO PILEM 27052007 eAuay TOC TMA ‘yuleg Dupissndsy Bulyy STOSZMA, ‘ySU “Wy, Ue IeYeN Ssisuarps0y vuleued LI8PSMA sn OSs sae R AM ‘sdaquieyy ‘u13}S$ DavUurnst eureueg S699T MA ‘Bares (‘bovf) papunnst ny S8787MA A LE9r YWS "DV suo tuaqavy sourddryiyg S8PZMA LID]. DIDAAAS-OJDUAAI eAet PO6ZZMA, 24 zSrs jaddoy ap ur, [PU] vamoumpuuto euly) SS66TMA, ‘ySUT “Wy, UT “U's [[LIOT auIn[g vupasung [eeas] SPp77MS/) fnH cs youpnyy “YI Sij04qsnp sourddyryd 619SMA ‘Lal MYOftaagsp 38280) PICD OLSETMA OHA S6SZ Oust [Sug potsaprsf-1ygjopv sILTag saurddytyg OZEZIMA ‘your sisuauddyryd uedef CT SOMA ‘your véadsp aHINVNVHdY Brqui0jo) 98P8ZMA A $99 puesng ‘TPUPIS 2217704 [Ize1g 906EE46A, ‘M's ZIURTTA “‘wyYyNYy vaqnys NIOIQ , NOILOATIOD 4 AAHINOA AOLIATIO|) Salads adooM Woaravadayy (panuijuo2) paururexa aeareul[ ys) Jo suautiseds poom prauq “ep ATAVL - é SWEITZER, ULMACEAE 1971] o3u0+ urs yseod ploy Ra10y seanpuopy ysnug seinpuoy ysoug $a}e}S paytuy sourddiiyg spurjsy [eajsny Tee Y eouIny) MON $9ze1S pou v67P7MA OCELEMA ST8SPIMA L6OVS7MA L6¢sIMSn OSLZEMSN) 7L9IOPMA f86E7MSN) LIZEMY “Alu TRAQUaD JeUONEN A < = Z fy A g¢ BuvyD Iq BurA "U's YARIS ‘u's IY gz oyuowLY “u's ypoyury Sth SAYS 6L9¢ uyof 4S 06 J9TIMJ9q ‘u's S[ATURCIL ‘US I[BON ‘Bug tuayuag ‘yON pyofimuay TED PyP9 “‘YOur| uzqapas [Buy wuxnvdos “sldq Sisuauts udqiyas ‘Tpurys uddiyas Your, sisuazis ‘Your|d suaosasis 1g YY pywiasyynwu “IBA DyvnaiUDpd “‘your[d CIPUa) Dypjnaiupg ‘youryg pyofi40] (9sog sisuatd dississtust *D Se) DyDIILADI JOURNAL OF THE ARNOLD ARBORETUM [vor. 52 576 yse0) PI9D COTETMA OHA bOSZ BuIA “AQPTNAL Stpunad VaTalLdo1oH] sesueyIy ZOZSEMA V ‘W'S SSOP ‘Your (auepy) wpranp VaTaLdI Wap] eauInr) MON SPIOZMA, | ‘yourlg syonbangns eouInt) MONT OL6ETMA DpiyimM eIUOpaTea MON SOSPIMA, yjueg ppiyim SaIpuy yseqy yong 6SETIMA DSOAdaU SaIpuy ysey yong SSEZIMA ‘POUL vsoasau eAet TOLTEMA €8Z sntuossuv[ Dywpidsna SaIpuy seq yong LSEZIMA zINy Dyvpidsna Wy LO9LZ7MA AN 98h WIS “DV oyof1piyfa9 PIsoUela TL O987Z7MA AN 6PT asnoyi9}EM "YoIphey) pyofiptyya9 VaadINNOUIS) vAUdy SOSTMA “GIR M Dd4vI0401UL vOLyV SLE6TMA a “W'S UEqeTTYIS ‘YOUR 07D)S14D AIWHOVLAVHD pyeurayens) S88ZZMYV V OFOz yIINYAS [‘yourfg (ugar) vuvoxam SNU]/) SC] DUDIIXAIML euleueg Z8SSIMSQ 7. go SlaquIey) ® W1a}S “UIgar'y vuvoixau VITILIOLAVHD NIOIYQ , NOILOATIOD q MAHINOA AOLIATION) saloadS doo WoaArdvaIayy (panuijuo?) paururexs seaovur]) Jo suaum2ads poom paliq “eh ATAV 577 SWEITZER, ULMACEAE 1971] BIyeUINS BARR] Pury Pury epHold epHoly epHoly PIqUIO]O. eAef OOIXIJ, euueg eIpuy osu0d TS8L£Z7M4S) OPZ9OTMA CvIZEMA 6TOZEMA 68ZTZ7MS) 606E7MA 669TEMA O8S9TMSN ESTOPMA O6Eh7MS 1) HOIW HSI4a “AIU() UPUOFT “AIU() UBUOFT SVTA AAV On avn ud ‘ms BIDIOg IS yeUIYRY 60bZ UvAIG ‘u's UPA BuIyD ‘ws UBM SUTYD “W'S WOSTELAA, ‘ws puesndg “u's sniuossuef 622189 divys 96 SIaquIey.) Y UII} ‘u's STUOCT “‘Pourlg Dyofysnsup aUIN|_ sisuauroquy VW L MOULD IDI “ULNIXB YY MOULD IDY SILTIOONAL pouonbo poyonbp ‘yourry “yy { va10nbp VaANVId ‘ded sisuayispag NOTALSOTTIAHG “DIP, Dsoytasndg VINOdSVaAVqg divys (‘[SWax{) DI10UN0M SILTAOVANVAIP uoWOW ® ary (‘wisg-"uu0gd) Dydydouunua VITANVZO'T ‘yourld vyoftasagut SIpuvas JOURNAL OF THE ARNOLD ARBORETUM [vou. 52 578 puvpsuq 16961T MA. OHA (stagsadupo 7 Se) vAqnis edoiny P906MYV (Staqsaquno * 2) SB) DAQn{s edoing S906MY (‘J Stagsadupa *f Se) Daqnjs $9721$ peruy) COVESMA sn paqnis $972} Poul) 98STZMA “spnyy Paqnjs $978}§ psuy) TZ7Z7EMy V ‘VINNY oYofissv49 [eeasy] 8Lb7cMSN {nH cg youpny,) OTTALSIA, S¥a2sauna vuepul L808TMSN avn DUDILAIULD $9}e}S Poul) STZEMV V T vuvotsawmp Hnosstyl SE697TMA A DIDID $978}$ Poul) Oz77fMV V DDD S$972}S poruy) pSO64V V "XIJAL DIDI saWty) eurqo LEST7MA “ysuy “Weyy UR SIZ suey, oun] 07D5.u0. PIUOPaed) MON, S86S TMA, “yooR, Pug ~ (your) mpsvpjsata sourddytyd 66SZ€4S0 d19 aunt” (""T) sypzuatso shay epuoly brrorMA A SIT Wig aun ("T) oypunsoeue shoy epuoyy LOF6PMA - L@Z siaquivyy ® usajs aunt g (-I[NYIS ®Y “wa0yY) DUDLYIADULD] Iopensy b9FOZMS 1) x | 986TT SI[OS-eISOIV ‘[purys (Jang) Dusdsadaqut 3s¥0D PIOD 899ZIMA avn 902T aust _ WIT sesuaumns ROY SS6EMA aun 0707084904 NIOIQ , NOILOATIOD q YAHINOA AOLIATION SaIOddS adooM woaravaday (panutju02) pauruiexs svaovw]) Jo suaumsads poom paiiq, “Pp ATAV = / SWEITZER, ULMACEAE 1971] PSOULIO.J $9}8}S paul) sa}e1S paup sour puelod $9}e1S povUuy saje}1S pou, ynots0uU0D purpsuyq O90bZMA 69064V SE7OPMA cTLecMSN) C996MY so90cMsn TOPOEMA ‘JSUT “WIT Ue aT OdAOL Vv WT) OHA ‘WS BYOLALT ‘W'S YPIUIYIS PZT Ay PLPs ALLL Ss “A "W'S YDOITET. I9ZT JasUIAITD ZLZI A9BUaAIT_D 76 PaIyoURy, 10Z yoef PLZI jlasuaAg{a VyARET Uspouacn (vaynf A se) DAQna CXUDTTAL Dayn{ “9 SB) “[YNJL Digna "I Dynund D4a20dd “qst[es D4aI04d soni, 227074 ‘boef nzjofraand Yus0y] Suaqiu yyws “yf 4ofpu aouey Vdsvr0s9vUM JARIY (‘AgNVIL) pypiui90 DItpuni{oOY 1oinq votpunyjoy [voL. 52 JOURNAL OF THE ARNOLD ARBORETUM 580 ay} 18 pasnoy mou ynq ‘A1jsa10g Jo JooyIg AjIs1aAtUy) a “UOTIATJOD [RHOWa prosay samef anureg aq} sv ‘UIsMOISTAA ‘UOsIpYPy ‘AlOJeIOGeT sjNporg jsesOq “S') TBA 24} I paqeooy AprauI0; wINIE[AX ay} 07 MA PUL UINTEqIIY AY} OF SIAJAI x» unsoUDK XY XapUuy Ut (L961) UAIIg Aq pepuatIUIOIAT asOY} MOT[OF SUOTEIAZIGGY , ‘MNAOLADQ4a FY XaPUy UL ($961) NABI Y MnofueyT Aq papusurwodaI asoy} MOTO} SUOHPIAXIAIY « ‘zeg ‘d uo Up ATAV], 0} BDUAIAJOI 99g , reyeN (qunyy) vpp44as uedef eTZez7MSN OAOL d M “UNIXEYL ("qeIS) zyDay via weet VAONTIZ ueysryed $SO6bMA aynqynsuy yoivasay i qsa10,J uvysIyeg “yourld 249% 0ar NIONYQ , NOILOATIOD q WAHONOA AOLOATION SaroddS aooM WOrdvaNaFT (panutqzuo2) paururexa avaovulyy) Jo suaumTIads poom paid: “Vp ATAVL 1971] SWEITZER, ULMACEAE 581 TABLE 4b. Dried leaf specimens of Ulmaceae examined *’” SPECIES COLLECTOR ORIGIN AMPELOCERA cubensis Griseb. Ekman 9470 Cuba edentula Kuhlm. Ducke 1736 Brazil hottlei Standl. Ll. Williams 9497 Mexico - onaagh an aspera Planch. Chiao 2935 China olga Planch. Edano 26773 Luzon CELTIS bungeana Blume Wilson 10459 Korea caudata Hance Callado 9592 Mexico iguanaea ( Jacq.) Sarg. Standley 1500 exico laevigata es, Thieret 16497 Louisiana laevigata va texana (Scheele) Sarg. Lodewyks 382a Missouri occidentalis L. Allard 7931 Virginia ss chdentalis var. crassifolia (Lam.) A. Gray Peck s.n. New York pallida Torr. Killip 32729 Arizona philippinensis Blanco Conteras 23396 Philippines pubescens Granat. Pickel 1177 Brazil pumila Pursh Harbison 5.1. Virginia reticulata Torr Johnson 8580 Arizona reticulata (as ; C. douglasii Planch.) Jahn 8449 Washington schippii Standl. Little 6216 Ecuador sinensis Pers. u sed inosa Ruiz Jorgenson 3853 Paraguay tenuifolia Nutt. Fernald & Long 13326 —_—Virginia tenuifolia Terrell & Barclay 3373 Alabama triflora Lam. Smith & Klein 11162 razil trinervia Lam. Webster 3679 Cuba CHarTacu ME dz 430 Fre aristata Planch. Zenker & Hau Caiacsoutn microcarpa Rendle Piemeisel & Kephart 21 Ken CHAETOPTELEA 5 mexicana Liebm. Standley 20002 Mexico GIRONNIER ss celtidifolia Gaudich. Beer 7082 pncoart ‘ China cuspidata K Lei 831 tons Planch, Low & Motan 94530 Malay Peninsula 582 JOURNAL OF THE ARNOLD ARBORETUM TABLE 4b. Dried leaf specimens of Ulmaceae examined *” (continued) SPECIES COLLECTOR ORIGIN nitida Benth. Taan 1629 Hong Kong parvifolia Planch. King 5369 Malay Peninsula reticulata Thw. Piolane 10768 South Vietnam HEMIP davidii ‘ain Planch. HOLOPTEL miateila Planch. LOZANELLA enantiophylla (Donn.-Sm.) Killip & Morton MIRANDACELT monoica «(Hes ) Sharp monoic PARASPONIA andersonii Mia. PHYLLOSTYLON brasiliensis Cap. rhamnoides Taub. PLANERA aquatica J. F. Gmel. PTEROCELTIS tatarinowii Maxim. TREMA amboinensis Blume canescens Blume lamarckiana (Beur.) Stand. micrantha (L.) Blume orientalis (L.) Blume Utmus alata Michx. americana L. car pinifolia Gleditsch crassifolia Nutt. glabra Huds. glabra (as U. campestris L.) hollandica Duroi Dorsett 564 Choudhury 108 Hinton 14286 Andrle 51 Andrle 103 Baumann 6138 Kuhlmann 17991 Dugand 5511 Palmer 15124 Chiao 3085 Santos 4656 Garnier 688 Small 8078 Dressler 1660 Henry 12189A Moore & Horton 6740 Hoy 521 i ces 10593 Palmer 24050 Norman s.n, nitz s.n. Sonnobree 3523 China India Mexico Mexico Mexico New Caledonia . Brazil Colombia Florida China Philippines Nicaragua Florida Mexico China Louisiana New York Utah Oklahoma Spain Poland Belgium [voL. 52 1971] . SWEITZER, ULMACEAE 583 TABLE 4b. Dried leaf specimens of Ulmaceae examined *"” (continued) SPECIES COLLECTOR ORIGIN laevis Pall. US1750849 Poland minor Mill. David 3A England montana With. Paulsen s.n. Denmark parvifolia Jacq. Howell 26645 California parvifolia Terakawa 136 Japan pumila Purdon 1048 China rubra Muhl. (as U. fulva Michx. Humphrey s.n. Wisconsin rubra (as U. fulva) Sheldon s.n. Minnesota Serotina Sarg. Biltmore Herb. 3634 eorgia thomasii Sarg. Toumey s.n Connecticut ZELKOVA scale carpinifolia (Pall.) Koch Sudworth s.n. District of Columbia serrata Makino Sudworth s.. District of Columbia sinica Schneid. Pei 3077 hina “All dried leaf specimens obtained from United States National Herbarium, Smith- sonian Institution, Washington, D.C. ”See reference to Taste 4b on p. 532. [voL. 52 JOURNAL OF THE ARNOLD ARBORETUM 584 PISy U1ajsoM ‘odoing [e1]Uad pue U1dYWON adoing vISY U19}SIM pur ‘BILIpY Uoyjiou ‘adoing sureJUNOJ AYIOY 0} JSAM pu EPLIOL] 0} pur[punojMaN, epllop,y Ulay}Nos pur skay eplopy eplOp.] UleyINOS pur shay ploy] valoy pue ‘eunyoueyy “euryd uleyION UOIsZaI ULIURIIITIPIP] LIUIOJI[VD UJaYINOS puv UOJZUTYSe AA 0} SPXAT, PUL OPP1OTO.) PIULOJI[VD UIIYINOS puv UOJUTYSeAA 0} SPXAT, PUL OPLIO]O) yeiQ pur ‘opesojo_ ‘sesuey 0} 3saM ‘“plopJ 0} Aassof MON $9}e}S peu sesuvy pue ‘eueqry ‘eulpored YON 0} YINos “eqoyUPPY 0} saqand) eULqey PUL PIBZI0I4) 0} assauUaT, puP PUI[OILY) YON epluoy,y pue sexay, 0} SIoul[] pue eueRIpuy UayINOS uedef pur va1i0y Raloy pur ‘euNyoueyA, “eUIYD Wisy}AOU pue [e1}U9_) evILiV O88LTV essLtV ‘W'S ‘TIZIMS ‘AYVIA 199Z U19}S ‘AYVIN 0992 U191S ‘AUVN ‘spny piqnia ‘pysy (YUIO|) DSOsaqns “WA DY Ofmid 409 YsHpay vyofmsduvs "TT pupa1sauD sawn, aunt (J) PysupsoM aunt (‘NYS Y‘wWsaoy) vuv1y2sDUd] > JONVY vWwaa], e87SV ‘your, (aoueEy) zpiavp VaATALdI NAL q9S-£6SV "Ue’T 12740fUANOF 6LLIZV (‘your uspjinop “2 se) vypjnayjas V9S-LESV "LIO, DgDININAL d66621V ysing vyrund VrsoslV ‘Slag DIDp40I "VA SYDJUApLIIO ALeogveVv "rT supyuaptrz0 SOIIV “B1vg (a[peog) Mmpous “AVA DIDSINAD] V-bS-Z11V ‘PITEM, Dg0s10an] ceSslV ‘zploy stsuaossal eS-Ep9V auintg puvasung SILTI) wAHONOA saloadS woravadayy qx POUTUTEXA aeIIvUTT/) JO SUdUATDIads poom pur Jer] PIAS. “Op ATAVL SWEITZER, ULMACEAE 585 1971] €00ZZ VINIOUIA “ATVONVNNY ANATIO) ALINOWWOd VINIOYIA NYAHLYON SOILVNAHLVI GNV SHONAIDS TVANLVN £0 NOISIAICG, “(€S61) aIWT pue ‘(6r6t) Aopleg “W “TJ ‘(OP6T) Jopyay ‘(EL8T) UOYuL[g WoIJ Udye} a1aM saduLY , ‘zes ‘d uo dp ATAV], 0} aUAIaJaI vag , *paqsi[ SI 9WIeU S,10799T[09 dy} d194YM ‘sIIYINOA WINLIeGIay AYVW YIM asoy} 3dooxa ‘s}JasnyoessePy ‘UOJsog ‘WINJa1OqIy PlOUly 9Y} 3B pazda][o a19M suawTads [TY , BYyseIqaN 0} Jsam ‘aassouuay, 0} Jaqand deqen() 0} yJ410U puP “e}OxRG YON ‘eyseiqaNn ‘ewoye yO ‘sexa], 0} Sam ‘saze1S pa}Uy Usla}sey uejsayin yp, uejsayIn Ty, pur ‘eulyd usayiIOU “eLIaqIS ula]seq adoing ulayjnos pue ulajsam ‘purpsuyq eISY U1a}saM ‘adoiny adoingq ASZ6LTV V996T7V 6£6STV 9f6STV 8£-b-06V OT6LTV TZ6LTV "‘BIVS UspMOoYy “XYOI, DAQna ‘AWT Das0qg4v “eA Dpuind “I ojeund “qsl[eS vsa204g T?d S#0a0g THI 92¢punj1 04 $2 586 JOURNAL OF THE ARNOLD ARBORETUM [ VOL 16. THE ECOLOGY OF AN ELFIN FOREST IN PUERTO a Be THE FLOWERING CYCLE AND AN INTERPRETA OF ITS SEASONALITY Lorin I. NEvLINe, Jr. : igati f the THIS STUDY WAS UNDERTAKEN as a part of the en Le dynamics of a forest located on Pico del Oeste, in the Sierra to stress one aspect forests — the reproductive cycle. } : Jant There are few organized records of flowering times for tropical p raphers or floristic workers, largely on t ; : udy seemed en. to be a unique sie eet He wering and fruiting data. These data, See cage sa of flowering and fruiting, have been useful in help te on a number of questions such as an ahiaks pda e ty, comparative studies with lowla properly the origin of the florist forest. The “flora”’ 40 species of Dicotyledons plus the non-flowering species (see Howard, 1968; Fu * This Study was Partially Supported by a srant from the National Science Foun- dation (GB-3975) to Dr. R. A. Howard, Principal investigator, 1971] NEVLING, ELFIN FOREST, 16 587 species being that few of the species are adapted to the prevailing climatic conditions of the elfin forest. FLOWERING Preliminary transects of the site were made in the fall of 1965 when tentative identifications were obtained from representative collections. These determinations formed the basis of a checklist of the plants of Pico del Oeste and representative specimens were labelled both at canopy and ground level (for details see Howard, 1968). This labelling proved to be of particular importance because of the number and variety of personnel who were to be involved in the maintenance of the records. The “flora” could be learned in a relatively short period and the number of mis- identifications was kept to a minimum. Weekly records of flowering and fruiting were kept from February 1966 to January 1968 inclusive, and the completeness of the record attests to the faithfulness of all persons con- cerne The weekly flowering and fruiting records were combined into a single cumulative monthly record which is presented in TABLE 1. The records are restricted to the plants growing on the site and do not include individuals of the same species growing at lower altitudes or in other geographic locali- ties. The scoring was on the basis of all or none. Quantity was not scored inasmuch as few individual observers had previous experience with the species under study and had no point of reference in this regard. There are a few abnormalities in the record which lead me to the conclusion that the two-year record, while useful, does not give an accurate indication of the sexual reproductive cycles of the plants of Pico del Oeste. Extended observations would tend to show more variability than is now shown for some of the species. The flowering record must be considered n as fruit development became apparent. In many instances, however, mature fruit did not s. There are a few additional mented upon. The s represented by a y early in re study i sg i in flower only trom one Period. Arthrostylidium sarmentosum 1s known : s coeeseauce may uate data failed to flower on the study site, although it flowers conspicuously at lower altitudes. Trichilia pallida and Hillia parasitica were represente oe and may not be regular members of es co Warum and Lobelia portoricensis were.each repr : specimen but as they aes flowered continuously no record eg One species, Vriesea sintenisii, the flowering record as Be nite ane May be exaggerated. This Vriesea is fewsflowered and t a det last for a few days only. In contrast the infignescene a | | [voL. 52 JOURNAL OF THE ARNOLD ARBORETUM 588 rT ae * re aE. x Se so DUDIUOW SIMO * a+ a+ xt at at at ato «xt at at at z WNIDIONLI wnipmolyrDAg AVANVAIHIACO at at at a+ a «+ oo at at at + at UNADIUYUD DiUjoaUuay AVAVAATIONIZ DIvp402 DiuvlDy AVAOVANOISOIG, * + at at at at at ee 2 2k * + * usimajuis Dasats A at + a+ a+ a+ *} * * * DUDIUOLII49G DINDUZNL) aVAOVITAWOUG et hs Macks: ob ae 6 re ae ee ee ea ee ee asuarmiuop wninyjup avdovay * a+ at at x et at * * « i at DUDIUOM DA0ISALG aVNTVd er ae ee ae re a a ES ee at SUuDIAS D1d9]IS at «+ + at at at at ce ee 4+ ve at SISUANDUNK S14DYI0aIA a ‘ ms nee at at at at at at a pt yonjsKjog xaavy AVAOVAAIAD a+ * * at at at st ee at “7 Dyofysnsun auyIDS] * * eho > et ooh > ee ee ee suayog snyjuvuyr] UNS OJUAULADS WUNIPYAJSOLY JAP at smiong uosogoipup AVANIWVA 99q.Ss AON ~pQ dag any Inf unf Aep idy iw qa uef Sa10adS puv ATINVJ 19890 [PP ON%_ JOF sprosa1 Suny pue Fursamoy ATYUOUr sanvpnumMy “| ATAV], 589 NEVLING, ELFIN FOREST, 16 1971] «+ at «+ at at *+ a+ at at at at at a+ «t+ at at at at a+ at «at at at *+ *+ at pyojoundog{v biakay) aVaOVaAH MSMAIUIS DIADAFIAD PW AVANVIAVADONV JL D4I949 DISASDANDG AVAIVNHOO DYyofiaund DISDq]DAdO T, AVAOVALSVTAD usimaquls xa] AVAOVITOAINOY Dpyyog DYyryr14 AVAOVITAY DyDjnNYIDgsS D3I0IO aVaOVaNV’T sisuanbunk vapid usnsy DAltd AVAOVOILAL) pypyjag vidoszay aVaOVaO]L suarsadogaD unwusoxpa AVAOVHLNVAOTHD pyofupupusay Diuosagag DyauisaDua DIWMOsag ag avaovaddig JOURNAL OF THE ARNOLD ARBORETUM [VOL. 52 590 a+ a+ at a+ at at Dyofiuuis4vs syoydosnpw AVANVLOAVS sisuanbunk piuayyo 4 g UstMazuis sisdoyjumqucy Sisuayinbn) Disipap AVAOVNISUAJ DSOUWLAIDA DIMUDULAUAO FT sisuag140j40g XKJVIOUO4) AVAOVOINY usmMaquls DINO PY oycydxyovd mmuorpy Dynjoanof DIu0NW UNUYDPSKUD UWNINDAIA WY unsojnuonbs unimoso2kjv9 AVAIVLVWOLSVTAJY sisuanbuisog viuaing usnsy sayguvsgaxpoy AVIV LAAT DApunrap vimosag AVIOVINODAG DuDlyIDgastAés DISH]) avaddILLay AON vO das any Inf unf ACI idy Ivy be f uef Sa10adg puv ATINVA (panuijuor) 2189O [PP ON 10J spr0d291 FuNmay pue FuraMoy ATyUOW sane[NUIND “[ ATAV], 16 591 , NEVLING, ELFIN FOREST 1971] “salias sty uy saodud snotaaid ut zayy (G1) BSfMaquES DIMapOoMemDsy SY payealy UVEq sYYy (R96 “89 * BNGZUM) [epunyy (qa) msmazurs sesdoyquniquedd SAIS STYI ut suaded snotaaid ul “uso wmnasd “gq SY payeary uvaq sey (6961 “POF 70S “GIy ploury “inof) Avies) wngnjorIo wnipiuoryoDsg , ‘Buymnay = (4) Ysyose {Buyamoy = (+-) uss snid , at at at at at + pycydxyovd vimvyt pw aVLISOdWOs) at at at at at at at at at at at at s1suant40z4od Dyaqo'T AVAOVINNVAWVS a+ at «+ *+ at at a+ + at a+ tb at sisuadnppons v14joyIKsg + + + * * at at at at at a at DUDLAIZAIG DIAJOYIRS | paiispapdg DUI AVAOVIAN Y at su + + at at at at at at at DUDIMOSUIZADUL DIIGSN AVAOVHLNVOY at at + at re at USIMAJUIS DIABUSA) ot ee Gack et “et ae et at Be ae a ke snnsiquo snjoaqdoyy AVANVIMANSA) - + os * * = + 2 2 + + + ppisi4 DINngagn I AVANVINONDIG =f -f- ale et ppuvngas vaomodg ] AVADVTNATOANOD) at at *t SNYOf1ayDS snyJUDIMAD FT aVAOVATO ea + I: pyjuvaniu sorojqursy AVANVIOTAWAS 592 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 52 and we are inclined to believe that they were incorrectly scored as beeibieee TABLE 1 has several additional shortcomings because it is a cumulative record. In the case of Haenianthus salicifolius (Oleaceae) flowering was not observed until late in 1967. When in flower and fruit this small a is quite spectacular and is unlikely to be missed. It is safe to assume t i it did not flower in 1966. Therefore, the records show when a a flowered but cannot be used to predict that any individual species wil : in flower in a particular month in any given year. A second type ‘ difficulty is based on the all-or-none scoring of flowering and oe good example is supplied by Dilomilis montana (Orchidaceae ) wine flowers for seven months of the year. It does reach a flowering pea a January and February, a fact that is not noted in the cumulative eg (and that could be established only at the end of the observation period). The records which were compiled from Pico del Oeste are by no means comprehensive and in order to determine the relative degree of accuracy additional flowering records for these species were compiled from npn: material from the entire island of Puerto Rico. The additional month y flowering records are as follows: Begonia decandra: February. Calyptranthes Rrugii: April. quamulosum: January, August. Mecranium amygdalinum : February, May, June. Miconia pachyphylla: September. Hornemannia racemosa: January, December. Micropholis garciniifolia: February, April, May, November, December. Ipomoea repanda: June, S be eptember. Hillia parasitica: All months of the year, The cumulative monthly flowering record for the plants on Pico del Oeste is given in F IGURE 1. It is presented against the background of the total cumulative records from all of Puerto Rico. In general, the pattern 1971] NEVLING, ELFIN FOREST, 16 593 No. of Species w oOo T rh oO t 3 J A 5 0 N D Months Ficure 1. Cumulative monthly flowering record of the plants of Pico del Oeste. Solid line record is based on observations and herbarium records from the Pico only. Dotted line record is based on all available data sources of the Pico species from all of Puerto Rico. for both is similar and we believe this indicates that our observations are relatively accurate and continued observations could be expected only to add some months of flowering for individual species. It appears as though the overall pattern would remain static. The most striking individual feature is the flowering peak (45 “naire Which is attained in the month of July. This may provide some practica advice for future collectors in the elfin forest, namely, they may a. Most effectively, in terms of the number of species in flower, in ge 3 equally interesting situation is found in the month of October when Secondary flowering peak of 39 species is found. : he question nr whether pease ahr number plays any role in the length of flowering must be considered since ploidy level has — telated by some with survival under extreme environmental we _— The term survival should imply the ability to reproduce in ieee pa another. Reduction of the available chromosome counts to the hap . ssa (Nevling, 1969) and plotting the potential months in flower tor species with a particular haploid number is not helpful. 594 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 n = 26—7 months n = 15— 6, 12 months 25 —0 14— 11, 11 13 —0 ea, 12—— 3,5, 9).12, 12,12 22 — 12 11 — 21— 10, 10 10 —0 20 — 3, 6, 7, 10 9 —0 19— 11, 12 8 — 10, 10 18 — 8, 12 7— 5,12 17— 9, 11 6 —0 16 —0O 5—9 There appears to be no relationship between flowering duration — absolute chromosome number. The data for ploidy level are too limited to evaluate, as most of the counts reported previously remain all that is available for a genus. By plotting the flowering records (FicurE 2) on the basis of Mono- 40} w o ' Dicots nm o ' No. of Species a F M A J J S 0 N D Months URE 2. Cumulative monthl Cc . Fis y flowering records of the Dicotyledons and Monocotyledons of Pico del Oeste. cotyledons vs. Dicot of July and October is sha se towards the end of the year. Interestingly, the lowest flowering level j i i < trast, follow a strong bimodal flowering curve. The primary peak is In July followed by a precipitous drop in flowering in August. The overall peak in July is due to the 1971] NEVLING, ELFIN FOREST, 16 595 Monocotyledons and Dicotyledons. The secondary peak, in October, which was also reflected in the overall flowering pattern, is due to the Dicotyledons alone as the Monocotyledons reach their low point during the month. Another informative way of presenting the flowering record is by habit (Ficure 3). I have arbitrarily sorted the species into five categories: trees, shrubs, climbers, and herbs with the latter subdivided into terrestrial and epiphytic herbs. In some instances a species is referable to more than 10 0 Trees : Sane Shrubs ea 0 uw pa 3) a o a Y Climbers Re AAT a 0 ° 3 a | MI P Epiphytic Herbs a ———— q Terrestrial Herbs ee, as ee J F M A ee Months . Oeste f ts of Pico del ng records ol plan shrubs; climbers; IGURE 3, C ative monthly floweri . Cumulative n ) preference: trees; F divided according to habit and habitat epiphytic herbs; terrestrial herbs. 596 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 one category, Dilomilis montana is found both epiphytically and terres- trially and consequently is scored in both categories. Some differences are apparent in the flowering behavior of the separate groupings. 48 The record for trees shows a remarkable drop in flowering beginning with the month of January and ending in the month of March; this is followed by a gradual rise with a slight peak in April. There is a decrease in May and a gradual rise to a high in the month of July. Then flowering drops off sharply and gradually goes to a secondary high in the months of October and November. The pattern for the shrub flowering is somewhat more complex, with a sharp decrease in January and February and a rather interesting increase in the month of March followed by decrease in April and a gradual rise to high months for June and July followed by decrease in August and an increase to the month of October. The high months are June, July, and October and in general this pattern follows that for trees. There was some question in the results for trees of the tertiary peak found in April. This tertiary peak also seems to be found in the shrubs but one month earlier, and so may not be an artifact of observation but rather a record of some importance. The pattern for the climbers is not especially revealing because of the small number of species. Otherwise, the high months are April, June, July, October, and November with some depression in the months of May, August, September, and December and the early part of the year. In general it does follow the flowering pattern found in both trees and shrubs. by distinguishing the ep them separately. that of the trees a 1971] NEVLING, ELFIN FOREST, 16 597 invalid and the wasps are regular members of the peak biological assem- blage but they do not appear to have a role in pollination. The failure to identify pollinators was not due to lack of diligence on the part of the observers but to an apparent lack of pollinators. That is not to say that the insect assemblage on the peak is small as there is a considerable array of phytophagous insects present and active. In only one instance was a definite pollinator correlated with a specific plant and that was the case of Prestoea montana, pollinated by a species of bee flies (Bombylius). Although this may have been the most disappointing aspect of the study it may explain, in part, the extraordinarily low percentage of fruit set. Those plants which depend on an outbreeding system would be seriously affected by lack of pollinators. This lack may be due to the great frequency of rainfall and consistently high humidity which dilutes nectar concen- tration and decreases attractiveness. In spite of being unable to identify the pollinators it is possible to speculate with some assurance upon the possible pollinators for most of the plant species. This can be done by considering the pollination syndrome for each individual species. In FIGURE 4, I have attempted to separate the flowering records of those species which 40 Insect/Bird Pollinated 30 u ov 8 a wn i > 20 2 loi ee oe Rd, a eae er ee Months “Chea 4. Cumulative monthly flowering records from Pico del Oeste plotted ean with bird/insect pollination syndrome vs. species with wind pollination rome, tg Projected as plants which have a bird- or insect-pollination syndrome tom those which have a wind-pollination syndrome. Although it is usually “asy to distinguish between bird- and insect-pollinated species this distinc- tion was rather vague in the possible bird-pollinated species so it seemed 598 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 appropriate to present these as a single plot. Janzen ae Se Pres: that the activity of insect pollinators peaks during the an Pel the activity of bird pollinators peaks during the wet season. panini record for the bird-insect pollinated species is very similar 5 i Sedathialls Dicotyledons, with a primary peak in July, a secondary sini d-ncilinated November, and a tertiary peak in the month of April. The eit ni at tee species are at a tremendous disadvantage in the climatic regi eng Pico because of rainfall. With high rainfall distributed pee only is year good opportunities for pollen transport are rare indee ll signifi- direct rainfall, which almost immediately cleans the air of pol “ ee. high cant problem, but also the accompanying long-term “i x dity . the humidity are influential. During such periods of high hum ses individual florets of wind-pollinated species do not have the “ia703 has to become dry and shed pollen in the normal manner. Ogle . mnie reported that petroleum jelly coated slides exposed at El Ver ua aaa Rico, yielded no results.2. We ran a similar Cape ment Maing Bx ollen gravity pollen samplers and found a considerable number of wee es, slid grains, and insect parts but it has not been possible to bring ' IS * a of the study to a conclusion. The actual number of projecte wi pte linated species is small within the total flora and this is not baie arse as the percentage of wind-pollinated species in the tropics is pn Seal in general than that which one encounters in temperate regions. 7. the plot of the flowering record is very reminiscent of that shown i terrestrial herbaceous herbs and in fact the two should correspond : sid f because most of the wind-pollinated species are terrestrial herbs. T ne a gradual rise in flowering from January to April followed by a long ane This plateau is followed by a general decrease u ateau extending through December is reached. visa d of many plants of temperate regions is contro se. Short-day and long-day plants are well know This may be partially because of the lessening we gest and shortest day of the year as one approaches to Rico this difference still amounts to more than two hours over the year. We ar Flowering times of tropical plants often at the onset of the rainy season” or during sarily in terms of months of the year. € are in a rather peculiar situation because * Strictly as an aside, Ogle, in her discuss important but unidentified components were ; lieve this is a lapsus for Rubiaceous type. Further, she indicates that Casuarina was first introduced to the island in 1923, although there is an herbarium specimen, from cultivation in Ponce (Sintenis 4372 (GH)), dated 1836. ‘ e ion of results, mentions that one a pollen grains of a Rutaceous type. J J Months ha 5. Flowering data of Pico del Oeste correlated with climatic data. Percent seas mean minimum relative humidity (in perc ent); otter ete . dn cm. age of hours of 100% humidity; so Hiamon i eee ven n.); “on line = flowering (in number of species) ; within sha oe minimu mean temperature (in C°); shaded area = absolute aes a hor car “ah ama differential (in C° ); wavy line = number of days 600 JOURNAL OF THE ARNOLD ARBORETUM [voL, 52 they are in a tropical area but the microclimate is such that a pede We seasonality does not seem to be well marked (especially as the mont os the lowest rainfall has 26 cm. of rain). The question of identifying physiological trigger of the flowering response in these plants has prov o be a difficult one. The flowering record has been plotted with several waeesigne variables (Ficure 5). All of the environmental data were supplied by a National Center for Atmospheric Research (NCAR) except those for tota rainfall which is based on data generated independently by the various observers from Harvard University. The median air temperature at pe Gl level gradually rises from a low in December, January, and erick Pe a high in September and October. This change is slight, the di * i between the high and the low being only three degrees centigrade. : absolute maximum and minimum temperatures have been plotted as well. The greatest difference between night (low) and day (high) cna a is found in the month of April with a secondary spread found in October, . The least difference of temperature, between day and night, is found in the months of May through August. a The figures for the mean minimum relative humidity vary aye a at little with a high of 99 percent plus and lows of 96 percent in April an 94 percent in November. The mean minimum relative humidity ne ee mirror a plot of the number of days of no rain for each month. The hig : was five days with no rain in April with a secondary high of four days 0 rainfall is extremely vari the months of May, include water deliver There are a number of days without rain ($) and an examination of the rainfall record on a da : as exceedingly heavy. Further, it is difficult to identify a wet season aS separate from a dry season. In order to help explain the April “water” records and to identify seasonality I reinterpreted 1971] NEVLING, ELFIN FOREST, 16 601 the data by subtracting all rainfall generated by major storms. This was accomplished arbitrarily by removing all rainfall of five centimeters or more delivered in a single day. In Ficure 6, the plot for an adjusted Rainfall > o ! w o Lf Months R Ficure 6. Monthly rainfall record of Pico del Oeste taken at canopy level. olid line with diamonds = actual measurements; solid line = rainfall record adjusted by subtracting all major storms (ie., days in which rainfall exceeded cm.). This procedure helps to identify wet/dry seasonality. rainfall record is given and the results are striking. It is possible to identify . Crypto-wet/dry seasonality: the wet season being from May Tough September inclusive, and the time from September to May being Considered the equivalent of a dry season. The wet season in Puerto Rico sfnerally is considered to extend from April through November (see espe- Cally Odum, Drewry, and Kline, 1970). In Ficure 7 I have extracted and replotted the rainfall data given by Odum et al. in their Fig, 49c for 602 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 El Yunque. This figure is based on data from 1958 to 1962 and ae es fairly well with our data of 1966-67 indicating an end of the wet io in October. Janzen (1967) has noted differing lengths of the wet season 80 fi a) 5 5 ra = £40 20 J F M A M J oR S 0 N D Months Figure 7. Monthly mean ra infall at El Yunque, Puerto Rico from 1958-1962. Adapted from Odum et al., 1970. Central America. On the basis of rainfall data from various parts of ae Puerto Rico, I believe that although the total rainfall increases as altitu increases (windward side of the mountains) the length of the wet ek decreases with altitude. The seasonality is masked by high total rainfa but it does exist and is an important environmental factor. All visitors to the elfin forest, particularly those wishing to take photo- graphs, are struck by the relatively low ligh Suggests that over 40 percent of incident ra cover as the Pico seems t the daylight hours. so that available pho by more than 40 per on the basis of energy exchange observations estimates that “it would appear that about half the maximum 1971] NEVLING, ELFIN FOREST, 16 603 Unfortunately, a full year’s data on solar radiation is not available but a high of 322 Langleys per day was recorded for April. A direct correlation between amount of solar radiation and flowering periods was not estab- lished. If we consider various environmental factors which may be important to arrive at a flowering peak in July it would appear that April might be a critical month because of its striking differences from the other months of the year. In addition to depressed rainfall, humidity, and high number of days without rain, it is on an increasing day length, the days increasing at the rate of approximately 60 seconds per day. It is possible that we are dealing with a flower bud initiation period in April so that flowering does not reach its peak until July. If true, this is a three month lag between the time of flower initiation and peak of flowering. The second highest peak of flowering occurs in October and if we assume a three month lag period between the time of flower initiation and peak of flowering, then the month of July should be significant. One is struck by the very high relative humidity, the high number of hours of one hundred percent humidity, the lack of any rain free days, and the very narrow band of temperature difference between night and day in July. In addition, day- length is decreasing at a rate of a little less than 30 seconds per day. For the July and October flowering peaks we are facing a set of contrasting environmental characteristics in April and July. April las typified by increasing day-length, depressed rainfall and humidity, increased are perature maxima, and decreased minimum temperature, while le typified by slightly decreasing day length, high rainfall, high paieeed and little difference between day and night temperature. — ae of the control of the actual flower bud initiation has not been resolve ta this study although the preliminary indications point to day — with the actual flowering modified by seasonality. dtoa While we have stated that plants of the tropics generally erate marked wet/dry seasonality by their flowering cycles we have a a + fact that the plants of the Pico del Oeste do not seem to do Sold haste 1). many of the species have very prolonged flowering periods (se aaa Ficure 8 is an attempt to sort out some possible pseee duration, those species which have flowering periods less than ten mon Je unless that With in addition, less than one break in the flowering cy¢ ich fit t cycle is clearly bimodal. The relatively a rai the crypto qualifications have been plotted in FIcuRE te ae ont . ep ee hs, May through Sepa é en months of the dry # ote wering initiations Which occur in May (2), June (4), and July of : : FE i : : ek ek according to their appearance (4) which accounts I ivides 604 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Dry Season Wet Season Dry Season ed we: oe ——< | as sas eo oo o——_.__...... \ or ee ree o—— eo qo ————— oe o—— a ecen ® eees pore ee —— | a ——— eo scorer SN seca Seale Es a Oe ae oo a F M A M J J A Months i i i i f less cycles of species with flowering periods te) “ie flowering, correlated with wet/dry Torralbasia cuneifolia, Pilea , krugi, Peperomia emarginella, Guzmania berte- roniana, Carex polystachy. a, Isachne angustifolia, Ichnanthus pallens. in only one of the seasons or in results are obtained: FLOWERING PERrop During dry season Ee shah Bik a Initiated in dry season, continuing into season a combination of seasons, the following No. oF SPECIES seis teed Initiated in dry season, continuing through wet season into next dry season. . 2 Initiated in wet season, continuing into ry eeaton cS a, - Af During wet season only 0 There are two species (Micropholis garciniifolia and Mecranium amyg- dalinum) which have bimodal flowering, both flowering once in the dry season with the additional flowering segment initiated in the wet season and continuing into the next dry season. number of species in flower and the number of months in flower, some additional interesting aspects may be seen. 1971] NEVLING, ELFIN FOREST, 16 605 % OF SPECIES IN % OF SPECIES IN No. Montus 1n FLrower No. or SPECIES FLOWER IN JuLY FLower IN OcToBER 12 10 100 10 id 7 100 100 10 2 100 100 9 7 100 85 8 5 100 80 *, 6 100 50 6 5 60 40 5 2 50 100 4 3 66 33 3 2 50 50 Z 1 100 100 1 1 0 0 4 0 0 Of immediate interest is the number of species (10) and percentage of species (18%) in flower throughout the year. The number of species in ower nine or more months is 26 or 47 percent of the flora. Even more striking is the number of species (42) in flower for six or more months. In percentage this is 76 percent of the flora in flower for a prolonged period and suggests some similarity to a weed flora. ; It 1s unlike such a flora, however, because the flowering of the Pico species is rather sparse and the fruit production is low or non-existent (for a discussion of the latter situation in Tabebuia rigida see Nevling, 1969). __ IGURE 9% is a representation of the flowering cycle of P Oe SORE but as a continuous record rather than as an interrupted linear ee In addition to presenting the yearly variation it is helpful also in scione ae the percentage of the flora in flower. The possible flowering potentia vad 55 species is 660 flowering months, the plants of the Pico realized Tes mately 64 percent of this potential. An additional 10 percent, or a . : of 74 percent, is obtained if one adds the flowering 2iratan i fatin Puerto Rico (see Figure 1). Thus, regardless of the method of calculating flowering time, the amount of flowering on the Pico 1s high COMPARATIVE STUDIES OF FLOWERING CYCLES The limited records of flowering times of tropical plants pagans the opening lines of this article are no more acutely felt than sib fat of a ing to compare the flowering periodicity eee wai: a rds for an different area, I have not been able to locate flowering aera can be additional elfin forest in the New World so that no compari d in th ted in the *The graphs in this paper, with the exception of FIcuRe 9, ee a ak traditional manner. This linear technique permits easy er cember records are activity but has a serious deficiency in that the January te fed as an attempt to Separated spatially and difficult to correlate. FIGURE 9 P eee of the percentage Overcome this deficiency by presenting a continuous yearly © of the flora in flower 606 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 made. Failing this it was hoped that a comparison of the elfin forest flora of Pico del Oeste with the adjacent rainforest flora at El Verde, Puerto Rico, could be made. Considerable effort was made to study the ecology of JAN sii aay I KY , a Ly OCT ) YY JUL FIGURE 9. Cumulative monthly flowering record of the plants of Pico del Oeste expressed in percentage of speci es in flower. Each circle equals 20%. the forest at El Verde, but unfortunately, the flowering records (Pinto, 1970; Bannister, 1970) are extraordinarily incomplete and are almost use- less for comparative purposes. Therefore it was necessary to direct our com- parisons elsewhere. It is Possible to generate useful flowering data from local floras, particularly county floras, from temperate regions. Such a duces a set of interesting an as data beco had their origins in the tropics, and assuming wet/dry seasonality existed at that time, is there any accommodate ] often in a leafless condition, and the trees in the leafless condition? Or, 1s 1971] NEVLING, ELFIN FOREST, 16 607 No. of Species J eo Wee Aes ieee Months T ‘cal areas. Wavy Ficure 10. Monthly flowering records from pease ae ae (adapted line = flowering of tree species from Puntarenas aie ‘records from Barro from Janzen from data by Allen). Dotted line = ie rae = records from Colorado Island, Panama, from Croat (+10). Das Pico del Oeste, Puerto Rico. 608 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 this resemblance simply the result of a shift to wind pollination? A host of other similar questions of prime biological significance also present themselves. These questions were not explored because it seems to be an unreasonable undertaking with such limited information on tropical floras. We were able to contrast the Pico flowering with data presented by Croat (1969) for a lowland flora of Panama and to a lesser degree with data from Allen (1956) from Costa Rica. The most accessible and usable data : tropical flora with which I am familiar are those recently published by Croat for Barro Colorado Island, Panama. His flowering data were based on herbarium records and direct observations. I have manipulated his data (see Croat’s Ficure 2A) to make them compatible with the Pico data and have plotted the two together (FicurE 10). Barro Colorado is an island about six miles square in the Canal Zone, the topographic relief is approximately four hundred and fifty feet. Croat has suggested that the forest type be referred to as semievergreen, moist tropical forest. He has recorded 1,181 species of phanerogams from the island. The climate 1S described as having a dry season from December to April, the rainfall then increasing gradually until a peak is reached in November. The island receives about 251 cm. of rain per year. As in the case of the Pico flora, Croat found that a surprising number of species have prolonged flowering times. His figure for species in flower for more than nine months of the year is about 41 percent or only slightly less than the 47 percent which we have reported for the Pico. In genera ours show similarity, records show a majo and March, which j ne) whereas the flowering on the t is of interest that our July peak flower during the month of February and also add to the general high 1971] NEVLING, ELFIN FOREST, 16 609 flowering peak during that month. If we question the way in which the composition of the flora of Pico del Oeste differs from that of Barro Colorado Island the difference seems to be attributable to two elements: the Pico flora lacks the annual herbaceous flora, and the massive or large and medium-sized trees. The major difference existing in the flowering between the Pico and Barro Colorado occurs during the dry season and seems to be accounted for in large part by the flowering of the massive, lowland, tropical tree flora. This floristic element often flowers in the dry season in a leafless condition. There is also considerable evidence accu- mulating from the tropical lowlands of Central America indicating that tree-flowering is strongly correlated with the dry season (Janzen, 1967). The primary advantage of flowering during the dry season when leafless, appears to be a reduction in “internal nutrient competition” between the reproductive and the vegetative phases. If this concept is correct, one might then expect no peak flowering to occur on the Pico during the “dry season” as there are no species flowering in the leafless condition. To test this possibility further I have extracted the graphic data presented by Janzen (which he in turn extracted from Allen) for a study of tree species from Puntarenas Province, Costa Rica. These data also are shown in FicureE 10. The data for the tree species show a very high peak of flowering in the early part of the year. The dry season in this area extends from November to April, and is followed by markedly reduced flowering ane April through October, although there is a secondary peak to be we in August (homologous to our July peak?) which may possibly - bn es by those species which are flowering in a bimodal condition. This is tur nd evidence that the Pico flora flowering record is strongly affected by the lack of a major tree flora. Some interesting flowering data a : ae (1966) from the “aie North Forest Reserve, Tanzania. The site : : i meters above sea level and has a mean annual rainfall eee oe has listed the flowering species by the five seasons se : s for com- and according to their habits. His results are tabulated her Iso have been published by Boaler parison, SEASONS Dry Pre-rains Early Rainy eek nal _ June-Aug. Sept.—Oct. Nov. Dec.—Feb. —May Hasir fe z CANopy TrEES 2 6 ust “s UnpDERWoop TrEEs 3 6 ae az a SUFFRUTICES — 17 23 : RU 9 6 3 4 ERBS i 4 4 BuLsIts Be 1 5 TOTAL 3 1 : 70 69 TOTALS 3 40 52 610 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Once again the trees (canopy and underwood) dominate the dry season flowering. The suffrutices then assume a major flowering role from the pre-rain season through the rainy season. The herbaceous plants (including sedges and grasses) begin to come into flower in the early rainy season and peak during it. Other than the shift in calendar months for the dry and wet seasons and the relative numbers of the various habit types of data pattern seems not unlike that given for Pico del Oeste. It is curious to note again that grasses and sedges, presumably wind pollinated, peak during the rainy season when apparently at a serious disadvantage. I believe that the comparative studies furnish evidence that the flowering pattern ex- hibited by the plants of Pico del Oeste is not basically different from that which one finds under other tropical conditions but the differences which appear are primarily the result of the absence of massive and medium-sized trees which typify the lowland rain forest. FRUITING It has been reported in several of the preceding papers of this series that considerable vegetative reproduction is to be found on the Pico. The environmental characteristics of the site are similar to those reproduced 30 Dicots hp oOo No. of Species — Oo — ae Monocots ae een a se Th [Se Lo! UN Months 0 ia 11. Cumulative monthly fruiting records of the species of Pico del este. in mist tents so widely employed by plant propagators today. It is not surprising that almost every broken branch develops a root system and has the possibility of becoming established as an individual plant. In actual total reproduction, asexual reproduction has clearly outstripped sexual reproduction during the course of this study. that one of the major tree species, Tabebuia ri one or a few clones which have, through vegetative reproduction, come almost to dominate the area. Extensive root grafting. among individuals of the same and of different species is so prevalent that it is no longer For example, it is possible gida, is represented by only 1971] NEVLING, ELFIN FOREST, 16 611 possible to isolate distinct trees. The effect of the peculiar environment of the Pico has been to permit or even to promote the amount of vegetative reproduction. The records for fruiting (FicurE 11) are more uneven than those for flowering, for a variety of reasons but mostly involving the observation team. There was a general tendency to score a species in fruit as soon as fruit-set was observable and perhaps Ficure 11 is a better record of that phenomenon than of fruiting. Fruiting implies that some fruits mature but in fact, very few species did carry fruit to maturation. The initial fruit-set was low, perhaps due to the paucity of pollinators, a lack of proper environmental factors or both. In those cases where fruit-set occurred the young fruits were subject to bird and insect predation to the degree that very few species brought fruit to maturation. In this context the existing environmental characteristics of the site must be considered as extremely harsh. Only Eugenia borinquensis, Symplocos micrantha, and Vriesea sin- tenisii seemed to fruit on a regular basis and produce seedlings which were observed. ‘ 5 We are faced with the fact that the flora on Pico del Oeste is not The additional fact that such a large percentage of species have prolonged flowering periods may indicate an adaptation permitting sexual reproduction whenever the prop- er conditions occur, regardless of season. Su have occurred during the more than two years of ou If this conclusion is correct, it may be that pattern can reproduce sexually in the ecologically s the elfin forest. The consequence 0 i i leading to reproduction i lowing of evolutionary processes © Se tramn Tel be the case as the 55 species of the e than one species occurring , and Psychotria (2). The endemic species would have developed. with few exceptions, the elfin forest : : les and the area se-pentiwane or occasionally supplied an ate from individual plants growing at somewhat lower nee and in other basis the propagules could be brought to sai apis a : cases the fruits could be carried by high wine ee aivaiuels of Hillia Storms. That this has happened is proven - a petites there is rare parasitica and Trichilia pallida found on the site. ly of propagules effective sexual reproduction im situ or a continuing SUPP from lower altitudes, the role of vegetative reprodu _In summary, it seems to me, on tive patterns shown by the plants 0 i d reproduc- = of the flowering ane Mr yee “el Oeste, that the flora is similar 612 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 to a tropical flora from lower altitudes minus some elements. The — of the Pico has been derived from only a portion of such a flora. ‘The missing elements are the herbaceous annuals and the massive and — sized trees. A firm indication of wet/dry seasonality remains in spite of t : high total rainfall of the area. The effective species filter of the ier and large trees is to be found in the total climatic conditions present es most especially in the low light intensity. The Pico del Oeste flora ey to be an aggregation of species with a common ability to survive under a low light regime and one which had its origin in the rainforest. BIBLIOGRAPHY ALLEN, P. H. The Rain Forests of Golfo Dulce. Univ. Florida Press, Gainesville. pp. 417. 1956. f : Anperson, E., & L. Husricut. A method for describing and comparing bloom ing-seasons. Bull. Torrey Bot. Club 67: 639-648. 1940. ; : Bacon, P. R. The Ecology of Caroni Swamp, Trinidad. University of the Wes Indies, St. Augustine (?). pp. 68. 1970. T BANNISTER, B. A, Ecological life cycle of Euterpe globosa Gaertn. is: BG Dum. A Tropical Rain Forest. U.S. Atomic Energy Commission. B-299- B-314. 1970. ‘ Baynton, H. W. The ecology of an elfin forest in Puerto Rico, 2. The micro- climate of Pico del Oeste. Jour. Arnold Arb. 49: 419-430. 1968. - The ecology of an elfin forest in Puerto Rico, 3. Hilltop and forest influences on the microclimate of Pico del Oeste. Jbid. 50: 80-92. 1969. CLaus, E. P. A study of the anemophilous plants of Puerto Rico. Bot. Gaz. 109: 249-258. 1948. Coster, C. Lauberneuerun trockenen Monsun- EST. “1923; g und andere periodische Lebenprozesse in oe Gebiet-Ost-Java’s. Ann. Jard. Bot. Buitenzorg. 33: 117 mmunities. Special Publication No. 1. 1-45 + aha Evolutionary patterns and strategies in seed plants. Taxon 970. : The structure of the tropical rain forest and speciation in the humid tropics. Jour. Ecology 54: 1-11. 1966. Foerster, J. W. The ecology of an elfin forest in Puerto Rico, 14. The algae of Pico del Oeste. Jour. Arnold Arb. 52: 86-109. 1971. F FULForp, M., B. CRANDALL, & R. STOTLER. The ecology of an elfin forest in Puerto Rico » 11. The leafy hepaticae of Pico del Oeste. Jour. Arnold Arb. 51: 56-69. 1970. Gates, D. M. The ecology of an elfin forest in Puerto Rico, 4. Transpiration rates and temperatures of leaves in cool humid environment. Jour. Arnold Arb. 50: 93-98, 1969. Git, A. M. The ecology of an elfin forest in Puerto Rico, 6. Aérial roots. Jour. Arnold Arb. 50: 197-209, : 1969 1971] NEVLING, ELFIN FOREST, 16 613 Hotttum, R. E. On periodic leaf-change and flowering of trees in Singapore. Gard. Bull. Straits Settlements 5: 173-206. 1931. Howarp, R. A. The ecology of an elfin forest in Puerto Rico, 1. Introduction and composition studies. Jour. Arnold Arb. 49: 381-418. 1968. . The ecology of an elfin forest in Puerto Rico, 8. Studies of stem growth and form and of leaf structure. /bid. 50: 225-267. 1969. . The “alpine” plants of the Antilles. Biotropica 2: 24-28. 1970. Janzen, D. H. Synchronization of sexual reproduction of trees within the dry season in Central America. Evolution 21: 620-637. 1967. Jones, C. E., Jk. Some evolutionary aspects of a water stress on flowering in the tropics. Turrialba 17: 188-190. 1967. NEvLinG, L. I., Jr. The ecology of an elfin forest in Puerto Rico, 5. Chromosome numbers of some 622 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 3. Style not divided but grooved at apex, hence with one bilobed stigma 1-2 mn weed Biariiewy ites ein eat. ol 4. B. succulenta. 1. Bourreria baccata Raf., Sylva Tellur. 42. 1838. Ficures 1E; 2A. Bourreria arborea foliis ovatis alternis P. Browne, Civil & Nat. Hist. Jam. 168 foe. T7536: Cordia bourreria L. Syst. Nat. ed. 10. 2: 936. 17 59, sensu stricto. Ehretia bourreria (L.) L. Sp. Pl. ed. 2. 1: 275. 1762. Bourreria rigida Miers, Ann. Mag. Nat. Hist. IV. 3: 204. 1869; Miers, Contrib. Bot. 2: 236. 1869. Crematomia elongata Miers, tom. cit. 308. 1869; Miers, Contrib. Bot. 2: 250. 1869. Crematomia molliuscula Miers, tom. cit. 309. 1869; Miers, Contrib. Bot. 2: 52. 1869 Beureria velutina sensu O. E. Schulz in Urban, Symb. Antill. 7: 62. 1911, maj., non (DC.) Giirke. Evidently widespread in Jamaica, where it ranges from sea-level to 3000 ft. (900 m.), this has been collected in the parishes of St. JAMES (Stearn 615); TRELAWNY (West & Arnold 782); MANCHESTER (Howard & Proctor 14328); CLARENDON (Proctor 10219; Proctor 8215; Lewts 2851); St. ANN (Alexander 564); St. CATHERINE (Harris 6763); ST. ANDREW and KinGsTon (McNab; Campbell 5774, 6151; Harris 11774; Webster 4996; Yuncker 17066, 18149), I have taken as the lectotype of Bourreria baccata Raf. a specimen from P. Browne in the Linnaean Herbarium 254.2 (LINN), of B. rigida Miers one from Houstoun (BM), of Crematomia elongata Miers one from Bancroft (K, BM; drawing of K specimen by Miers), of C. molliuscula Miers one from McNab collected near Halfway Tree (K, BM; drawing of k specimen by Miers). 2. Bourreria velutina (DC.) Giirke in Engler & Prantl, Pflanzenfam. IV. 3a: 1893, as “Beureria”; O. E. Schulz in Urban, Symb. Antill. 7: 62. 1911, pro parte min. Ficures 1E; 2B. Ehretia velutina DC. Prodr. 9: 508. 1845. Crematomia velutina (DC.) Miers, Ann. Mag. Nat. Hist. IV. 3: 310. 1869; Miers, Contrib. Bot. 2: 252. 1869. Small tree or shrub to 4 m. soft whitish erect or ascending and gray. Leaves scattered or ‘(S}Op) Dsouan Diuasinog *4 *(4¥}8) Dutjnjan “g ‘(S\Op) DyDI9nqg HI4asanog “| *(S\OP) PmursstyNsay DIysAOfaUaNo J ‘ad = ‘(sieys) 40j021q "y *(S\Op) vaununjzs “7 ‘(saxenbs) paqnjsqns “AVA DYI1AJOASD “J S(SI}IID) VY IAZOAJSD “AVA DYI1AZO4]SD Dy4ofaudnoy “ *(Sar3D) tapas “J *(SIOP) masavjyrsynm +7 > (1¥}8) Donazo4y20 “J puL DYO{IDAO DUDJUOWaDUAEgD, *¢ ‘(SIBIS) Suarsaonnjs “+ (S}OP) DYOfianD] DUDJUOWaDULaQD] “Yy ‘Bowel ul saideds jo uoNNquysIp Zuimoys sdveyy *] aNNo1y ¢ | | | j [N 1! | TB Lid i) “ay | | | of ——__—_{— woIVWVE ou OS ayes Sop 624 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Inflorescence terminal, loosely corymbose, 6—20-flowered, completely pu- berulous with soft erect hairs to 0.5 mm. long. Calyx at anthesis narrowly campanulate, about 6-7 mm. long, with acute teeth 1.5—-2 mm. long, completely puberulous like the inflorescence with soft erect hairs, in fruit outspread and deeply divided with the teeth 3-4 mm. long. Corolla white when fresh, brown when dried, glabrous; tube about 6 mm. long; lobes broadly obovate, rounded, about 4 mm. long. Stamens exserted; filaments above insertion about 2.5 mm. long, anthers 2 mm. long. Style divided at the apex into two branches 0.5-1 mm. long. Fruit 7-10 mm. in diameter, red. Jamaica. St. CATHERINE: Port Henderson, Lane 423 (xk). Port Henderson Hill, 50-250 ft., vii. 1958, vii. 1954, Webster & Wilson 4930 (BM; IJ). Near Port Henderson, 250 ft., xi. 1957, Yuncker 17477 (BM); Port Henderson Hill, E. of Great Salt Pond, vii. 1958, Powell 558 (BM; IJ). This puberulous species has hitherto been confused with the widespread sericeous Bourreria baccata. Apparently it is confined to the low coastal scrub on the hot dry “dog-tooth” limestone of the Healthshire Hills of St. Catherine, which the earlier collectors Sloane, Bertero, and Lane could Fda RE 2. Photographs showing char: > is 40. A. B, acteristic | apagir-P8 in a of Bour- lecto , P. Browne i tne Rat, hair-covering of pedicel and calyx (from a te f mn einnaean Herb, 245.2) ; B. B. velutina (DC. ) Giurke, C eine of pedicel and ae (from W ebster & Wilson 4930. BM). easily have reached by boat from P is in Sloane’s herbarium vol. between December 1687 and ort Royal. The earliest gathering , fol. 36, lower specimen (Bm), collected se 1689. De Candolle described it in 1971] STEARN, NOTES ON JAMAICAN PLANTS 625 1845 as Ehretia velutina from a specimen collected in Jamaica by Bertero. As his description would also apply to the closely related B. baccata, Dr. Werner Greuter kindly examined the type specimen for me in the De Candolle Prodromus Herbarium at the Conservatoire botanique, Geneva, and compared it as regards pubescence with modern Jamaican material of both taxa. He reports that “the type of Ehretia velutina has patent hairs on both calyces and pedicels. The type specimen is in complete agreement with some specimens from the Port Henderson area in our herbarium and especially with Yuncker 17477.” This establishes con- clusively that the name B. velutina belongs to the extremely local species with spreading hairs of which Miers published a good description from Lane’s Port Henderson specimen under the name Crematomia velutina and not to the widespread species (B. baccata) to which other authors, notably O. E. Schulz, have applied it. For this reason a detailed description of B. velutina, sensu stricto, is given above. FicureE 1F. Nat. Hist. 2: 96. 3. Bourreria venosa (Miers) Stearn, comb. nov. Jasminum periclymeni folio, flore albo Sloane, Voy. Jam. 204, f. 1. 1725. Crematomia venosa Miers, Ann. Mag. Nat. Hist. IV. 3: 305. 1869; Miers, Contrib. Bot. 2: 248. 1869. Beureria velutina var. venosa (Miers) O. E. Schulz in U 5°63, 197i: Also widespread in Jamaica, Bou rban, Symb. Antill. rreria venosa ranges from sea-level to 3000 ft. (900 m.) and has been collected in the parishes of HANOVER (Britton 2252; Proctor 23906; Stearn 146); WestMorELAND (Harris 10236; Britton & Hollick 2069, 2074); St. James (Proctor 23075); St. Euizasetu (Britton 1200; Harris 9703, 9806; Stearn 1034; Webster & Proctor 5326); TRELAWNY (Harris 8788; Howard & Proctor 14164; Howard, Proctor & Stearn 14658; Proctor 24488; Robbins 127; Webster, Ellis, & Miller 8388); MANCHESTER (Adams 9843; Harris 8239, Proctor 10590); St. ANN (Harris 8981, Prior) ; CLARENDON (Lewis 5025; Webster 5112): Sr. CATHERINE ( St. ANDREW (Newill 2795; Proctor 10209); an 4824), It also occurs in the Cayman 1715, 2069; Kings GC 310), LITTLE CAYMAN MAN Brac (Proctor 29032). an unlocalized specimen collected in drawing by Miers of k specimen onal herbarium retained by him “Ocho Rios”: hence, Ocho It should be noted that ho collected in Jamaica 9 on the death of a se of The type of Bourreria venosa is Jamaica by R. C. Alexander in 1850 (K; is in Bm); a duplicate of this, in the pers but bequeathed to Kew in 1903, 1s labeled Rios, on the coast of St. Ann, is the type locality. Dr. Richard Chandler Alexander agent ae i took the surname of Frio I it a penalise tm collections are sometimes cited as tho Alexander, sometimes of Prior. 626 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 4. Bourreria succulenta Jacquin, Enum. Syst. Pl. Carib. 14. 1760; O. E. Schulz in Urban, Symb. Antill. 7: 55. 1911. Mespilus americana laurifolia glabra fructu rubro mucilaginoso Commelin, Hortus Med. Amstel. 1: 153. t. 79. 1697. The only localized Jamaican gatherings referable to Bourreria SUCCU- lenta were collected in St. Andrews Parish at the Grove on a rocky hillside at 750-800 ft. (200-240 m.) by Harris (5760, 12084) and these may possibly belong to an aberrant form of B. venosa in which the style branches have failed to separate. The species is recorded from Cuba, Hispaniola, Puerto Rico, St. Croix, St. Martin, Barbuda, Antigua, Guade- loupe, Dominica, Martinique (locus classicus), St. Lucia, St. Vincent, Grenada, Barbados, Tobago, and Curacao. BORAGINACEAE CorpIA section VARRONIA IN JAMAICA Cordia section Varronta (P. Browne) G. Don, Gen. Syst. Garden. Bot. 4: 382. 1837, consists of shrubs with small flowers clustered in dense heads or spikes. Its members in the West Indies, Mexico, and Central America south to Venezuela are the subject of a critical and detailed revision by Ivan M. Johnston, published in the Journal o f the Arnold Arboretum 30: 85-104. 1949, in which he distinguished 16 species. In 1950 (op. cit. 31: 177-179) he added two more, both Jamaican. The few departures from Johnston’s conclusions, as below, are not meant to diminish their value as a whole; Johnston’s scholarly studies of tropical American Bora- ginaceae have provided a sound basis for all later work on the taxonomy and the nomenclature of the family. The following is a synopsis of the Jamaican species. 1. Inflorescence cylindric, 1.5-3.5 cm. long 2, Leaves densely hairy on lower side. Calyx with evident short stiff hairs eee ee 1. C. brownet. spicuous. Calyx minutely hairy, MO en eae ee: appearing almost glabrous, except under a stron ‘ : 1. Inflorescence globose. Sait . C. jamaicensis. irs on upper side of mature leaf without conspicuously broadened bases. Peduncles mostly axillary. Calyx lobes acute, without linear appendages. sighs ade 3. C. linnaet. e of mature leaf mostly with broadened and often swollen bases. Peduncles terminal or internodal. Calyx lobes elongated into linear appendages 1,5-6 mm. long. 4. Calyx (exclusive of appendages 5-6 m Leaves ovate to very broadl ng. m. long) about 5 mm. long. y ovate. Peduncle with rusty brown hairs. Corolla about9 mm. long... 4. C. clarendonensis. - Calyx (exclusive of appendages 1.52.5 mm. long) 4 mm. long. Leaves lanceolate to ovate, mostly narrowly ovate. Peduncle with whitish hairs. Corolla 4-6 mm. long. ss 1971] STEARN, NOTES ON JAMAICAN PLANTS 627 5. Mature leaves rigid, thick, the lower side usually with deep depres- sions between the veins; hairs on the upper side with raised swollen es. 5. C. bullata. Mature leaves thinner, the lower side normally flat and without depressions between the veins; hairs on the upper side without raised swollen bases: 352255 6. C. globosa var. humilis. un 1. Cordia brownei (Friesen) I. M. Johnston, Jour. Arnold Arb. 31: 177. 1950; Stearn, Proc. Linn. Soc. London 170: 142. 1959. Mountjolya Brownei Friesen, Bull. Soc. Bot. Genéve II. 24: 180. 1933. Jamaica. Parishes of St. CATHERINE, ST. ANDREW and KINGSTON, PortLANp, St. THomas; also on GRAND CAYMAN. For map of Jamaican distribution, see Stearn (op. cit.). 2. Cordia jamaicensis I. M. Johnston, Jour. Arnold Arb. 31: 179. 1950; Stearn, Proc. Linn. Soc. London 170: 142. 1959. ELIZABETH, TRELAWNY, MAN- Jamaica. Parishes of St. JAMEs, St. { distribution see Stearn, loc. cit. CHESTER, ST. ANN, CLARENDON; for map 0 3. Cordia linnaei Stearn, sp. nov. FicurE 3. Ulmi angustifoliae facie Baccifera Jamaicensis folits superne scabris, subtus villosis, floribus flavis perpusillis, fructu botryoide monospermo Plukenet, Imag. Bot. 393. 1696; Phytogr. 4: ¢. 328, f. 5. 1696. ee tain Lantana corymbosa L. Sp. Pl. 2: 628. 1753, P-P- quoad synon. Plukenetu, n C. corymbosa (Desv.) G. Don, 1837. Varronia lineata L. Syst. Nat. ed, 10. 916. [May-June] 1759, ry Nov] Brownei excluso, nomen illegit.; L. Pl. Jam. Pug. (sist. Elmgren) 9. |Nov. 1759; reimpr. in Amoen. Acad. 5: 394. 176 Cordia lineata (L.) Roemer & Schultes, Syst. I. M. Johnson, Jour. Arnold Arb. 30: 91. 1949. 2 Frutex 1.5-3 mm. high. Ramuli graciles brunneoli, et get sm b * : : a j i ilis albidis, Dasl erectis fulvis 0.3 mm. longis hispiduli, vetustiores aoa ee taeviter inenti i i sorum muni rominentibus petiolorum foliolorum delap p ‘ lata, apice longe acuminata, margine sparse petiolata, dissita; lamina lanceo ; ; | rotun- leviterque acute serrata praecipue supra medium, <0 Tiere Conte 4 data, 2.5-10 cm. longa, 0.8-3.5 cm. lata, nervis — G aias. subtus parum prominulis cee Se oa haud bulboso-inflatis supra setis brevissimis et setis a gh intermixtis appressis asperula, su 1-3 mm. longi. Inflorescentiae p™ is hispidulus, quam 1-4 cm. longus, pilis densis patentibus at ©: ine ee ae ie folium subtentum multo brevior, bast petioli sgt ie ee Jato termi- latae adnatus, capitulo solitario globoso Sees natus. Calyx cupularis, 2,5-0 mam. oe a leniter lobatus, lobis 0.5 mm. longis mucro “eet ca. 3-4 mm. longa, ore 1.5 mm. diametro, alba, 0. Sie Veg. 4. 464. 1819, nomen ile git.; 628 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 ENSSTITETE 19% JAMA A FLORA OF JAMAICA Parte Si. ABIES Name scotty: Pastures behind HOPE GaRDRia. Rititede. BOOW?IOD ft. ate. October 22, 1956 Haheint, Opom lela, “h a FIGURE 3. Cordia linnaei Stearn, holotype (pm) Proctor 15789. 1971] STEARN, NOTES ON JAMAICAN PLANTS 629 ae bee ew vix 1 mm. longo. Stamina ca. 2 mm. supra basin corollae : ees lamenta ca. 1 mm. longa; antherae ca. 0.5 mm. longae semi- xsertae. Fructus ca. 3 mm. longus, in vivo ruber, calyce aucto maximam partem velatus. Type: Jamaica: St. ANDREW, pastures behind Hope Gardens, 600-700 ft. 22. x. 1956, Proctor 15789 (BM, holotype; 1J, isotype). Jamaica. Parishes of St. ExLizABETH (Britton 694); MANCHESTER (Harris 9434); CLARENDON (Proctor 16714); St. CATHERINE (Proctor 7815), St. Mary (Proctor 20606); St. ANDREW and KincsTon (Adams 5479, Alexander 555; Powell 748; Proctor 15789, 18265; Yuncker 17767); PortLanp (Fredholm 3271), and St. THomas (Britton 3618, Proctor 5535, 7425). For records outside Jamaica, for Cuba, Hispaniola, Mexico, and Central America to Venezuela, see Johnston, loc. cit. 91. ‘ The seemingly paradoxical act of describing as new a species which has been known since the 17th century, but which nevertheless lacks a legitimate name, is intended to cut through the nomenclatural tangle and make a fresh start by associating a new name with a modern type specimen of known provenance. Johnston began his account of what he called Cordia lineata by referring to it as “a very well marked but previously unrecognized species here associated with a neglected specific name long ago proposed by Linnaeus. The plant is known only from Mexico, Central America and the northern West Indies and is to be found in herbaria mistakenly identified as C. ulmifolia, C. corymbosa or C. polycephala.” Reluctant to give this a new name, Johnston resuscitated a very old one, which unfortunately is illegitimate. The species concerned was known to and Patrick Browne (ca. 1720-1790) and from 1758 onwards to Carl Linnaeus (1707-1778), who had, however, seen no specimen when he drafted the account of Lantana corymbosa for the Species Plantarum 2: 628. 1753: corymbosa 6. LANTANA foliis alternis, flo Periclymenum rectum, salviae foliis majo villosis alternatim sitis, flore & fructu minoribus. S p. 83. t. 194. f. 3. Ray. dendr, 30. Ulmi angustifoliae facie baccifera jamaicensis, villosis alternatim sitis, floribus parvis perpusi spermo. Pluk. alm. 393. t. 329. f. 5. Habitat in Jamaica. ; floris structura nulla mihi certitudo. Refertur ex fide Sloanei; De Here confusion begins. This account is entirely based on the — works of Sloane and Plukenet, an s ecimens from whic reserved in the British Museum (Natural their illustrations were made are pre: ; History). The original of Sloane’s illustration, drawn Leonard Plukenet (1641-1706) ribus corymbosis.t ribus oblongis mucronatis subtus loan. jam. 104. hist. 2. foliis superne scabris, subtus llis, fructu botryoide mono- coun 630 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 in April 1701, is in Herb. Sloane (vol. 6, fol. 86) and belongs to the species later named Cordia polycephala (Lam.) I. M. Johnston. The original of Plukenet’s illustration is also in Herb. Sloane (vol. 102, fol. 132) and belongs to the species named C. linnaei above and called C. lineata by Johnston. Since Sloane’s illustration shows a corymbose in- florescence and Linnaeus states that his own entry had been inserted through the testimony of Sloane —refertur ex fide Sloanei — Sloane's illustration (op. cit. t. 194, f. 3. 1714) must be taken as the lectotype of . corymbosa. However, the Linnaean epithet corymbosa cannot now be legitimately transferred to Cordia; its use is barred by C. corymbosa (Desv.) G. Don (1837) based on Varronia corymbosa Desvaux (1809) ; the earlier C. corymbosa Willd. ex Roemer & Schultes (1819) was pub- lished as a synonym and therefore is not valid. In 1756 Patrick Browne published his Civil and Natural History of Jamaica. He was well acquainted with Linnaeus’s works and adopted his methods except for the new binomial nomenclature. Here Browne defined on p. 172 a genus Varronia, now regarded as congeneric with Cordia, including two species. Under the first species, his Varronia fruticosa foliis rugosis ovatis subhirsutis serratis alternis, capitulis subrotundis, he cited as a synonym Linnaeus’s Lantana foliis alternis floribus corymbosis, ie. L. corymbosa L. (1753), and Sloane’s Periclymenum rectum etc., i.e. its lectotype, although the plant figured (¢. 13, f. 2) and described by him is that now called Cordia globosa var. humilis (Jacq.) I. M. Johnston. Linnaeus accepted Browne’s genus Varronia in his Systema Naturae, ed. 10. 2: 916. 1759, his account being as follows: VARRONIA. Cal. denticulatis recurvatis. Drupa nucleo 4-loculari. lineata. A. V. fol. lanceolatis lineatis, spicis oblongis. Lantana corymbosa Spec. pl. 628. Brown. jam. t. 13. f. 2. bullata. B. V. fol. ovatis venoso-rugosis, spicis globosis. Sloan. jam. t. 195. f. I To be noted in this generic diagnosis is the phrase “Cal. denticulatis recurvatis”: this refers to the recurved calyx lobes characteristic of V. globosa, mentioned by Browne as having “limbus in quinque lacinias tenuissimas longas reflexas vel intortas divisas” and shown in his figure. 1971] STEARN, NOTES ON JAMAICAN PLANTS 631 net’s figure. Actually Patrick Browne had collected three species, i.e. C. linnaei, C. bullata, and C. globosa var. humilis. In the second edition (1762) of the Species Plantarum Linnaeus com- pletely altered the earlier concept which he had called Lantana corymbosa. This now became Varronia lineata, with the original C. corymbosa as a synonym but with a new diagnosis, a citation of Plukenet’s figure and a citation of the Elmgren dissertation. Thereby, in Johnston’s opinion, Linnaeus restricted the application of the epithet /ineata to the species here called C. linnaei but he did not in so doing make the illegitimate epithet /ineata legitimate. Johnston accepted Plukenet’s plate (not Sloane’s) as typifying both Lantana corymbosa and Varronia lineata and he adopted the name Cordia lineata published by Roemer and Schultes in 1819. Since the species concerned (C. linnaei) had no legitimate name under Cordia in 1819, Roemer and Schultes’s use of the name Cordia lineata might be regarded as publication of a new name starting from 1819 instead of a new combination based on an illegitimate name of 1759. Unfortunately they cited as synonyms Varronia humilis Jacq. and V. polycephala Lam., which provided epithets they could have adopted at that time in accordance with their taxonomy although in fact these epithets relate to different species; consequently their name C. lineata is also illegitimate. The next possible name seemed to be Cordia adnata DC., Prodromus 9: 493. 1845, based on a specimen of uncertain origin. However, a photograph of the type in the Prodromus Herbarium at Geneva shows a plant with several-headed peduncles, as well as slightly larger heads and less Fieve leaves than C. linnaei, agreeing in habit with specimens of C. polycephala from Puerto Rico. Since Johnston’s mate name I have accordingly named it C. the same alphabetical sequence and to reca works of Carl Linnaeus. “Cordia lineata,” therefore, appears to have no legiti- linnaei both to maintain it in ll the references to it in the 4. Cordia clarendonensis (Britton) Stearn, comb. nov. : ; Friesen, Varronia clarendonensis Britton, Bull. Torrey Bot. Club 41: 16. 1914; Friesen, Bull. Soc. Bot. Genéve II. 24: 155. 1933. ; Jamaica. Parish of CLARENDON, at Peckham Woods, 760 m., where Harris (no. 10995) collected it in July, 1911. ; 9: 5. Cordia bullata (L.) Roemer & Schultes, . Veg. 4: 462. 1819; I. M. Johnston, Jour. Arnold Arb. 30: 96. 1949. Periclymenum rectum, salviae folio rugoso eres 7 este : Cat. Pl. Jam. 163. 1696; Voy. Jam: 2% 7%) Js ey, Jam. Pub. Varronia bullata L. Syst aa: pa . oe Acad. 5: 394. 1760. (sist. Elmgren) 9 (Nov. : ; Man- Jamaica. Parishes of St. JAMES, TRELAWNY, ST. ELIZABETH, * undo, bullata Sloane 632 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 CHESTER, CLARENDON, St. ANN, St. CATHERINE, KINGSTON and St. ANDREW, ST. THOMAS, 6. Cordia globosa (Jacq.) Kunth, Nova Gen. 3: 76. 1818. umilis (Jacq.) I. M. Johnston, Jour. Arnold Arb. 30: 98, 116, 1949, Periclymenum rectum, salviae folio rugoso minore bullato, flore albo Sloane, Cat. Pl. Jam. 163. 1696; Voy. Jam. 2: 81. ¢. 194. f. 2. 1725. Varronia fruticosa foliis rugosis ovatis subhirsutis serratis alternis, capitulis subrotundis Browne, Hist. Jam. 172. t. 13, f. 2. 1756. Varronia humilis Jacq. Enum. Pl. Carib. 14. 1760; Select. Stirp. Amer. 41. 1763. Cordia humilis (Jacq.) G. Don, Gen. Syst. Garden. Bot. 4: 383. 1837. Florida, Cuba, Cayman Islands, Jamaica (all parishes), Hispaniola, Mexico, Guatemala, Salvador, Nicaragua, Panama. Cordia globosa is a widely ranging species of which the nomenclatural type probably came from the lesser Antilles; the name C. globosa var. globosa is accordingly used for the southern geographical race distinguished by Johnston and stated to range over the lesser Antilles, Venezuela, and Brazil. TOURNEFORTIA Tournefortia astrotricha DC. Prodr. 9: 520. 1845; Griseb., Fl. Brit. W. Ind. Isl. 483. 186. Var. astrotricha. Tournefortia astrotricha was described by De Candolle from a specimen collected in 1821 by Carlo Giuseppe Bertero (1789-1831) in Jamaica, probably on the Long Mountain west of Kingston. It has relatively large leaves, up to 20 cm. long and 11 cm. broad, densely hirsute below with arranged, whence the epithet astrotricha, as indicated by De Candolle’s description of the leaves “ tomentosis pallidis.” It grows in dry rocky places along the southern coast of Jamaica in the parishes of St. THomas (Adams 5493) ; St. ANDREW and Kincston (Har- rts 8930; Graham 8047; Campbell 6413 ; Stearn 832; Webster & Wilson 4851; Webster et al. 8374; Vuncker 17,750); St. CATHERINE (Britton & Hollick 1833 ; Harris 9330); CLARENDON (Lewis & van der Porten 3134; Lewis 2447; Howard 12002); and Mancuester (Proctor 27539). It reappears in the Cayman 7 In St. Elizabeth and on the northwest coast in Hanover and St. James a very similar plant to Tour. inde nefortia astrotricha occurs which has relatively €w hairs on the lower leaf surface. Before visiting Jamaica, I had de- 1971] STEARN, NOTES ON JAMAICAN PLANTS 633 scribed this (though not published it) as a new species confined to the vicinity of Montego Bay. Specific rank now seems too high for it. Var. subglabra Stearn, var. nov. Folia subglabra, subtus pilis sparsis brevibus plusminusve appressis munita. Shrub of straggling sparse habit up to 5 m. (15 ft.) high; branchlets stout, those of current year sparsely pubescent, of previous year glabrous and prominently bossed with leaf-scars, later furrowed. Leaves short- stalked; blade narrowly ovate to narrowly elliptic, the apex acute, the margin entire, the base narrowly cuneate, 2.5—20 cm. long, 1.5~8 cm. broad, both sides glabrous except for a few scattered curved appressed hairs, the primary veins in 7 to 9 pairs; petiole 0.5—-1 cm. long. Jnflorescence several times branched, many flowered, sparsely pubescent with short more or less appressed hairs, the ultimate branches 4-20 cm. long. Calyx about 2.2~5 mm. long, green; segments lanceolate, sparsely pubescent. Corolla scented; tube green, glabrous in the lower half, minutely strigose in the upper half, about 4-6 mm. long; limb 5-parted, about 3—4 mm. across, white becoming reddish, with the segments suborbicular, about 2 mm. long, 2-4 mm. broad. Anthers about 1.2 mm. long, inserted about 2.2-2.5 mm. above the base of the corolla tube. Gynoecium about 2.5-3 mm. long; stigma about 1 mm. broad, mounted on an evident style ag 1.5 mm. long. Fruits subglobose, in the dried state about 4 mm. long an broad, when fresh to 7 mm. long and 9 mm. broad, white, not blotched. 1908, Harris 10351 (BM, XK, isotypes; NY, holotype). Near the mouth of Great River, west of eed . , ili 1 (ny). Hanover: Oran Boe rie ay Pag eo, Great Pedro Bluff, 50 ft. 11. v. 1956, Stearn 1029 (A; BM). GraNp CayMaN: near Beach Club 8. vi. 1967, Brunt 2080 (BM). Hotel, West Bay Road, Georgetown, Probably the Grand Cayman plant collected by caer ee an pendently evolved from 7. astrotricha var. soto a ’ directly derived from the western Jamaican ate anne with rather large Tournefortia astrotricha belongs to a group © sre 3.5 cm. broad and —— site veer ae ee Suoiibiealsx or oblong sometimes up to 30 cm. long an : Aue ies, i.e. T. astro- corla lobes abot a lon 2 ad. Te i ene Gr tricha DC., T. bicolor Swartz, +. ; may be distinguished as follows: 1. Anthers with their tips protruding from th of the corolla. .......--- a T. staminea. protruding. 1, Anthers completely enck 2. Stigma almost or quite ses 634 JOURNAL OF THE ARNOLD ARBORETUM [ VoL: ($2 3. Hairs of corolla 6-8 mm. long, very abundant. Leaves very hairy. Be ER be GEN eet cs U, VEtr tee nt ees eg a T. hirsutissima. 3. Hairs of corolla 4 mm. or less long, fewer. Leaves almost or quite b Ue ee OY ai) vi dis T. bicolor. 2. Stigma mounted on an evident style or prolongation of the ovary 1-1.5 SR Oe Oe ee a ee . astrotricha. 4. Leaves densely hairy beneath..................... var. astrotricha. 4. Leaves very sparsely hairy beneath................. var. subglabra. These taxa have distinctive ranges (FicuRE 1 C and D). Thus Tourne- fortia astrotricha is exclusively coastal, while T. staminea inhabits the western upland and montane region at 800 ft. (240 m.) to 1600 ft. (480 m.) in the parishes of Hanover, Westmoreland, St. James, and Trelawny, and 1. bicolor the eastern region at 1000 ft. (300 m.) to 3500 ft. (1060 m.) in the parishes of St. Mary, St. Andrew, Portland, and St. Thomas. The “chiggernut,” T. hirsutissima, probably occurs in every parish and ranges from sea-level to 700 m. SOLANACEAE SoLANUM* Solanum americanum Miller, Gard. Dict. ed. 8, art. Solanum no. 5. 68 Var. nodiflorum (Jacq.) Edmonds, comb. et stat. nov. Solanum nodiflorum Jacq. Coll. Bot. 2: 288. 1789; Icon. Pl. Rar. 2: ¢. 326. 1789. Revision of the South American taxa of Solanum L. sect. SOLANUM ($ Maurella Nees, § Morella (Dunal) Bitter) typified by S. nigrum L. has made apparent the need to reduce onspecific. Early authors placed sensu lato, which breaks up, however, into a number le on correlated morphological and cytological fea- easonably treated as species. The name S. nigrum of taxa, distinguishab B’ge ss Fo eS oo x5 4 p80 sak o Te M3 Po oad RE om a5 ¢ Q BER eT ako, introduced from Virgini in the British Museum (Natural Histor * By Jennifer M. Edmonds. 1971 ] STEARN, NOTES ON JAMAICAN PLANTS 635 a specim . : _ eamaager pee cultivated in Vienna, now in the British Muse Guin. Bin pai “te is glabrescent; Professor G. T. S. Baylis agent wr the pollen (mounted mea Zealand 85: 381. 1958) the mean diameter of Be toes apree tall in lactophenol blue) of this specimen as 23». The the name S. sie ex floral and vegetative features. Acuacdeeaite Ditliest specific-t num has, on grounds of priority, to be adopted as the S. nigrum com ans whee: ' peas lants of the diploid group within the Meciniens ae 2 hitherto distinguished as S. nodiflorum. Jamaican mien, being ee nodiflorum. ee ae es: separating Solanum americanum from S. nigrum, sensu ait ; A prate ta significant. Thus 5. americanum has smaller flowers Miiting dis mm. across and anthers ca. 1-2 mm. long, erect Ecicad aval s, pia globose berries 5—7 mm. wide with constantly well Dien ata s and smaller seeds (only 1-1.5 mm. Jong), as well as smaller DOA vs ccs ape y with its diploid state. S. nigrum has larger flowers Blick o: olla 10-18 mm. across and an inion green ovoid berries 6-10 mm. wide wit mi Sedaapanionic = reflexed, and larger seeds (1.7— ies o iameters of pollen grains of Jamaican species of S. america- r. nodiflorum collected by Stearn, measured by Baylis after mount- ing in lactophenol blue and recorded on the sheets (pm) are Stearn 113, 271, 22.1p (21.5-23.7), ange Mma) 181, 243p (23.7-25.8p); Beato, Sheds 365, 24.1p (21.5-28.0n), 432, 19.8p (19.4-21.5p), thncs ‘6 (19.4-21.5p), 584, 23.9 (23.7-25.8u).. The anthers on these ‘amon vary from 1.0 te 17 mm. but are mostly 1.5 mm. long. These ai Nor ents fit in with Baylis’s finding in New Zealand material of “Seng um a mean pollen diameter of 21p and a range of 20-27. Baylis the range of pollen diameter in S. nigrum, similarly mounted in lactophenol blue, to be 27-38». SCROPHULARIACEAE Alectra fluminensis (Vell.) Stearn, comb. nov. A scagoachaaag wus tebts Richard, Actes Soc. Hist. crophularia fluminensis Vellozo, Fl. Flumin. 263. 182 87. 183 . 1846; Griseb. Fl. Brit. W. - 273. t. 47. 1862. Nat. Paris 1: 111. 1792. 5;-F1L. Flumin. Ic. 6: ¢. Alectra brasiliensis Benth. in Ind. Isl. 429. 1862; J. A. Sc Alectra melampyroides (Rich.) Kuntze, Rev. Gen. : lin 15: 435. 1944, no A. melampyroides Benth. 1846. h.) Chodat & Hassler, Bull. Herb. Boiss. Hl. + & Wilson, Sci. Surv. Porto 291. 1904 Pennell ex Britton +, FL Trinidad & Tobago 2: M elasma melampyroides (Rich.) Rico Virgin Is. 6: 188. 1925; R. E. D. Baker in 285. 1954. This species has long been known from P and Trinidad and is apparently widesprea the Lesser Antilles, uerto Rico, th America d in northern Sou 636 JOURNAL OF THE ARNOLD ARBORETUM [von. 52 (Guyana, Venezuela, Brazil, Paraguay) but was first found in Jamaica, in Hanover parish on Dolphin Head, in 1960 by Dennis Adams (no. 8641) and later in Westmoreland parish at Seedy Pond, southeast of Mount Grace, in 1966 by George Proctor (no. 27703). Vellozo’s epithet flumi- nensis refers to Rio de Janeiro (Flumen Januarii ), Brazil. Although Melasma Bergius (1767) and Alectra Thunb. (1784) have from time to time been treated as congeneric, H. Melchior in his papers, “Beitrag zur Kenntnis der Gattung Melasma” (Notizbl. Bot. Gart. Berlin 15: 119-127. 1940) and “Die Gattung Alectra Thunb.” (loc. cit. 423-447. 1941), has provided reason to keep them apart. ACANTHACEAE Justicia The definition of the genus Justicia L. and the typification of its name provide interlocked problems of taxonomy and nomenclature for which every possible solution is certain to cause nomenclatural disturbance. As commonly accepted, e.g. by E. C. Leonard (Contr. U. S. Natl. Herb. 31: 487-645. 1958), R. W. Long (Jour. Arnold Arb. 51: 302-309. 1970) and the authors of most Floras, Justicia includes some 300 species found in €cognizing the possibility of its later The name Justicia was first used by Linnaeus in 1737 (Genera Plan- 1971] STEARN, NOTES ON JAMAICAN PLANTS 637 tarum 4, no. 12). His heading reads “JUSTICIA Houst. A. A. Adhatoda Tournef. 79. Ecbolium Riv. 1. 129.” The cryptic “Houst. A. A.” refers to an intended publication on new genera by William Houstoun (1695— 1733). Made acquainted with Houstoun’s manuscript and drawings by Philip Miller, Linnaeus expected this would form part of the Philosophical Transactions of the Royal Society of London, the title of which he latinized as Acta Anglica and then abbreviated to “A. A.,” but Houstoun’s work was not published until 1781, after the death of all concerned, by Sir Joseph Banks under the title Reliqguiae Houstounianae. Houstoun based his genus Justicia on American species now placed in the genus Dicliptera. Linnaeus took from him only the name Justicia commemorating Houstoun’s fellow countryman James Justice (d. 1762 or ’63), Clerk of the Sessions at Edinburgh. He based his own generic description in the Genera Plan- tarum on the Asiatic species later named Justicia adhatoda L. (Adhatoda zeylanica Med.), merely substituting the name Justicia for Paul Hermann’s Adhatoda used by Tournefort (Inst. Rei Herb. 175. ¢. 79. 1700), because this was derived from the Tamil vernacular name “Adathodai,” Sinhalese “Adathoda,” meaning “goat not touch,” in allusion to its bitter leaves, and Linnaeus rejected such nomina Barbara. His reference “Riv. 29” is to a plate under the name Ecbolium in Rivinus, Icones Plantarum, quae sunt Florae irregulari monopetalo (1690), which illustrates Justicia adha- 54) designated this as its lectotype, thereby making Justicia, sensu stricto, equivalent to Adhatoda Miller. In the Species Plantarum (1753) Linnaeus expanded his concept t - genus Justicia to include ten other species, although he did not ‘eggs : ir generic description in the Genera Plantarum (1754) to fit aig n these species differ so much among themselves that later authors have pu them in different genera as indicated below. nade lanica Medicus, 1790; A. vasica Nees, 1832. Adhatoda: later Adhatoda zeylanic ety Alston; B. Homeswom ni 5 hs Ta Ecbolium: later Ecbolium viride a ‘ a7ay2 unpesg pue viuojdgQ jo saiads uvouauy yINOS “gq ‘| (snyzUDIDYI ed _ e Pee 646 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 genera Oplonia and Psilanthele (Acanthaceae)” in Bull. Brit. Mus. (Nat. Hist.), Bot. 4: 259-323. 1971. Forsythiopsis was based on a Madagascan species, F. baroni Baker, later found to be conspecific with the earlier Jus- ticia vincoides Lam., which Nees had included in Anthacanthus. Despite their wide geographical separation (cf. Ficure 5), the American and Mada- gascan groups resemble each other so closely as to leave only the glabrous style of the one and the hairy style of the other for their morphological separation and their generic distinction. Both groups are heterostylous. It seems highly improbable that these manifold resemblances between them are the result of convergent evolution from different stocks rather than parallel or slightly divergent evolution from the same stock. If it is the latter, as I believe, then this advanced genus must have developed before the breakup of the Gondwanaland landmass in the Cretaceous period over 70 million years ago. The taxa distinguished in the Jamaican flora are O. acicularis (Justicia acicularis Swartz, Anthacanthus acicularis (Swartz) Nees), O. armata with two varieties, var. armata (J. armata Swartz) and var. pallidior (Psilanthele minor Lindau), O. jamaicensis (P. jamaicensis Lindau) and O. microphylla (Justicia microphylla Lam.). The genus Psilanthele Lindau is restricted to Ecuador; Jamaican species included in it belong to Oplonia. RUELLIA Ruellia sceptrum-marianum (Vell.) Stearn, comb. nov. Pedicularis sceptrum-marianum Vellozo, Fl. Flum. Descr. 269. 1825; Icones : #. 104, 1831. Ruellia acutangula Nees in Flora (Regensburg) 21 (Beibl. ii.): 61. 1838; Hooker f. in Curtis’s Bot. Mag. 104: t. 6332. 1878; Lindau in Urban, Symb. Antill. 2: 196. 1900. Arrhostoxylum acutangulum (Nees) Nees in Mart. FI. Bras. 9: 58. 1847; Nees in DC. Prodr. 11: 210. 1847. of its ever having become naturalized in Jamaica MAPS hich I have examined specimens. xploration of Jamaica during the past tor of the Institute of Jamaica, enough 1971] STEARN, NOTES ON JAMAICAN PLANTS 647 STeARN, W. T. 1958. A key to West Indian mangroves. Kew Bull. 1958: 33-37. [ Avicennia germinans. | _ 1959. A botanist’s random impressions of ee Proc. Linn. Soc. net 170: 134-147. [Columnea, Cordia, Ipomoea. | . Jamaican and other species of re (Sapotaceae). Jour. ved Arb. 49: 280-289. [Bumelia americana. | _ 1969. A synopsis of Jamaican Myrsinaceae. Bull. Brit. Mus. (Nat. ar Bot. 4: 143-178. [ Ardisia, Wallenia, Myrsine 1969. The Jamaican species of Columnea and “Alloplectus (Gesneria- ceae). Bull. Brit. Mus. (Nat. Hist.) Bot. 4: 179-236. [Columnea, Allo- plectus, Achimenes. _ 1971. A survey of the tropical genera Oplonia and Psilanthele (Acan- thaceae). Bull. Brit. Mus. (Nat. Hist.) Bot. 4: 259-323. [ Oplonia. | DEPARTMENT OF BOTANY British Museum (Nat. Hist.) CroMwELt Roap, Lonpon S. W. 7 ENGLAND EXPLANATION OF PLATE PLATE I jlosa Stearn: A, s staminate inflorescence A inden of tee erigiiel ate ms leaf; c cluster of rn ge owers; a. G, fru ees f, b, saa of leaf. K’ F. segregata (Jacq.) Krug & Urban: a, upper sidé shoe 9015.] Plate [A-E from Harris 9017, F-H from Harris 9056, K from ‘Harris drawn by Derrick Erasmus. PLare ul an . r nm) au tie j nel ei WA - ‘ - Aces wat FUN » “s oN v y 4 yj . Z a y ba y ? EY, Ci > / ) A / Aa pe i Ne i; OH IN, 7 ie =. A CR) 5 UY feo), ! ‘1 y ft . 4 yy A 40% \ iy iy . 4 Te oS ' iS uty : . ES, t?) % \ A) here 2 - 3 ‘ j AES } A ref. $y ® ' Ns t } I uG \ ADs. bs it i VL t= - A eh 52 D 2 0) = a XV v vi O ef Jour. ARNOLD Ars. VoL. 52 STEARN, FoRESTIERA THE LINACEAE IN THE SOUTHEASTERN UNITED STATES * KENNETH R. ROBERTSON LINACEAE S. F. Gray, Nat. Arr. Brit. Pl. 2: 639. 1821, “Lineae” fA. P. de Candolle, Théorie Elem. Bot. 214. 1813, “Linées”], nom. cons (FLAX FAMILY) 5 mbers] with simple and mostly alter- nate (rarely all opposite or whorled) leaves; stipules present or absent. Inflorescences various, cymose, racemose, corymbose, or paniculate [rarely solitary]. Flowers perfect, regular, 5|4]-merous; sepals and petals free h or slightly connate at the base, the sepals quincuncially imbricate, the on and often clawed; disc absent; stamens Herbs [shrubs, trees or woody cli ia sometimes present; ovary times partially or completely sion of false septa, ovules Js, mostly free or superior, syncarpous, 5|2- subdivided into 10 [4-8] locules by the intru 2 per carpel, epitropous, styl sometimes partly to completely united; filiform, Fruit a septicidal capsule [or drupe] ; laginous epidermis, embryo straight [or slightly curved]. (Linaceae sensu Hutchinson, 1967; Linoideae of Winkler, excluding Nectaropetaleae, and of Scholz.) TYPE GENUS: Linum L. About 13 genera and 275-350 species of worldwide distribution, but with relatively few species in the Southern Hemisphere. Four genera ag represented in North America: Linum, with about 41 (including 4 or 5 1 Prepared for a generic flora of the southeastern the Arnold Arboretum and the Gray Herbarium 0 through the support of the National Science Foun investigator, Carroll E. Wood, Jr.). This treatment vn, Arnold Arb. 39: 296-346. 1958). T project of ade possible The descriptions are rimarily on rmation in brackets. References I have not seen are m Arkansas, and Louisiana. arked area, with additional info by an asterisk. preparation of uals for their assistance in the nd careful editing, a R I am very grateful to several individ : the manuscript: Dr. Wood for his most aay suggestions 7 Dr veeis Dr, Bernt for thorou hly reading the manuscript, ° i emice G. Schubert 10 iew and thoughtful comments, and Mrs. Nancy stration g (Wayne State University) for bis revie 2 Dunkly for checking the bibliographies and for her accurate typing. a a : was drawn by Karen Stoutsenberger largely from preserved material collec y Dr. Wood. 650 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 introduced) species primarily in the Southwest and on the Atlantic and Gulf coastal plains; Radiola J. Hill (Millegrana Adanson), with R. linoides Roth introduced and established in Nova Scotia; Hesperolinon (Gray) Small, with 12 species largely confined to California; and the monotypic Sclerolinon C. M. Rogers, with S. digynum (Gray) C. M. Rogers ranging from Washington south to California. Reinwardtia Dumort. has been introduced into the West Indies from India. These genera, plus the Asian Anisadenia Wall. and Tirpitzia Hallier f., comprise the tribe Lineae (mostly herbs or subshrubs, stamens the same number as the petals, fruits capsular), which is generally of temperate regions. The tribe Hugonieae Planchon (trees, shrubs or woody climbers, stamens twice the number of the petals, fruits drupaceous) is mostly tropical and contains six genera, including the South American Hebepetalum Bentham and Roucheria Planchon. The family is evidently not indigenous to southern Central America, adjacent northern South America, and most of the West Indies. Closely related to the Linaceae, or sometimes included within it, are the Humiriaceae, Ctenolophonaceae, Txonanthaceae, and Erythroxylaceae, which share the simple leaves; regular flowers: diplostemonous androecium or stamens in only | whorl, rarely numerous, with the filaments basally united; nectiferous disc absent or small; and superior ovary. The Gera- niaceae, Oxalidaceae, Malpighiaceae, Zygophyllaceae, and Polygalaceae are also sometimes considered to be related to the Linaceae. The Linaceae have recently been placed in the Linales (Cronquist), the Geraniales (Winkler, Scholz, Thorne, and Takhtajan), and the Malpighiales (Hutchin- son, in his Lignosae). Hallier placed the Linaceae in his Guttales as a derivative of the Ochnaceae and considered it as the evolutionary center of origin for several lines, including the Passiflorales, Myrtales, Sapotales, Santalales, and some families of the Tubiflorae. Airy Shaw suggests that through Anisadenia the Linaceae are connected with the Plumbaginaceae, and Erdtman (1969) says that on palynological grounds these two families are close. The floral anatomy and morphology of Linum and certain other genera have been the subject of numerous papers (cf. Sharsmith and Narayana). The stamens in the Lineae are monadelphous, with a cup from which the filaments arise. In some species and genera nonvascularized structures, usually called staminodia, alternate with the stamens. Sharsmith found in Hesperolinon, Radiola, and a few species of Linum that the petals arise in exactly the same position as the staminodia found in rs of this family should be made ng and be pressed immediately. Reported chromosome numbers are Hugonia (3 spp.) 2n = 12, 24, 26; Radiola linoides, 2n = 18; Reinwardtia trigyna, 2n = 20, 22; Sclerolinon 1971] ROBERTSON, LINACEAE 651 digynum, 2n = 16; and Linum, 2n = 12, 16, 18, 20, 24, 26, 28, 30, 32 34, 36, 52, 54, 60, 62, 68, 72, 84. NGS Pollen grains of the Linaceae are diverse and of some taxonomic impor- tance. The tribe Lineae has 3-, 4- or 6-colpate, pantocolpate, or panto- porate grains with a granular to baculate sexine, while the Hugonieae have mostly tricolporate grains with the sexine usually strongly baculate and often tegillate. The pollen of Anisadenia pubescens and Reinwardtia indica was reported as nonaperturate by Erdtman, but Saad describes both as pantoporate. Pollen morphology supports the removal of Ctenolophon and the Humirioideae from the Linaceae. Pollen of the woody members of the family is of a more advanced type than that of the herbaceous members. especially Hugonia, and the Plumbago group of the Plumbaginaceae. Palynological similarities to the Geraniaceae and Oxalidaceae have also been noted. Anatomical studies indicate that the Linaceae sem group and its division into several families is warran and Erythroxylaceae are considerably Linaceae and Humiriaceae have numerous, the latter family more primitive than the former and similar to Ctenolo- phon. Within the Linaceae, Indorouchera has There is no evidence from wood anatomy that the woody members of the family are derived from herbs. su lato is nota natural REFERENCES: Barton, H. Linacées. Hist. Pl. 5: 42-66. Hartoc, The natural history of plants. 5: 42-66. London. 1878.] BENTHAM, G., & J. D. HOOKER. Lineae. Gen. Pl. 1: 241-246. 1862. [Treatment by BenTHAM; includes Erythroxylaceae. | BoLKHovskIKH, Z., V. G qveseva, & O. ZAKHARYEVA. aeruahd numbers of flowering plants. (Russian and English prefaces.) 926 PP. Leningrad. 1969. [Linaceae, 412-414. ] ; Carus, J. F. The medicinal and poisonous flaxworts of India. sites €s seid Nat. Hist. Soc. 42: 167-170. 1940. [Hugonia, ‘Linum, Reinwaratia, 0 cheria. | CaANDOLLE, A. P. pe. Lineae. Prodr. 1: poets ee ae Cuarrecasas, J. A taxonomic revision of the Humiriaceae. re SAO Herb. 35: / 5-214. 24 pls. 1961. {Humiriaceae distinct from Linaceae, but related to its woody members. ] ; Davis, G. L. Systematic embryology oF the . . 161, 162; references.) mia acne 2 eee i W ed. Vistas in botany. 4: 23-54. 1874. [English transl. M. M. s. x + 528 pp- New n. W. B. L, any. 7 pls, 1964. [Linum grandiflorum, L. rigidum, 38 and pl. 1, Linaceae and Plumbaginaceae, 50.] : . Pollen morphology and plant taxonomy. eT ies ae ba print and new addendum. xiv 3: 543 bp. New. 0m 241-245; references to earlier papers. | - Handbook of palynology- 486 pp- chart. Copenhagen. 1969. [Linaceae, 652 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 105; Linum usitatissimum, 258, 259; L. hirsutum, 294, 295; L. alpinum, 304, 305; L. perenne, 306, 307.] Guosu, S. S. Linaceae. Jn: K. A. CHowpHury & S. S. GuosH, Indian Woods 2: 1-3. pl. 31(184). 1963. [Wood descriptions of Hugonia mystax L., Hatter, H. Beitrage zur Kenntnis der Linaceae (D. C. 1819) Dumort. Beih. Bot. Centralbl. 39(2): 1-178. 1921. [Expands Linaceae to include Humi- riaceae, Erythroxylaceae and other families. | HeEcNAvER, R. Chemotaxonomie der Pflanzen. Band 4. Dicotyledoneae: Daph- niphyllaceae-Lythraceae. 551 pp. Basel & Stuttgart. 1966. [Linaceae, 393- 401; many references. | Hermscu, C., Jr. Comparative anatomy of the secondary xylem of the “Grui- nales” and “Terebinthales,” of Wettstein with reference to taxonomic grouping. Lilloa 8: 83-198. pls. 1-17. 1942. [Linaceae, 92-96, pl. 1; b ‘Telationships between Linaceae and Humiriaceae, Erythroxylaceae and thers. | J others. HutTcuHINson, J. The genera of flowering plants. Vol. 2. xii + 659 pp. London. 1967. [Linaceae, 595-600. Family restricted to Hugonieae and Lineae (including Anisadenia).} Knut, P. Handbook of flower pollination. Vol. 2. (Transl. J. R. A. Davis.) viii + 703 pp. Oxford. 1908. [Linaceae, 214-217. In German ed. cf. 2(1): 225-227. 1898.] LEINFELLNER, W. Beitrage zur Kronblattmorphologie. VII. Die Kronblitter einiger Linaceen. Gsterr. Bot. Zeitschr. 102: 322-338. 1955. [Linum, Hesperolinon, Hugonia, Reinwardtia.| Mavritzon, J. Zur Embryologie einiger Gruinales, Sv. Bot. Tidskr. 28: 84-102. 1934. [Radiola linoides, 84-91.] Mertcatre, C. R., - Cuatx. Linaceae. Anat. Dicot. 1: 268-273. 1950. [Including Ixonanthes & Ctenolophon. | Narayana, L. L. A contribution to the floral anatomy and embryology of Linaceae. Jour. Indian Bot. Soc. 43: 343-357. 1964. [Durandea pentagyna, } j » L. rubrum, Reinwardtia trigyna. ment with complex classification schem species. | Rao, D., & L. L. Narayana. Embryology of Linaceae. Curr. Sci. Bangalore 34: 92, 93. 1965. | Roucheria, Ochthocosmus, Ixonanthes.| REICHE, K. Linaceae. Nat. Pflanzenfam. III. 4: 27-35. 1896. (Supplements PILGER.) e and descriptions of all known G linon, a new genus in the Linaceae. Madrofio 18: 181-184. 1966. [Monotypic genus based on Linum digynum Gray from western North America. ] Saap, S. I. Pollen morphology and sporoderm stratification in Linum. Grana Palynol. 3(1): 109-129, 7 pls. 1961. [Addendum discusses phylogenetic trends o: Stratification in Linum, Reinwardtia, Hugonia, Durandea; similarities found between pollen of some Linaceae and Plumbaginaceae and Polemoniaceae. | . - Palynological studies in the Linaceae. Pollen Spores 4: 65-82. 1962. [ Linum, Hesperolinon, Radiola, Anisadenia, Reinwardtia, Indorouchera, Hebepetalum, Hugonia, Durandea, Pilbornea. | 1971] ROBERTSON, LINACEAE 653 ScHoiz, H. Linaceae. Jn: H. Melchior, Engler’s Syllabus der Pflanzenfamilien. ed. 12. 2: 253, 254. 1964. [Linaceae includes Linoideae (Hugonieae, Anisa- denieae, Lineae), Ctenolophonoideae, Ixonanthoideae, Humirioideae. SHarsMiTH, H. K. The genus Hesperolinon (Linaceae). Univ. Calif. Publ. Bot. 32: 235-314. 1961. [Information on floral morphology of Linum, Anisa- denia, Reinwardtia, and Radiola as well as Hesperolinon. | SMALL, J. K. Linaceae. N. Am. Fl. 25(1): 67-87. 1907. SoukcEs, R. Développement de l’embryon chez le Radiola linoides Roth. Bull. Soc. Bot. France 84: 297-306. 1937. Urpan, I. Lineae. Jn: C. F. P. Martius, Fl. Brasil. 12(2): 457-471. pls. 97-101. 1877. Warsure, E. F. Taxonomy and relationship in the Geraniales in the light of their cytology. New Phytol. 37: 130-159, 189-210. 1938. [Linaceae, 202-205.] Wius, J. C. A dictionary of the flowering plants and ferns. ed. 7. (Revised by H. K. Airy Shaw.) xxii + 1214 pp. + liii pp. (Key to the families of flowering plants.) Cambridge, England. 1966. [Linaceae, 656; family related to Plumbaginaceae. | WINKLER, H. Linaceae. Nat. Pflanzenfam. ed. 2. 19a: 82-130. 1931. Tribe LrnEaeE [Eulineae Planchon] 1. Linum Linnaeus, Sp. Pl. 1: 277. 1753; Gen. Pl. ed. 5. 135. 1754. Perennial or annual herbs [or shrubs], sometimes woody at the base, with one or several + erect stems arising from a taproot or rootstock, usually unbranched below the inflorescence, stems mostly glabrous, infre- quently scabrous [or pubescent], and often striate. Leaves ange Mer alternate, sometimes opposite below, infrequently all opposite [or whor } ; ovate or obovate to linear, entire, denticulate or glandular toothed, mostly sessile with one midvein or several prominent parallel veins; stipules florescences terminating the main stems, to many flowered [rarely solitary |, ng as scorpioid cymes, panicles, racemes [white or red]. Sepals 5, free or united apices, rarely obtuse, + equal or the inner or] all or only the inner glandular too ee prominent veins, persistent or infrequently deciduous, . d Stamens 5, alternate with the petals, basally ae ” ee igigel the ovary, nectar glands often on the outse - eg aa 2-locular staminodia alternating with stamens or absent; ant eee aa a eete at anthesis, opening lengthwise. Stigmas 5, distinc He Dae Se [to filiform] ; styles 5, free oF sian ted nce ssa S-carpellate but stylous [or heterostylous]; ovarTy a cine | SE NO ‘ See i 10-locular by the intrusion ele nie the os septa sometimes ciliate, ovules 2 in each carpel, pen . , 654 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Ficure 1. Linum. a-m, L. Carteri var. Smallii: a, plant with flower and immature capsules, < 1/4; b, flower, X 2; c, adaxial surface of pubescent ligule with small abaxial pouches vi surface of outer sepal, : flower (petals : : 6. n, L. flo var. floridanum: nearly mature capsule with persistent sepals (two removed) and distinct styles, a car- tilaginous plate lacking, & 6. axile, epitropous. Fruit a Septicidal capsule, often beaked, dehiscing into 10 one-seeded or 5 two-seeded Segments, the latter type with triangular 1971] ROBERTSON, LINACEAE 655 cartilaginous plates at the bases, usually dehiscing first between the false septa, rarely indehiscent; seeds sometimes retained within the capsule segments. Seeds compressed, smooth or slightly scalariform, mucilaginous when moistened; endosperm formation of the Helobial type, copious, scanty or absent in the mature seed; embryo straight. Embryo sac devel- opment of the Polygonum type, embryo development of the Solanad type. (Including Cathartolinum Reichenbach.) LECTOTYPE SPECIES: Linum usitatissimum L.; see J. K. SMALL, N. Am. Fl. 25: 67. 1907. (Ancient Latin name for flax and linen.) — FLAX. A genus of 150-225 species widely distributed in temperate and sub- tropical regions, particularly in the Northern Hemisphere. Major centers of distribution are the Mediterranean region, the Balkan Mountains, Anatolia, the southwestern United States and Mexico, and the Coastal Plain province of the eastern United States. About 41 species occur in North America. In our area about fifteen species and seven varieties represent three major groups characterized by yellow, white, or blue flowers. The genus has not been examined on a world-wide basis since 1931, and the subgeneric classification is in need of re-evaluation, particularly the relationships between the North American species and those of South America, South Africa, and the Mediterranean region. All species indigenous to our area belong to the “yellow-flowered” group, which is further characterized by small to medium-sized flowers, capitate stigmas, glands on all or the inner sepals, short pedicels, and all or at least the upper leaves alternate. These species are often placed in section CaTHARTOLINUM (Reichenb.) Griseb., but are distinct and more closely related to the South American and Old World species referred to section i i i long Reiche. These North American species, @ menses iomerii treated in the genus Cathartolinum Reichenb. by Small, who set up rank, although some have cons! have refined this scheme and recognize only five The line Schiedeanum complex, characterized by —— st usual presence of stipular glands, persistent and er : de tit staminodia alternating with petals, essentially distinct styles, e oe a separating into ten one-seeded segments, includes ~- oe ok greatest diversity in east-central Ne praarerae po pots L. arenicola (Small) Winkler, 2n = 36, aistingt™ lar glands, is endemic ; Florida, and on the sip 0 ; ) Engelm. ex Gray, — bs cet and Guatemala, and L. a, Planchon, 2n = 72 ‘of central Mexico. Linum bahamense —— ot eo edaneda ‘ group of central Mexican species than to L. arenicord. groups to the pine wo ly related to a 656 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 The perennial Linum virginianum complex, comprising seven species and distinguished by the capsule separating into 10 one-seeded segments, absence of staminodia, persistent sepals with glandular teeth present only on the inner sepals or absent, distinct styles, and absence of stipular glands, is largely confined to the eastern half of the United States. Linum inter- cursum Bickn., 2n = 36, with capsules ovoid and pointed at the apices, the septa ciliate, and the false septa clearly incomplete, occurs in the semishade of oak or pine woods along the Coastal Plain from Massachusetts south to the Carolinas and inland to Tennessee, northern Georgia, and Alabama; an isolated population is found in northwestern Indiana. Dis- tinguished by subglobose capsules with rounded apices, nonciliate septa and inner sepal margins with conspicuous stalked glands, L. medium (Planchon) Britton var. texanum (Planchon) Fern. (L. Curtissii Small), n = 36, is widespread in the East and also occurs in the Bahamas. The tautonymic variety, 2n = 72, with thicker and wider leaves, is largely confined to the shores of the eastern Great Lakes. Particularly variable in southern Florida, this species needs further study. Linum virginianum L., 2n = 36, has small subglobose capsules with umbilicate apices, sparsely ciliate septa, complete false septa, early dehiscing capsules which leave many specimens without fruit, and sepals glandless or with inconspicuous sessile glands. It is widespread in open woods of the northern and eastern United States, becoming infrequent in the southern Atlantic Coastal Plain. Closely related to the two preceding species, L. striatum Walt., 2n = 36, is distinguished by its predominantl inflorescences, nonciliate ocarpum C. M. Rogers, 2n = 36, has Questionably distinct from the latter variety, Mobile, Alabama, and possibly Georgia, has predominantly 0 abruptly pointed apices and nonciliate septa e Linum sulcatum “complex” includes only L. sulcatum Riddell, 2n = 30, a common annual i iri 1971] ROBERTSON, LINACEAE 657 of 15 are characteristic of the L. rigidum complex. Rogers (1969) suggests that although “‘it is perhaps unlikely that L. su/catum is on a direct line of evolution connecting the L. Schiedeanum and L. rigidum complexes, . . . plants resembling L. sulcatum may have served as the link between the otherwise very distinct L. Schiedeanum and L. rigidum complexes.” In our area, only L. sulcatum has basally united styles and all five sepals glandular- toothed and persistent. Certain collections from the Southeast with dark herbage and racemose inflorescences are sometimes segregated as var. Harperi (Small) C. M. Rogers (L. Harperi Small). Typical members of the annual L. rigidum complex have capsules separating into five 2-seeded segments, no staminodia, sepals deciduous and at least the inner glandular-toothed, and styles united nearly to the summit, The group is composed of ten species, mostly of the southwestern United States and northern Mexico; two tetraploid populations are endemic to southern Florida, and at least three species occur in the extreme western part of our area. The Florida populations have been considered to either varieties of L. rigidum Pursh, 2n = 30, or a distinct species. Rogers (1968a), drawing on the biosystematic work of Mosquin & Hayley, treated the populations with puberulent or scabrous stems, stipular glands, petals 9-11 mm. long, and found only in the pinelands southwest of Miami as L. Carteri Small var. Carteri, 2n = 60. The other populations in Dade, Collier, and Monroe counties with + glabrous stems, stipular glands usually lacking, and petals 11.5-17 mm. long were identified as L. Carter . Smallii C. M. Rogers, 2n = 60. Linum Carteri is most closely relat 4 L. rigidum vat. Berlandieri (Hooker) Torrey & Gray, 2" = 30, ss reddish pigment in the petals and thinner-walled capsules, age southern Texas to eastern Colorado, northern Kansas, and western _ all), and Louisiana. Linum alatum (Small) Winkler (L. aun sinignonle ce ig 2n = 30, with unique glandular-toothed sepals with scarious margins, largely confined to southern Texas an P collection labeled “s.w. La.” oer opment we area in Louisiana is L. imbricatum al.) : ex Torrey & Gray), 20 = d glandless sepals and sparsely 30, with fringe ili eaves. bet ; “The sat neo-mesric ant complex is similar sa soe cea complex, but staminodia are absent, and the base . se uthwestern United 13. None of the three species in this eee = ern Mexico approaches ou . : aor apres group, section ie AO agen Griseb., includes only Linum catharticum Ls in - ute flowered” group, annual has numerous features in common with sig i: that it is not closely but the observations of Rogers and Sharsmith in ge ay eats nd related to the flaxes indigenous to North America. nd Newfoundland with European species is established sen att “Pennsylvania. No other scattered locations west to Michigan eo cant : site leaves. S. F. Gray species in the region has small white flowers and oppo? 658 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 placed L. catharticum in his CHAMAELINUM, but the subgeneric category of this taxon is uncertain (cf. Brizicky, Taxon 18: 650. 1969); the name predates Grisebach’s sectional combination. The predominantly European “‘blue-flowered” group, section LINUM (Eulinum Griseb.), with medium to large flowers, linear to capitate stigmas, glandless sepals, long pedicels, all leaves alternate, and stipular glands absent, has only one or two species indigenous to the New World. Wide- spread in western North America, L. Lewisii Pursh, 2n = 18, is sometimes cultivated in the East and has been reported from Cave Mountain, West Virginia (Allard). Plants with sepals less than 5 mm. long are sometimes segregated as L. pratense (J. B. S. Norton) Small, 2x = 18. This variable complex has been considered conspecific with the European L. perenne L., 2n = 18, 36, the only consistent difference between them being that L, Lewisii, like all the New World species, is homostylous, while L. perenne is heterostylous. The latter species is sometimes encountered as a garden escape. Other Old World species introduced to the eastern United States are the red- or pink-flowered L. grandiflorum Desf., 2n = 16, a native of Algeria, and L. usitatissimum L. (L. humile Mill.), 2n = 30, flax. One of the oldest cultivated plants, flax is grown both for the fibers from which linen is made and for the seeds which yield linseed oil. Section Ciiococca (Bab.) Planchon contains only the South American L. selaginoides Lam., and is perhaps best regarded as a distinct genus. The Old World sections, SyLtInum Griseb. and DasyLINUM (Planchon) Juz., do not occur in North America. Rogers (1969) suggests a monophyletic origin for the North American yellow-flowered flaxes with the L. Schiedeanum complex, which includes the most primitive members of the group, giving rise to three lines — the L. virginianum complex, the L. neo-mexicanum complex, and the L. sul- catum and L. rigidum complexes. East-central Mexico, the center of distribution for the L. Schiedeanum complex, is postulated as the region of establishment and diversification of the North American yellow-flowered group. Rogers points out that some South African species are very similar to L. Schiedeanum, and preliminary evidence (personal communication) indicates similarities between the South American species and certain Mexican and African species. —. 2 — — in North American yellow- ied Wik hae ogers) urther emphasized the distinctness SoG. Koutras a — va possesses the a-carotenoids leutin and f-carotenoid violaxanthin Th ae st Leal camden car ie supports the presumed welatibtalits nar a ry pease ae virginianum complexes. et ds = se epecah key rateg to North America are homostylous, while nile hee ak eterostylous. In some heterostylous species, ee , e filaments and styles are of different lengths while in others, such as L. gr, difl é 8 , grandiflorum, only the styles vary in length. / 1971] ROBERTSON, LINACEAE 659 Charles Darwin first recognized that heterostyly is an out-breeding mech- anism since the long-styled and short-styled forms are mostly self-sterile but cross-fertile; homostyled flowers are largely self-fertile. (See Baker’s interesting paper for a misunderstanding between Darwin and Thomas Meehan, but also involving Asa Gray, about heterostyly and self-compat- ibility in L. perenne and L. Lewisit.) The pollen morphology of Linum species is diverse. Aperture types are 3-, 4- or 6-colpate, pantocolpate and pantoporate. The tricolpate grains have been classified on the basis of aperture shape and size and the nature and texture of the nexine covering the colpi. The exine is thin, undulate or thick and bears surface processes that vary in size and shape. These processes may be monomorphic, dimorphic or polymorphic; in a few species the processes are united to form a tegillum which can bear suprategillar processes. The sexine shows a trend from a granular composition to the formation of distinct bacula. Ockendon found in the L. perenne group a correlation between aperture number and ploidy level, and between the morphology of the exine processes and the stylar condition of the flower. Yermanos & Gill (1967) report that in eight species the pollen was 3-colpate in diploids and 4-colpate in their autotetraploids. t 0 some extent pollen morphological trends follow the subgeneric classification of Linum. The thick-walled polyporate pollen of the L. rigidum and L. sulcatum complexes is evidently the most specialized in the genus. The North American and European species of Linum are in ania cytologically, and definite base numbers have been scab in ee sections and species complexes. Polyploidy and aneup 32 er : - Linum Schiedeanum complex, 18, 31, in most groups. Base numbers are: ; et ian Fah 34: L. virginianum complex, 18; L. neo-mexicanum — 8: bition catum and L. rigidum complexes, 15; section as Gidlactiies Linum, 9, 15; section Linastrum (Old World spe er” aa ich SYLLINUM, 14; section DASYLINUM, 6, 8; and L Or a ueikinted New Zealand, reported as having an = 84. Numerous older u _ The hybridization studies of and doubtful counts have been excluded Ns a a cicliweal Gill, Gill & Yermanos, and Mosquin & Hay si ial tes rearrangements, particularly translocations, have played a major role the evolution of Linum species. — Hybridization in nature is evl aio . been Peels irs ‘ial L. virginianum complexes. a jonny i produced in the L. rigidum complex and in both the d P f the L. perenne tatissimum groups of section Linum? Autotetraplo! ou hats Haploid group (2n = 36) can hybridize with L. . . aes plants of the latter species have also been Sani mostly Old World reports of unsuccessful hybridization attempts inv species. dently rare, but has been reported in the : : tween L. Race as obtained a hybrid be = - es os European L. perenne group, — 3 rson D. J. Ockendon (pe tylous membe Lewisii and L, Leonii, a homos did not flower. 660 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 One Linum hybrid cross is often used as an example of the requirement that the hybrid embryo, endosperm, and seed coat be mutually compatible. In the L. perenne 9 x L. austriacum 8 cross, the hybrid embryo develops normally but cannot sprout through the seed coat. If removed, it grows into a vigorous and fertile F; hybrid. In the reciprocal cross, the embryo is inhibited and will not develop unless dissected from the seed and placed in a nutrient solution. REFERENCES Most of the voluminous literature on the agricultural aspects of Linum usitatissimum has been omitted. Under family references see BOLKHOVSKIKH et al.; Davis; ERDTMAN, 1964, 1966, 1969; HEGNAUER; HUTCHINSON; KNUTH; LEINFELLNER; NARAYANA; PLANCHON; SAaD, 1961; SHARSMITH; SMALL; and WINKLER. AxpaBa, V. C. The structure and development of the cell wall in plants. I. Bast fibers of Boehmeria and Linum. Am. Jour. Bot. 14: 16-24. pls. 3-6. 1927. [L. usitatissimum, 21, pl. 5. Hee ALLARD, H. A. Linum Lewisii and Bouteloua curtipendula in West Virginia. Castanea 3: 13, 14. 1938. [L. Lewisii reported from Cave Mountain, Pendleton County. ] Baker, H. G. Charles Darwin and the perennial flax—a controversy and its implications. Huntia 2: 141-161. 1965. | Misunderstanding between Darwin and Thomas Meehan, and involving Asa Gray, caused by the confusion of L. perenne and L. Lewisii and their different floral biologies; comments on the personalities and allegiances of the three men. ] : Bart, G., & M. B. E. Gopwarp. Influence of chromosome size on the radio- sensitivity of Linum species. Canad. Jour. Genet. Cytol, 11: 799-802. 1969. [11 spp.] —& lescens, and L. umbers. Inheritance of seed and 6 » W., & A. StAutin. Handbuch der Samenkunde. Introd. + 656 pp. 358. 6 spp. of Linum, 356- Byrne, J. M., & C. HeImscu. The root 1011-1019. 1968. [10 spp. of sects. of cortical initials; three North Am apex of Linum. Am. Jour. Bot. 55: Linum and SyLLinum have two tiers erican yellow-flowered spp. have only one. Carton, A. Etudes phyllotaxiques sur 55 CHITTENDEN, R. J. Cytoplasmic i 1927. [When gene for “ male-sterile plants result. | Dane, K. D. Conséquences de mutilations expérimentales sur la croissance, lanatomie, l’hydratation et la floraison de quelques espéces de lin. Bull. Soc. Bot. Suisse 69: 346-467. 1959. [L. austriacum, L. angustifolium, re usitatissimum. | DARWIN, C. On the existence of two forms, and on their reciprocal sexual relation, in several species of the genus Linum. Jour. Linn. Soc. Bot. 7: 1971] ROBERTSON, LINACEAE 661 69-83. 1864. [Heterostyly and self-incompatibility in L. grandiflorum, L. perenne, L. flavum. Davis, P. H. Materials for a Flora of Turkey II: Linum Linn. Notes Bot. Gard. Edinburgh 22: 135-161. pls. 6-8. 1957. [Includes discussion of sects. SyLLINUM and DasyLInvuM. | _ Linaceae. Fl. Turkey 2: 425-450. 1967. [Includes 38 indigenous spp. in 4 sects of Linum. |] Dittman, A. C. Classification of flax varieties, 1946. U.S. Dep. Agr. Tech. Bull. 1054. 1953.* Dopce, C. R. The present status of flax culture in the United States. Yearb. U.S. Dep. Agr. 1897: 471-486. 1898. [Includes historical data and discussion of Linum spp. in cultivation. ] Dorasami, L. S., & D. M. GOPINATH. An embryological study of Linum mysorense Hyene. Proc. Indian Acad. Sci. B. 22: 6-9. 1945. DostAL, R. Correlative curvatures of the Linum hypocotyls and growth regu- lators. Beitr. Biol. Pflanzen 45: 257-270. 1968. [L. usitatissimum, refer- ences to earlier papers. | j : Durrant, A. The environmental induction of heritable change in Linum. Heredity 17: 27-61. 1 pl. 1962. [Heritable changes induced in L. usitatessimum by fertilizer treatments behave in crosses and grafts like orthodox genetic types; also see Evans et al.| . Eastman, W. The history of the linseed oil industry in the United States. 277 . Minneapolis. 1968. f Bee x. Vaacala differentiation in the vegetative shoot of Linum. I. saa procambium. Am. Jour. Bot. 29: 738-747. 1942; II. The first aes xylem. bid. 30: 248-255. 1943; III. The ae of See lca 579-586. [L. perenne; L. usitatissimum included in 12). . Plant i i and ed, xx + 767 pp. New York, London, & a 1965. [Linum used to illustrate anatomical features; numerous drawings Evans, G: Me _ Durrant, & H. REES. Associated nuclear — - sea induction of flax genotrophs. Nature 212: 697-699. -_ [ - oe gests that DNA changes are associated with the induction differences; also see DURRANT Eyre, J. V., & G. SmiTH. Some no flax. Jour. Genet. 5: ay oh le to hybridize eight Old Worle 477 : ieginia. Rhodora Ferwatp, M. L. Midsummer vascular pit rae As. frond 427-431. 37: 378-413. pls. 384-393 ; peng ec L floridanum, ‘a intercursum.] pl. 396; L. medium, L. neg pedtoae + 344 pp- London & New York. Free, J. B. Insect pollination of crops. XI 1970. [L. usitatissimum, 179-182 tes on the Linaceae. The cross pollination of 1916. [Attempts made, mostly unsuccessful, ; ic signi floral pigments sien ys re “heen Rot et oe ar Co aicen a = mt 1 3 ont er 1 al prone oo a * cific hybrids; mitosis ag agreed , nega nalvciesde fatty acid co Cf. GILL 7 hereon Cytogenetic studies in the genus Linum. I. Hybrids 662 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 among taxa with 15 as the haploid chromosome number. Crop Sci. Madison 7: 623-627. 1967; II. Hybrids among taxa with nine as the haploid chromo- some number. /bid. 627-631. ast, Girotamti, G. Relation between phyllotaxis and primary vascular organization in Linum. Am. Jour. Bot. 40: 618-625. 1953. [L. usitatissimum ‘Punjab’.] . Leaf histogenesis in Linum usitatissimum. Ibid. 41: 264-273. 1954. GrisepacH, A. H. R. Lineae. Spicillegium florae rumelicae et bithynicae 1: 115-119. 1843. [Linum sections EvuLINuUM, SYLLINUM, CATHARTOLINUM Harris, B. D. Chromosome numbers and evolution in North American species of Linum. Am. Jour. Bot, 55: 1197-1204. 1968. [36 taxa. HyeLMovist, H. The flax weeds and the origin of cultivated flax. Bot. Not. 1950: 257-298. 1950. [Mimics of flax, found mostly in flax fields. | , Hosuscu, L. Untersuchungen iiber die Kreuzungsschwierigkeiten bei den Bas- tarden zwischen Linum austriacum und L. perenne und deren Nachfol- gegenerationen. Zeitschr. Indukt. Abst. Vererbungslehre 67: 389-434. 1934. HorrMann, W. Lein, Linum usitatissimum L. In: T. Roemer & W. Ruporr. Handbuch der Pflanzenziichtung, ed. 2. 5: 264-366. 1961. | Extensive review of literature. | Ho.pen, D. J. Factors in dehiscence of the flax fruit. Bot. Gaz. 117: 294-309. 1956. [Indehi scent, semidehiscent, and completely dehiscent varieties of ts usitatissimum. | Juzerczux, S. V. Linaceae. In: V. A. Komaroyv. Fl. URSS (In Russian.) 14: 84-146. 1949. [See ix-x for classification scheme. ‘ KANTor, T. S. Comparative-embryological research of some cultural and wild species of flax. (In Russian; English sum Sada 53: 276-307. 1933. | Haploid-diploid twins. | Kuanna, A. N., & K. B. SINGH. Behavior of pollen-tube growth in some inter- specific crosses in the genus Linum. Agra Univ. Jour. Res. Sci. 5(1): 169- 178. pl. 3. 1956. [Crosses between L. usitatissimum and L. grandiflorum, L. marginale, L, angustifolium, and L. africanum. | Larpacu, F. Ectogenesis in plants: Methods and genetic possibilities of prop- agating embryos otherwise dying in the seed. Jour. Hered. 20: 200-208. 1929. (Cf. Zeitschr. Bot. 17: 417-459. 1925.) [Hybrids involving six Old World Linum spp. | Luspock, J. A contribution to our knowledge of seedlings. Vol. 1. vii + 608 pp. London & New York. 1892. te. campanulatum, L. monogynum, L. perenne, Martin, A. C. The comparative internal morphology of seeds. Am. Midl. Nat. 36: 513-660. 1946. [Linum, 624, 625.] € shoot apex of Linum usitatissimum L. and origin and ud primordia and foliar traces. Pakistan Jour. Sci. Res. 7: 159-164. 1955.* YLEY. Evolutionary relationships of the southern Linaceae). Canad. Jour. Bot. 45: Nestiex, H. Beitrige zur systematischen Kennan der Gattung Linum, Bei. Bot. Centralbl. 50(2): 1971] ROBERTSON, LINACEAE 663 CAIHARTOLINUM, EvLinum, CitococcA, LINASTRUM, SYLLINUM; also Hesperolinum. | NieuwLanp, J. A. Notes on priority of plant names. Am. Midl. Nat. 3: 150-158. 1913. [Incorrectly takes Nezera Raf. as older than Cathartolinum Reichen- bach. | OckENDON, D. J. Biosystematic studies in the Linum perenne group. New Phytol. 67: 787-813. 2 pls. 1968. [Biosystematics of a European complex; scanning electron microscope photomicrographs of pollen. | Taxonomy of the Linum perenne group in Europe. Watsonia 8: 205- 235. 1971. Ossporne, W. P., & W. H. Lewis. Chromosome numbers of Linum from the southern United States and Mexico. Sida 1: 63-68. 1962. [14 taxa. Pattwat, G. $. The development of stomata in Linum usitatissimum L. Curr. Sci. Bangalore 30: 269-271. 1961. [ Syndetocheilic type. | Panpey, K. K. Studies in autotetraploids of linseed (Linum usitatissimum L.). I. Growth rate. Lloydia 19: 120-128. 1956; II. Morphology and cytogenetics. Ibid. 245-268. Piessers, A. G. The variation in fatty acid composition of the seed of Linum species. Canad. Jour. Genet. Cytol. 8: 328-335. 1966. [Also chromosome numbers for 7 spp. . Ray, C. Cytological studies on the flax genus, Linum. Am. Jour. Bot. 3i: 241-248. 1944. [Chromosomes of 30 spp.; 28 varieties of L, usitatisst- mum. | Rickert, H. W. Wildflowers of the United States. Vol. 2. The ee States. Part 1. x + 322 pp. pls. 1-116. New York. 1966. [Linum, pee 267, pl. 96; 5 spp. illustrated in color.]; Vol. 3. Texas. Part : > : Is. 1-81. New York. 1969. [Linum, 176-178, pls. 51, 52; 7 SPP- illustrated in color, including L. rigidum and L. imbricatum.| ie Rocasn, A. R. Interspecific hybrids in Linum. (In Russian; English su . Vestn. Gibridiz. 2: 84-88. 1941.* : Z Rocers, C. M. Yellow flowered species of Linum In eastern eS ae Brittonia 15: 97-122. 1963a. [12 Spp.; taxonemmt ee trations, maps, exsiccatae. | ‘ - Studies in Linum: L. imbricatum i L. 50-55. 1963b. [Maintains both spp. ; 4. 66: 278, _ Linum coahuilense, a new species from northern Meri i vinkler : 279. 1964a. [Later placed in synonymy. ars ree 1: 328-336. 1964b. _ Yellow-flowered Linum (Linaceae) in Texas. “canals 14 spp.; key, maps.] : ; North ie ae species of Linum 1n Central ong ee, anal America. Brittonia 20: 107-135. 1968a. [25 SPP-; key , and exsiccatae. | ce in Florida. Z ne ee of Linum dort ae cack treated Sida 3: 209, 210. 1968b. [Florida popu” as L. Carteri, with 2 vars.] é Rhodo . The Linum bahamense complex. #! : Linu _ Relationships of the North ae ae — 08 Torrey Bot. Club 96: 176-190. 1969. [Su spp. | - Linaceae. Jn: D. S. CorREL udsonioides. Rhodora 65: 70: 439-441. 1968c. sy m (fax). Bull. yellow-flowered 1 & M. C. JOHNSTON, Manual of the plants 664 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 of Texas. xv + 1881 pp. Frontisp. + 1 map. Renner, Texas. 1970a. [ Linum, 897-901. | . Some observations on the history of botanical collecting in eastern North America. Castanea 35: 78, 79. 1970b. [Number of collections and collectors of Linum from 1820 to 1955.] ———. Changing abundance of two species of Linum in eastern North America. Mich. Bot. 10: 113-116. 1971. [Based on relative number of herbarium collections; L. virginianum and L. medium var. texanum. | B. D. Harris. Some new chromosome numbers in Linum. Madrono 18: 179, 180. 1966. [5 taxa.] & . Pollen exine deposition: A clue to its control. Am. Jour. Bot. 56: 1209-1211. 1969. [In triploid hybrids the spore nucleus evidently does not participate in exine deposition and pattern determination SaaD, S. I. Phylogenetic development in the apertural mechanisms of Linum pollen grains. Pollen Spores 3: 33-43. 1961. [Old and New World spp.; drawings and photomicrographs. | Sizova, M. A. Study of the process of fertilization in interspecific crosses of flax. (In Russian.) Trudy Prikl. Bot. Genet. Selekts. Sborn. Puskinskikh Lab. No. 1: 303-311. 1958.* Soto, O. The photoperiodic influence on growth and fiber development of flax plants. (In Japanese with English summary.) Formosan Agr. Rev. 37: 144-152. 1941.* SOUEGES, R. Embryogénie des Linacées. Développement de l’embryon chez le Linum catharticum L. Compt. Rend. Acad. Sci. Paris 178: 1307-1310. 1924. (Cf. Bull. Soc. Bot. France 71: 925-938. 1924. TaMMgs, T. Die Flachsbliite. Rec. Trav. Bot. Néerl. 15: 185-227. 1918. . The genetics of the genus Linum. Bibliogr. Genet. 4: 1-36. 1928. [List of chromosome counts, bibliography. TANNER, F. W. Microbiology of flax retting. Bot. Gaz. 74: 174-185. 1922. [ Clostridium amylobacter. | gira W. A revision of North American Linaceae. Trans. Acad. Sci. St. ouis 5: 7-20. pls. 3, 4. 1887. [Linum only Tscuierscu, B. Zur Bildung der c es ee s yanogenen Glyk A. 157: 358-364. 1966 B ykoside aus Aminosauren. Flora UpHor, J. C. T. Dictionary of economic plants. ed. 2. 591 pp. Lehre. 1968. [Linum spp., 314, 315.] VAUGHAN, J. G. The structure and utilizati i : tion of oil seeds. . New York. 1970. [L. usitatissimum, 140-143. ] oe VazartT, J. Infrastructure de ’ antipodial. Compt. Rend. Acad. Sci. Paris D. 266: 211-213. 6 pls. 1968. : G. Tutt, V. H. Heywoop, et al., eds., sd a [Linum by D. J. Ockenpon & S. M. * “'; WALTERS; 36 spp. of Linum in 4 sections. | xX . . . . g AVIER, K. oe bent asian of electron and light microscopic studies PP. 1967. Ph.D. Thesis, W S iversit Detroit.* [Unpublished Cf. ab bog paged ome gid - Ct. abstr. Am. : : New. World apg. coattian m. Jour. Bot. 53: 634. 1966; Old and & C. M. Rocers Pollen m . 0 Rhodora 65: 137-145. 1963. [Ma photomicrographs. ] tphology as a taxonomic tool in Linum. inly North American yellow-flowered spp.; 1971] ROBERTSON, LINACEAE 665 Yapava, T. P., & H. R. Karta. Taxonomic relationship among some indigenous and exotic species of genus Linum. Jour. Res. 5(2, Suppl.): 67-75. 1968. [Hybridization attempts between 15 spp.; specific determinations ques- tionable. | ome _ Chromosome numbers and study of meiosis in some species and hybrids of Linum. Ibid. 76-78. 1968. [L. usitatissimum, L. corymbi- ferum, L. grandiflorum, L. M uelleri. | Yermanos, D. M. Variability in seed oil composition of 43 Linum species. Jour. Am. Oil Chem. Soc. 43: 546-549. 1966. & K. S. Gut. Induction of polyploidy in Linum species. Crop Sci. Madison 7: 423-427. 1967. [Tetraploids induced in mostly Old World species; morphology, cytology, and fertility studied. | & ——. Cytology of autotetraploids of L. usitatissimum L, and L, angustifolium Huds. and their amphidiploid hybrids. Crop Sci. Madison 9: 249, 250. 1969. [Quadrivalent frequencies in hybrids comparable to those of autotetraploids — evidence that the genomes of these species are closely homologous. | ARNOLD ARBORETUM Harvarp UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 664 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 of Texas. xv + 1881 pp. Frontisp. + 1 map. Renner, Texas. 1970a. [ Linum, 897-901. | . Some observations on the history of botanical collecting in eastern North America. Castanea 35: 78, 79. 1970b. [Number of collections and collectors of Linum from 1820 to 1955.]| ———. Changing abundance of two species of Linum in eastern North America. Mich. Bot. 10: 113-116. 1971. [Based on relative number of herbarium collections; L. virginianum and L. medium var. texanum. | & B. D. Harris. Some new chromosome numbers in Linum. Madrono 18: 179, 180. 1966. [5 taxa. ] & . Pollen exine deposition: A clue to its control. Am. Jour. Bot. 56: 1209-1211. 1969. [In triploid hybrids the spore nucleus evidently does not participate in exine deposition and pattern determination Saab, S. I. Phylogenetic development in the apertural mechanisms of Linum pollen grains. Pollen Spores 3: 33-43. 1961. [Old and New World spp.; drawings and photomicrographs. | Sizova, M. A. Study of the process of fertilization in interspecific crosses of flax. (In Russian.) Trudy Prikl. Bot. Genet. Selekts. Sborn. Puskinskikh Lab. No. 1: 303-311. 1958.* Soto, O. The photoperiodic influence on growth and fiber development of flax plants. (In Japanese with English summary.) Formosan Agr. Rev. 37: 144-152. 1941.* SOUEGES, R. Embryogénie des Linacées. Développement de l’embryon chez le Linum catharticum L. Compt. Rend. Acad. Sci. Paris 178: 1307-1310. 1924. (Cf. Bull. Soc. Bot. France 71: 925-938. 1924. Tammes, T. Die Flachsbliite. Rec. Trav. Bot. Néerl. 15: 185-227. 1918. ———. The genetics of the genus Linum. Bibliogr. Genet. 4: 1-36. 1928. [List of chromosome counts, bibliography. TANNER, F. W. Microbiology of flax retting. Bot. Gaz. 74: 174-185. 1922. [Clostridium amylobacter.] TRELEASE, big A revision of North American Linaceae. Trans. Acad. Sci. St. Louis 5: 7-20. pls. 3,4. 1887. [Linum only. | ame yeti der cyanogenen Glykoside aus Aminosduren. Flora Urnor, J. C. T. Dictionary of economic pl i ants. ed. 2. k ; 8. [Linum spp., 314, 315.] P - 591 pp. Lehre. 196 Vaucuan, J. G. The structure and utilizati i ’ zation of oil seeds. . New York. 1970. [L. usitatissimum, 140-143.] tie taal Vazarr, J. Infrastructure de |’ ovule du lin, Linum usitatissimum L. mplexe Wa ek pe ee Acad. Sci. Paris D. 266: 211-213. 6 petit Fl. Eine © eon > Inet, G. Turin, V. H. Heywoop, et al., eds., Watters; Radiola by S. M re [Linum by D. J. Ockenpon & S. M. XAVIER, K. S. PRONE 6 re ii ALTERS; 36 spp. of Linum in 4 sections. ] ot En ‘cotta eh — of electron and light microscopic studies Detroit.* incadieiic on st Fe Thesis, Wayne State University, New World spp. examined.] str. Am. Jour. Bot. 53: 634. 1966; Old and C. M. Roce Rhodora 65: a . én morphology as a taxonomic tool in Linum. photomicrographs. ] 63. [Mainly North American yellow-flowered spp; 1971] ROBERTSON, LINACEAE 665 YVapava, T. P., & H. R. Katia. Taxonomic relationship among some indigenous and exotic species of genus Linum. Jour. Res. 5(2, Suppl.): 67-75. 1968. [Hybridization attempts between 15 spp.; specific determinations ques- tionable. | _ Chromosome numbers and study of meiosis in some species and hybrids of Linum. Ibid. 76-78. 1968. [L. usitatissimum, L. corymbi- ferum, L. grandiflorum, L. M uelleri. | Yermanos, D. M. Variability in seed oil composition of 43 Linum species. Jour. Am. Oil Chem. Soc. 43: 546-549. 1966. & K. S. Gitt. Induction of polyploidy in Linum species. Crop Sci. Madison 7: 423-427. 1967. [Tetraploids induced in mostly Old World species; morphology, cytology, and fertility studied. | & ———. Cytology of autotetraploids of L. usitatissimum L. and L. angustifolium Huds. and their amphidiploid hybrids. Crop Sci. Madison 9: 249, 250. 1969. [Quadrivalent frequencies in hybrids comparable to those of autotetraploids — evidence that the genomes of these species are closely homologous. | ARNOLD ARBORETUM HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 666 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 A REVISION OF THE BORAGINACEAE OF WEST PAKISTAN AND KASHMIR * S. M. A. Kazmi ** 31. Caccinia Savi, Cose Botaniche 1. ¢t. 7. 1832. Anisanthera Raf. Fl. Tellur. 3: 80. 1836. Type sPEciEs: C. glauca Savi. Annual (?), biennial, or perennial hispid herbs; leaves alternate; in- florescence terminal; calyx usually divided for 2/3 to 3/4 of its length, never to the base, much enlarged in fruit; corolla tubular to hypocrateri- form, tube narrow, long; faucal appendages well developed, large; stamens inserted adjacent to the bases of the faucal appendages, exserted at the apex of the corolla tube, unequal, 4 stamens very small, one very large; style filiform; nutlets 4, usually 1-2 developed, marginate or immarginate, -++ compressed, smooth. Species about 6, distributed in southwest and central Asia. C. macranthera (Banks & Soland.) Brand, Pflanzenr. IV. 252(Heft 78) : 90. 1921; Riedl in Rechinger, Fl. Iranica 48: 227. 1967. Borago macranthera Banks & Soland. in Russell, Nat. Hist. Aleppo, ed. 2. 2: 246. 1794. Cynoglossum longiflorum Lehm. Asperif. 162. 1818. Anisanthera ciliata Raf. Fl. Tellur. 11: 60. , Caccinia russellii Boiss. Diagn. Pl. Orient. Nov. I. 2(11): 134. 1849. Type: Aleppo, Russell s.n. (Bm). Erect biennial (f. Boissier) to perennial herb. Stems fistulous, fleshy, angular, glabrous, simple or branched above, 20-80 cm. long. Leaves linear, linear-lanceolate, oblong or oblong-lanceolate, obtuse to acute, 7— 23 cm. long, 640 mm. broad, covered on the upper surface and margins more densely than below with short, stiff, curved bristles, arising from large tuberculate bases, lower surface sometimes glabrous the bristles con- fined to the midrib. Inflorescence paniculate to thyrsoid-compound, ter- minal or axillary, short in flower, later elongated in fruit; bracts leaf- like, narrower, and shorter. Pedicels erect, bearing few to many bristles, to 15 mm. long in flower, horizontal to subreflexed, to 25 mm. long in * Concluded from volume S2y.p. 522, a ** tatiana! formerly a Mercer Research Fellow of the Arnold Arboretum of arvar * niversity, 1S presently engaged in studies on the Boraginaceae of North Africa. He will welcome material for i i i : identificatio i of papers concerning the family, n and will appreciate reprints 1971] KAZMI, BORAGINACEAE 667 aia Calyx to 8 mm. long in flower, divided for 3/4 of its length, lobes anceolate, to oblong-lanceolate, acutish, 5—6 mm. long, to 2.5 mm ‘broad at base, margins of the upper half ciliate, those of the lower half setulose midrib prominent with few scattered bristles, calyx enlarged to 35 mm. lon ' in fruit, divided to nearly half its length, lobes triangular, to 17 mm. 78 and 12 mm. broad at base. Corolla tube 9-18 mm. long, lobes linear- oblong, spreading, 5-9 mm. long; faucal appendages puberulous, subor- bicular, + 1mm. long. Stamens exserted, anthers unequal, one + 5 mm. long, its filament 2 mm. long, the other 4 to 2 mm. long with filaments to 1 mm. long; filaments inserted at the throat of the corolla tube, adjacent to the bases of the faucal appendages. Nutlets grayish to grayish brown, ovoid to suborbicular, 5-11 mm. long, dorsally rugulose, irregularly tu- berculate or minutely hirsute, margins dentate, teeth broadly triangular. DISTRIBUTION oF SPECIES: Syria, Iraq, Iran, Armenia, Caucasus, Tur- kestan, Pamir Alaj, Afghanistan, West Pakistan. ts habit, leaf texture, and the Caccinia macranthera is very variable in i two varieties may be distin- position of its inflorescence. The following guished. Var. macranthera. Caccinia macranthera vat. eu-macranthera Brand, Pflanzenr. IV. 252 (Heft 78): 90, 1921 Biennial (teste Boissier). Stems €a. 3 branches divaricate. Leaves oblong-lanceolate, broad. Inflorescence large, paniculate, usually stems. 0 cm. tall, branched near the base, g—11 cm. long, 10-15 mm. developed to the bases of As that of the species. t Pakistan, I could not find Afghanistan and DISTRIBUTION OF VARIETY: In the collections of this species from Wes reported from east any specimen of this variety. It is may be expected in the western parts of our area. Var. crassifolia (Vent.) Brand, Pflanzenr. IV. 252 (Heft 78): 92. 1921; Riedl in Rechinger, F'. [ranica 48: 228. 1967. ifoli 00. Borago crassifolia Vent. Descr. Pl. Jard. Cels. 100. ¢. 100. 18 a Caccina glauca Savi, Cose Botaniche 1, ¢. 1, 1832; Ledeb. Fl. Rossica 3: 17. 1847: Boiss. Fl. Orient. 4: 277. 1875. ii Boi i : 132. 1849. C. celsii Boiss. Diagn. Pl. Orient. Nov. I. 2(11): 132. 18 C. erassijolic (Vent.) ©. ROeiemtaen 37: 647. 1849; M. Pop. Fl. URSS 7 Acta Horti Petrop. C. Koch, f. normalis O. Kuntze, "Iv, 252 (Heft 78): 92. se inearifoli tze, ibid. C. crassifolia (Vent.) C. Koch, f. jinearifolia O. Kuntze, -. C. phon Nb C. Koch, f. persica, f. obtusifolia, and var. echinata M. Pop. Acta Horti Bot. Univ. Jurjev. 12: 240, 241. 1911. 668 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Type: Persia, inter Hamadan et Tehran, Bruguiére et Oliver, s.n. (Pp). Icon.: Savi, l. c. #. 1. 1832, under C. glauca; Vent. 1. c. t. 100. 1800, under Borago crassifolia. Perennial. Stems simple to branched, profusely leafy, 30-90 cm. tall. Leaves very variable, linear, lanceolate or oblong, 7—23 cm. long, 6—40 mm. broad. Inflorescence thyrsoid to paniculate-thyrsoid, terminal, rarely de- veloped on the lower parts of the stems. DISTRIBUTION OF VARIETY: as that of species, excluding the western parts of the area. West Pakistan: Gitcir AGeNcy: Northern Hindukush, Giles s.n. (k). QUETTA Dist.: Spin Karez, Crookshank 152 (x); Quetta, Peddie 318 (xk), Schmid 163 (k), Narai Kotal, R. R. Stewart 3754 (K); Urak, R. R. Stewart 28064 (K, RAW). Reported from: Baluchistan, Stocks s.n. (K?) 32. Anchusa L. Sp. Pl. 133. 1753; Gen. Pl. ed. 5. 64. 1754. Type species: A. officinalis L. (lectotype species). corolla tube in the regular flowers, flowers. Style filiform; stigma capi- subhorizontal, areola large, basal to subventral, with fleshy thick rugulose collar. Species 30-40, centering in the Mediterranean region, but extending through Europe, Africa, and western Asia . KEY To THE SpEctEs a. Biennial to perennial; corolla 13- 15 mm. long, tube straight, limb regular. A. 1. A. italica Retz. Obs. Bot. 1: 12.1779; Ledeb. FI. Rossica 3: 119 1847; Boiss. Fl. Orient. 4: 154. 1875; M. Pop. Fl. URSS 19: 306. 1953; Riedl in Rechinger, Fl. Iranica 48: 233. 1967. A. azurea Mill. Gard. Dict. ed. 8: 9, 1768, nomen confusum. A. paniculata Ait. Hort. Kew. 1: 172. 1789, TYPE: not indicated. Icon.: Bot. Mag. 48: ¢. 2917. 1820; Reichenb, Icon. Fl. Germ. 18: ¢. 106. MCCCVIL. fig. 5-8. 1858, Biennial to perennial herb. Stems usually solitary, branched in the upper 1971] KAZMI, BORAGINACEAE 669 part, 30-60 cm. long, densely covered with white, spreading trichomes of unequal length, to 4 mm. long, arising from tuberculate bases. Basal leaves sessile, reduced in size upwards. Inflorescence terminal or axillary, branched, short, scorpioid in flower, elongated in fruit: bracts linear to linear-lanceolate, to 20 mm. long, 2 mm. broad. Pedicels erect, hairy, 1-3 mm. long in flower, elongated to 10 mm. in fruit. Calyx divided to the base, lobes linear to linear-lanceolate, to 10 mm. long and 0.5-1.5 mm. broad in flower, not enlarged in fruit, densely covered with trichomes, Corolla rose- purple when young later turning to pure deep blue, 13-15 mm. long, limb spreading, 7-8 mm. in diameter; faucal appendages large, ciliate. Anthers included + 2.2 mm. long; filaments + 1.5 mm. long, inserted below the faucal appendages on the corolla tube. Nutlets erect, cylindrical, prom- inently reticulate-rugose, minutely tuberculate between the wrinkles, ca. 4 mm. long. DistripuTIoNn: Central, southern, and southeastern Europe, North Africa, Turkey, Syria, Caucasus, Iran, Afghanistan, Turkestan, West Pak- . istan, and Kashmir. West Pakistan: Hazara Dist.: Dungagali, Kazmi s.n. (pes). Quetta Dist.: Quetta, Duthie 8686 (BM). RAWALPINDI Dist.: Murree, R. R. Stewart s.n. (RAW). Kashmir: Kotli, Hilal Khurd, A. Rashid 26999 (BM). Gusuleac (Bull. Fac. Stunte Cernaute 1: 273, 274. 1927) distinguishes three varieties of Anchusa italica, viz. var. italica, var. macrocarpa seer & Hohen.) Gusuleac, and var. kurdica Gusuleac. The : SENG i ed from West Pakistan and Kashmir belong to the typical bane other two varieties differ from the typical one in having pense “ene or less equal length on the leaf surfaces with the areas pig ane: in of the trichomes densely pubescent. Var. macrocarpa as en its dis- Turkish kurdistan, Caucasus, and Iran; whereas var. Aurdica tribution in Cyprus, Syria, Palestine, Iran, and Iraq. 2. A. ovata Lehm. Asperif. 1: 222. 1818; Ried! in Rechinger, Flora Iranica 48: 237. 1967. 7 } 7 1753: Lycopsis orientalis L. Sp. Pl. 139. L. micrantha Ledeb. in Pand. Beitr. 65. prviart Ps L. taurica Stev. Bull. Soc. Nat. Moscou 40: 5 fe ale Anchusa orientalis (L.) Reichenb. Icon. Fl. Germ. 40: : i Petrop. 10: 216. Lycopsis arvensis L. subsp. orientalis O. Kuntze, Acta Horti Petrop 1887. Dds Type: Habitat in Caria, Lehmann, 5.1. ay 1858, under A. orientalis. Icon.: Reichenb. Icon. 18: t. 109. fig. es ‘ anched, to 50 cm. Annual erect herb. Stems solitary to few, simple or br: ? 670 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 long, densely covered with spreading white trichomes of unequal length, to 2 mm. long, arising from tuberculate bases. Basal leaves petiolate, oblong-ovate, roundish at apex, entire to broadly and distantly dentate at margins, attenuated towards the base, including petioles to 12 cm. long 2 mm. broad, covered on both surfaces with patent or subappressed trichomes with tuberculate bases; cauline leaves subsessile to sessile, sometimes decurrent, lanceolate to broadly ovate, obtuse to subacute, re- duced in size upwards. Inflorescence foliate, terminal, short in flower, elongated in fruit. Pedicels slender, hairy, erect, very short in flower, later elongated to 12 mm. in fruit. Calyx divided to the base, lobes linear- lanceolate, acute, densely covered with stiff, long, spreading trichomes with tuberculate bases, 4-5 mm. long, 0.1-1 mm. broad in flower, elon- gated to 8(-10) mm., sometimes to 2.5 mm. broad at base in fruit. Co- rolla white to blue, infundibuliform, equalling or slightly exceeding the calyx, 5—6 mm. long, tube geniculate, limb shorter than the tube, to 2 mm. long, irregular, divided 1/2 to 2/3 of its length, lobes unequal; faucal ap- pendages small, papillose. Anthers ovate, obtuse, ca. 0.8 mm. long; fila- ments very short, inserted 1.5-2 mm. above the corolla base. Nutlets transversely ovate, beaked, dorsally carinate, reticulate-rugose, minutely tuberculate between the wrinkles, areola subventral, large, margins thick, elevated, entire to slightly rugulose. DistripuTion: Bulgaria, Rumania, Turkey, Armenia, Caucasus, Trans- caucasus, Iraq, Iran, Afghanistan, West Pakistan, Kashmir, northwestern India, northeast Africa. Gitcrr AcENcy: Baltistan, Indus River Valley, Webster & Nasir 5888 (cu); ; Rupal to Gorikot, 2250-2350 m., R. R 30458 (w); North Waziristan, Khaisota, J. L. Stewart $8.47 Kashmir: Ladak, Dras, 3000 m. R. R vile Koelz 2678b (cu, us); Pitug, Koelz 6342 (GH, Us); Pirpanjal Range, Nil Nag, ; Kishanganga Valley, near Bhagtaur, 2200 nagar, Nishat Bagh, R. R. Stewart 23070a (GH); Baramulla, 1500 m., R. R. Stewart 13901B (GH). Miscellaneous: Tibet occ., 12000-14000 ped., Herb, Ind. Or. Hook. f. & Thoms., Thomson s.n. (GH). 33. Nonea Medicus, Philos. Bot, 1: 31. 1789. TYPE SPECIES: not indicated. Rea biennial or perennial hispid or villous herbs; leaves alternate. orescence lax, racemose, elongate in fruit, foliate. Calyx 5-fid, short- 1971] KAZMI, BORAGINACEAE 671 ly or for 1/2 its length, never divided to the base, enlarged in fruit, tube enclosing the nutlets; corolla campanulate-infundibuliform to subcylindric; faucal appendages distinct, fimbriate, or reduced to fimbriae or sometimes to pilosity only in throat, limb actinomorphic to slightly zygomorphic, lobes 5, imbricate in bud, short to long and spreading; stamens 5, included, anthers oblong, obtuse; ovary deeply 5-lobed, style filiform, stigma short or obscurely bilobed. Nutlets 4, erect with basal areola or more or less transversely ovate with subventral areola, usually rugose, areola of thick, elevated collar, entire to denticulate at margins. Species about 25, chiefly Mediterranean, distributed in Europe, Asia, and North Africa. Key TO THE SPECIES Corolla white. oie b. Corolla 8-9 mm. long; nutlets 4 mm. long, areola 1-1.5 mm. in diameter, collar 0.5-0.7 mm. high, denticulate at the margins. .. 3. N. edgeworthii. b. Corolla 6-7 mm. long; nutlets 2.5-3 mm. long, areola 1.5-2 mm. in di- .5—2 mm. high, not denticulate at the margins. ....... ve ameter, collar 1.5—2 gh, MeN i ~ i) ©) ° a = 1X) @ .<*) 3. . oO ips = Oo o a 3 as =f z F . i ee pees nage short, distinct to indistinct, regular to zygomorphic; areola 1-1.5 a in poe ‘ ¢/ : ie Plants biennial ty perennial; corolla 10-20 mm. long, fimab site "3 ally equal half the length of the tube, never een ee Ae 2 mm. in diameter: .... cc ee ee ee ee A. 1. N. caspica (Willd.) G. Don, Gen. Syst. 4: 336. 1838; am in Rechinger, Fl. Iranica 48: 250. 1967. Onosma caspica Willd. Sp. Pl. 1(2): a eee Lycopsis caspica Lehm. Asperif. 2: 256. snopnee Anchusa picta M. Bieb. FI. — : ae ik Lycopsis picta (M. it re - CA “Mey Index Sem. Hort. Petrop. 43. Nonnea picta (M. Bieb.) Fisch. » AL . : 152. 1860. ona diffusa Boiss. & Bhuse, Mem. aap se site eT 6. Nonnea nigericans “nigrescens” Auct. Fl. . Tyre: In arenosis versus Mare Caspicum, peeing lit sually + 3 : : ‘ Annual herb. Stems striate, usually many, some MT cm, long, f t, simple, ? umbent, rarely erect, s : homes, to 0.5 ed sce minute, thin, crisped, Wiig USES deawly nt ith or without minute tuberculate ases, (nae ott waned yee stout, pungent, spreading trichomes, arising S ’ 672 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 inent tuberculate bases, 1-2 mm. long. Basal leaves petiolate, linear-lan- ceolate, oblong-lanceolate or oblong, entire to variously dentate, usually acute to rarely obtuse, including petioles 2-5(-6) cm. long, 3-10 mm. broad, covered densely with trichomes more or less like those on the stem, upper surface with more and longer trichomes than the lower surface. In- florescence short in flower, later elongated. Pedicels pubescent, very short in flower, elongated to 4 mm. in fruit. Calyx divided for 1 /3 to 1/2 its length, 5-7 mm. long in flower, lobes linear-lanceolate, acute, densely pubescent, enlarged and inflated in fruit, to 12 mm. long. Corolla red when young, turning to dark purple at maturity, 7~12 mm. long, limb distinct to indistinct, usually zygomorphic. Nutlets transversely ovate, 3-4 mm. long, dorsally slightly rugulose, covered with minute, slender trichomes, denser near the areola, areola longitudinally plicate, ridges ra- diating towards the body of the nutlet, 1-1.5 mm. in diameter, ca. 1.5 mm. high. DISTRIBUTION OF SPECIES: European Russia, Caucasus, Iraq, Iran, Tur- kestan, Afghanistan, and West Pakistan. The species is very variable in the form of cauline leaves and in the form and color of the corolla. The following subspecies may be disting- uished. KEY TO THE SUBSPECIES crisped-dentate. b. Corolla dark blue, cylindrical, hardly surpassing the calyx, limb not dis- rere ate ee pe la. subsp. caspica. b. Corolla red when young, later blue, infundibuliform, usually slightly zygomorphic, limb distinct. ............0 1b. subsp. zygomorpha. a, Cauline leaves ovate-lanceolate to oblong, entire to flabellate-dentate. : violet, limb short, indistinct, leaves entire to oQ © } ° ad oy Oo aes 5 a = 5 ie) jek) Bes t ie) Qu co ° oO Q g =) = rt) 5 e =] o.: —_ ie wn =) iz) i > = : S =" > la. Subsp. caspica. Low growing, 5—10(-15) cm. tall, leaves linear -lanceolate, entire to + lacerate, corolla short, subcylindric, limb minute, not evidently distinct. DISTRIBUTION OF SUBSPECIES: As that of the species. West Pakistan: CHITRAL Srare: Chitral Mastuj track, Buni : 2465 (pm). Lawore Disr.: Lahore, Schlagintwei = ei ely 2 rom Quetta to Ziarat, Kazmi 1578 (PEs). RAWALP : i Farm, Rawalpindi, Y. Nasir 3356 (GH), ee 1971] KAZMI, BORAGINACEAE 673 Ib. Subsp. melanocarpa (Boiss.) H. Riedl, Oesterr. Bot. Zeitschr, 110: 527. 1963, in Rechinger, Fl. Iranica 48: 251. 1967. Nonnea melanocarpa Boiss. Diagn. Pl. Orient. Nov. I. 2(11): 96. 1849, Nonnea szowitsiana Stev. Bull. Soc. Nat. Moscou 24(1): 574. 1851, Nonnea melanecarpa f. macra Boiss. ex Lipsky, Acta Horti Petrop. 26: 471. 910, Type: Circa Hierosolyman, Boissier s.n. (c). Ascendent, suberect or rarely procumbent herb, 12-15(-—20) cm. tall; leaves oblong to oblong-lanceolate, entire to lacerate, corolla small, sub- campanulate, limb much shorter than the tube. DISTRIBUTION OF sUBSPECIES: Palestine, Iraq, Iran, Caucasus, Afghanis- tan, Pamir Alaj, West Pakistan. West Pakistan: QueTTA Dist.: Sariab, s. Quetta, 1700 m., Rechinger 28815 (w). lc. Subsp. zygomorpha H. Riedl, Oesterr. Bot. Zeitschr. 110: 527. 1963, in Rechinger, Fl. Iranica 48: 252. 1967. Type: Ad confines Mesopotamiae et Arabiae borealis inter Meskene et Der-es Sor in tergo El Hilu inter vicos Sabcha et Tibne, Handel-Maz- zetti 548 (w). Procumbent to ascendent herb, 10-20 cm. tall; leaves ee a usually deeply lacerate; corolla rosy red turning violet, br mm. ope limb distinct, nearly half as long as the tube, usually evidently zygo phic. . DISTRIBUTION OF sUBSPECTIES: Iraq, Iran, Afghanistan, West Pakistan. akistan: Karat Dist.: Kalat, R. R. Stewart 600A (raw). Lawore Dist.: Lahore, April 1938, H. Choudhuri s.n. (GH). ld. Subsp. schmidii H. Riedl, Oesterr. Bot. Zeitschr. 110: 529. 1963, in Rechinger, Fl. Iranica 48: 252. 1967. Tyre: Persia: Prope Qazvin, 1200 m., Schmid 5024 (c-holotype). Icon.: Riedl 1. c. fig. 8. 1963. i . Plants 15-25 cm. tall; stems deflexed to eperte pele eh oc, cerate; trichomes lax and a eee . hae ssc I] -red when young, : ‘ omorphic. half 7 pita of the corolla tube or even shorter, never zygomorp DIsTRIBUTION oF suBSPEcIES: Afghanistan, Tran, : has been reported from eastern Afghanistan . ed in West Pakistan. Turkestan. The subspecies schmidti and southern Iran and may be expect 674 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 2. N. kandaharensis Riedl in Kgie & Rechinger, Biol. Skr. 13(4): 226. 1963, in Rechinger, Fl. Iranica 48: 248. 1967, Type: Afghanistan: Kandahar, Shin Ghasi, zwischen Kandahar und Girischk am Hilmend, Bit. blau, 30. IV. 1935, Kerstan 176 (1at-holo- type). Annual low growing herb. Stems many, procumbent to ascendent, sim- ple to branched, 5-15 cm. long, covered with long, spreading trichomes intermixed with short ones. Leaves 20-50 mm. long, 3-6 mm. broad; basal leaves petiolate, linear-lanceolate, acute to obtusish, attenuate to- wards the base; cauline leaves semiamplexicaul, covered on both surfaces with trichomes more or less similar to those on the stems. Inflorescence short and dense in flower, later elongated in fruit, foliate, these leaves similar to the upper cauline leaves 1 1 /2—2 times longer than the calyces. Pedicels 2 mm. long. Calyx divided for half of its length, lobes lanceolate, 4-5 mm. long in flower, densely pubescent, elongated to 7-8 mm., in- flated in fruit. Corolla purple-violet, blue when dried, cylindrical-cam- panulate, 6-8 mm. long, tube slightly dilated at throat, with long exserted fimbria, limb 1/3-1/2 the length of the tube, + 3 mm. long. Nutlets transversely ovate, reticulate-rugose, 2-3 mm. long. DISTRIBUTION: Afghanistan, West Pakistan. West Pakistan: Reported from (Riedl, /. c.): Katar State: 40 km. N. Kalat versus Mastung, 2000 m., Rechinger 28364 (w); QueTtA Dist.: Yaro prope Bostan, 1450 m., Rechinger 28933 (w). I was unable to examine the specimens of NV. kandaharensis cited above. From the description it appears to be closely related to NV. caspica (Willd.) G. Don, from which it is distinguished by having a much longer corolla limb, and the corolla throat bearing long exserted fimbria. ES N. edgeworthii A. DC. in DC. Prodr. 10: 30. 1846; Riedl in Rech- inger, Fl. Iranica 48: 249. 1967. Type: India bor. -occid. ad Malwa, Edgeworth s.n. (K). Icon.: Riedl, Oesterr. Bot. Zeitschr. 110: 531, fig. 9. 1963. Annual herb. Stems striate, usually many, dent to erect, simple to branched, to 35 cm arsely, especially on the trichomes arising from prominent -spathulat id- dle gradually narrowed Sei P ate, obtuse, from the mid ‘ » Sparsely so on the lower surface ng those on the stem; middle cauline leaves ses- 1971] KAZMI, BORAGINACEAE 675 sile, lanceolate, entire to irregularly and distantly dentate, obtuse to sub- acute, attenuate towards the base, to 10 cm, long and 15 mm. broad; upper cauline leaves shorter and narrower, margins bearing long trichomes, acute to acuminate, with slightly dilated bases. Inflorescence short in flower, elongated in fruit. Pedicels very short in flower, pubescent, re- curved, elongated to 3-4 mm. in fruit. Calyx divided for nearly half of its length, 5-6 mm. long in flower, densely pubescent, elongated to 13 mm. in fruit, lobes lanceolate in flower, saccate, triangular, acute, to 5 mm. long in fruit. Corolla white, + infundibuliform, 8-9 mm. long, limb hardly shorter than the tube, indistinctly zygomorphic, lobes short, rounded. Nutlets dark brown, transversely ovate, + 4 mm, long, back convex, slightly rugose, glabrous, with white spots, sometimes sparsely pubescent near the areola, above the areola ventrally keeled, ridge con- tinued on the dorsal surface and shifted to one side, acute at apex, areola I-1.5 mm. in diameter, collar 0.5-0.7 mm. high, longitudinally plicate, plicae radiating towards the body of the nutlet, caruncle small. DistriBuTIoN: West Pakistan, Kashmir, northwest India. West Pakistan: Atrock Dist.: Dhok Pathan, R. R. Stewart s.n. (raw). CHI- TRAL STATE: Chitral village, 1600 m., Stainton 2207 (BM); Lutkor, 2600 m., Bowes Lyon 798 (pm), JueLuM Dist.: Mt. Tilla, Kabir 20430 pula _ RAM AGENCY: Kurram Valley, Afandi 10150 (PES); Parachinar, a roo 28938 (RAW), Harsukh 14437 (x). Kxyper Acency: Khyber rng * a ESHAWAR D1st.: 7 miles from Peshawar, towards Khyber March, 1938, R. R. Stewart s.n. (GH). Kashmir: Reported from Jammu. 4, N. turcomanica M. Pop. Spiske Rast. Herb. Fl. tog pte 1953. Fl. URSS 19: 330. 1953; Riedl in Rechinger, Fl. 249. 1967. Type: Turcomania: In ditione Aschchabad, ope iomenos Wy Naile talem prope pagum Bagir, ad margine agrorum, 29 Apt ye 3575 (1E-holotype, a, w-isotypes). rect to ascendent, slender, simple y with thin, spreading eA : i ite tri- 0.1-0.2 mm. long, intermixed sparsely with some slender, stiff, w chomes with tuberculate bases. attenuate at base, 4-5 mm. long, ee ae ‘cages. ohideg:lanbeolate: 60 lanceolate, shorter, unequally gins, both surfaces covered de upper surface and the mar ae bases. Inflorescence of short or 676 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 fruits. Pedicels very short in flower, pubescent, elongated to 3 mm. in fruit. Calyx in fruits saccate-ovate, densely covered with trichomes simi- lar to those on the leaves, 6-8 mm. long, teeth acute, 1-2(—3) mm. long. Corolla white, narrowly subtubular, 6-7 mm. long, tube equalling the flowering calyx, ca. 4 mm. long, limb hardly dilated, narrowly campanu- late, 2 mm. long, lobes small, obtuse, ca. 1 mm. long. Nutlets brown, 2.5- 3 mm. long, incurved, subhorizontal, ventrally keeled above the areola, ridge continued through the acute apex on the dorsal side, shifting to one side, dorsally convex, slightly and sparsely rugulose, minutely puberulous all over the surface, areola large, 1.5-2 mm. in diameter, with a raised collar, ca. 1 mm. high, entire at the margins, caruncle small, mammiliform. DistriBuTion: Iran, Afghanistan, Turkestan, West Pakistan. West Pakistan: Kurram Acency: Kurram Valley, Parachinar, 1500-1700 m., Rechinger 30928 (w). Nonea turcomanica is known from only one collection from Kurram Valley in West Pakistan. A close relative of this species cannot be sug- gested in our area, except that it is annual and has white flowers like NV. edgeworthii. It differs from N. edgeworthii in leaf shape, length of corolla, and in its areolae. The nutlets are slightly rugose with large high col- lared areolae, entire at their margins, The description given above is based on the isotype in the Arnold Arboretum herbarium. 5. N, pulla (L.) DC. FI. France, ed. 3. 626. 1805; Riedl in Rechinger, FI, Tranica 48: 247, 1967. Lycopsis pulla L. Syst. Veg. ed. 10. 910. 1767. Anchusa pulla (L.) M. Bieb, F. Taur.-Cauc. 1: 125. 1808. Type: “Hab. in Tartaria, Germania” Herb. No. 190.2 (xin). Perennial herb. Stems ascendent to erect, solitary to many, simple to branched, to 45 cm. long, densely covered with spreading to subappressed thin trichomes to 0.5 mm. long, with or without tuberculate bases, inter- mixed with long, pungent, white spreading trichomes, 1-2 mm. long, arising from prominent tuberculate bases. Basal leaves petiolate, lan- ceolate, 1-7 cm. long, 8-15 mm. broad, covered on both surfaces to some extent with trichomes similar to those on the stems; cauline leaves sessile, dilated to subamplexicaul at base; upper cauline leaves shorter R a o a onal 3 wn 3 m. in half its length, lobes nar- gated 10-20 mm. in fruit. » J-4 mm. long, usually puberulous, dorsally er with a thick, longitudinally 1971] KAZMI, BORAGINACEAE 677 plicate collar, ridges radiating towards the body of the nutlet, caruncle large. DISTRIBUTION OF THE SPECIES: Central and eastern Europe, Turkey, Caucasus, Iraq, Iran, Afghanistan (?), West Pakistan. Nonea pulla is very variable in the form and indument of its cauline leaves, and in the length of the corolla. Riedl (/. c.) has classified these variations into four subspecies, viz. subsp. pulla, macrantha, monticola, and rudbarensis. Subspecies pulla is confined to Europe, subspecies mon- ticola is spread over Turkey, Turkish Armenia, Iraq, and Kurdistan, whereas the other two subspecies are distributed in areas from Iraq to northwestern India. Subspecies rudbarensis occurs in West Pakistan. Subsp. rudbarensis Rech. f. Ann. Naturh. Mus. Wien 55: 17. 1947; Ried] in Rechinger, Fl. Iranica 48: 248. 1967. Type: North Persia: Gil: Rudbar, 200-500 m., Bornmiiller 7677 (w- holotype). This subspecies is distinguished from the typical subspecies in having perennial rhizomes; dwarfed, often deflexed stems; narrowly lanceolate basal leaves with rigid trichomes, arising from tuberculate bases, ie mixed on the inflorescence with stipitate glands; and the fruiting calyx with broadly triangular teeth. . istan DISTRIBUTION OF SUBSPECIES: Iraq-Kurdistan, Iran, West Pakistan, western India. r West Pakistan: Arrock D1st.: between Campbellpore and Seon inelcomy 300-400 m., Kazmi 2733 (pes). Lanore Dist. Between pane “Lahore, Kasmi ca. 400 m., 10 miles towards Shekhupura, Kazm1 29 ob erent Dist: Ra 2922 (pes): Muridke, ca. 400 m., Kazmi 2955 (PES). : ince se Hips walpindi, ca, 500 m., R. R. Stewart 1510 (Us). vane pant South Waziris- Ind. Or. Hook. f. & Thoms., Thomson s.n, (cH). Reporte tan, Kaniguram, J. L. Stewart, Blatter, Fernandes. i : ir 1 ent, The specimens cited above agree in the form Hf cea ha ay Fe tan, and in the length and color of the corolla with af gon oe cmc In habit these are not dwarf but erect, to 50 cm. ta stems, and leaves to 11 cm. long. m. & DC. in C. B. Clarke referred these plants to Nonea pulla sensu 1a Hook. f. Fl. Brit. India 4: 169. 1883. : 608. 1963. 34. Decalepidanthus H. Riedl, Oesterr. Bot. Zeitschr. 110: 6 Type species: D. sericophyllus H. Riedl. Very likely a perennial herb. —— tt strongly contracted. Calyx divided to te; in the calyx, lobes longer than pg eee inserte serted slightly above the corolla base. alternate. Inflorescence Corolla tubular, included ] appendages 10, in- pane ae the middle of 678 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 the corolla tube, anthers small, ovate; filaments long, much exserted. Style slightly exserted from the corolla, glabrous, filiform, stigma deeply bilobed. Nutlets 4, free, slightly keeled dorsally, areola basal, fixed on the plane disciform gynobase. One species endemic in West Pakistan. D. sericophyllus H. Riedl, Oesterr. Bot. Zeitschr. 110: 608. 1963, Type: West Pakistan (NW-Frontier Province): Nathia Gali, N. of Rawalpindi, ca. 1800 m., 1907, Sir Harold Deane s.n. (x-holotype). Basal parts of the plant not seen. Stems densely and softly pilose above, herbaceous, internodes to 1 cm. long; middle and cauline leaves sessile to subamplexicaul lanceolate to linear-lanceolate, acute, narrowed towards the base, 5-10 cm. long, 5-8 mm. broad, midrib sunken on the upper surface, raised and prominent below, secondary veins, usually sev- eral, indistinct, running towards the entire margin, the upper surface covered densely with soft, antrorsely appressed trichomes, the lower sur- face with + pallid, silky trichomes; cauline leaves similar, smaller, and axillary, 2.5-3.5 cm. long, 2-4 mm. broad. Inflorescence terminal, ge- minate to ternate, rarely solitary in the axils of the upper cauline leaves, strongly contracted, bearing 5 to 7 flowers, elongate and lax in fruit; bracts lanceolate, 2-3 mm. long, + 0.5 mm. broad. Pedicels indistinct to 1 mm. long in flower, elongated to 1.5(—2) mm. in fruit. Calyx divided to the base, lobes linear, appressed-pilose, 3.54 mm. long, to 0.5 mm. broad, not at all to slightly elongate in fruit. Corolla subtubular, slightly shorter than the calyx, 3.5 mm. long, lobes triangular, acute to acuminate, ; faucal appendages 10, 5 alternating y ovate, 0.7-0.8 mm. long; filaments slender, filiform, 3.5-4 mm. long, exserted, inserted on the middle of the corolla tube. Style filiform, 4.5-5 mm. long, stigma deeply bilobed. Nutlets (only very immature ones seen) dorsally slightly carinate, smooth, attached basally to the plane gynobase DISTRIBUTION: West Pakistan. West Pakistan: Hazara Dist.: I did not see the type or any other authentic material of this species, - Bs Sy genus is based. The description given here is according to Nathiagali, Deane s.n. (xk). 35. Gastrocotyle (Bunge) Benth. & Hook. Gen. Pl. 2: 853. 1876. Anchusa L. Sect. 5. Gastr 851. Etr. 7: 405. 1 ocotyle Bunge, Mém. Acad. Sci. St. Pétersb. Sav. TYPE SPECIES: G. hispida (Forssk.) C. B. Clarke, 1971] KAZMI, BORAGINACEAE 679 Annual or perennial (?) herbs; calyx divided to the base; corolla sub- tubular, faucal appendages small; stamens 5, included, anthers ovate, obtuse; ovary 4-lobed, style short, subcapitate. Nutlets ovoid, erect, incurved, reticulate-rugose, areola ventral with elevated, dentate margins affixed to the carpophore which is half the length of the nutlets. Pakistan and India. G. hispida (Forssk.) C. B. Clarke in Hook. f. Flora Brit. India 4: 168. 1883; Riedl in Rechinger, Fl. Iranica 48: 254. 1967. Anchusa hispida Forssk. Fl. Aegypt.-Arab. 40. 1775; DC. Prodr. 10: 50. 1846; Boiss. Fl. Orient. 4: 158. 1875. Type: In desertis Kahirinis, Forsskdl s.n. (c). Species two, distributed from North Africa and Asia Minor to West d ers axillary, solitary, or in sma a hardly elongated in fruit. Calyx divided to the pee larged in fruit. covered with thick trichomes, ca. 2 mm. long, hardly enlarg ; , 3 mm. long, tube equalling Corolla pale purple to light blue, subtubular, + Hie hidin agtend- the calyx, limb 1 mm. long, divided nearly to the base, lo faculty ‘strongly ing; faucal appendages villous. Nutlets brown, ek ooo oiikr thick, convex, rugulose, minutely tuberculate, areola ventral, large, elevated, dentate collar. ; ets: Distrrsution: North Africa, Sinai, Arabia, ee De ecsits Caspian to Soongaria, Afghanistan, West Pakistan, India. . bster & Nasir West Pakistan: Gitcrr AceNcy: Baltistan, Skardu, i ee mith eg S 5767 (GH) Lawore Dist.: Lahore, April, 1936, wi RAWALPINDI Dist.: near (GH), PesHAwar Dist.: Peshawar, Nath ie ce aamaaaecs Punjab, Herb. Agriculture Farm, Rawalpindi, Nasir 5357 ve ‘ Ind. Or. Hook. f. & Thoms., packet ae Kashmir: Reported from Lower Jhelum sant (R. R. Stewart, Check list of the flowering Pp. — unpublished). GH). : i i 1700 m. ley; Mirpur, plains to ca. is of West Pakistan and Kashmir . 63. 1754 36. Myosotis L. Sp. Pl. 131. 1753; Gen. Pl. ed. 5 : ioides L. Type species: M. scorpt Leaves alternate; in- ; herbs. Annual, biennial or perennial, pubescent ne 680 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 florescence scorpioid when young, later racemose, ebracteate; flowers usually white, blue, or purple; calyx lobed to deeply divided, not at all to slightly enlarged in fruit; corolla usually hypocrateriform, rarely cam- panulate or infundibuliform, lobes 5, convolute in bud, obtuse, spreading; faucal appendages 5, distinct; stamens 5, included, anthers ovate, obtuse; ovary deeply 4-lobed, style filiform, stigma small, disciform, mucronate; nutlets 4, usually ovoid, rarely subtrigonal-pyramidal, usually compressed, glabrous, smooth, shining, areola small, gynobase usually flat, sometimes slightly elevated. A cosmopolitan genus of about 60 species, with two important centers of distribution, Europe and New Zealand. KEY TO THE SPECIES e Calyx sparsely covered with short, rigid, appressed trichomes of equal length, never uncinate at apex. b. Style much shorter than the See 1. M. caespitosa. be Styse equaling the cays 2. M. palustris. Calyx densely covered with trichomes of various lengths, usually those at r x. c. Plants annual, pedicels much shorter than the fruiting calyx. d. Calyx persistent, erect, appressed to the inflorescence axis; connate part of the calyx covered with spreading, curved to uncinate trichomes, sometimes intermixed with straight ones, trichomes on the lobes ap- pressed, 3 Qu Q po | e QO a 5 c a wn 3 8 =n =] ay = °o QO. oO Og » fo") — OQ =) S =] ber) ee fa) oo pot) at om ° ae — is @o ie) © on “<¢ tal Q ° ? t apex, margined only on the upper half, calyx rounded or cuneate at the base. f. Corolla white to indistinctly bluish, calyx broadly rounded at the base, nutlets broadly margined at the top, not keeled ventrally near the a 6 Corolla blue te pedicels, nutlets tinctly keeled ventrally near the MOE cer. 5. M. alpestris. i Ms d, margins narrow, calyx al- 8. Plants fleshy, less branched, calyx wide open in fruit, corolla meter j : oe We romp wie is, 44°F olan; . M., silvatica. g. P mS not fleshy, much branched, calyx lobes more or less in- curved in fruit; corolla to 2.5 mm. in diameter. __. 8. M. arvensis. 1. bas a ag Schultz, Prodr. Fl. Star ocr. 10: 105. 1846; Ledeb. Fl. Rossi i : . FL ca 3: 144. 1847: Boiss. FI. Orient. 4: 235, 1875; C. B. Clarke in Hook. f. Fl. Brit. India 4: gard, Suppl. 1: 11. 1819; DC. 1971] KAZMI, BORAGINACEAE 681 173. 1883; M. Pop. Fl. URSS 19: 366. 1953; Riedl in Rechinger, Fl. Iranica 48: 257, 1967. | M. scorpioides L. subsp. caespitosa (Schultz) H i i Mitt-Eur. 5(3): 2164, ba a M. lingulata Lehm. Asperif. 110. 1818, nomen nudum. M. uliginosa Schrad. in Mert. & Koch, Rohlings Deutschl. Fl. 2: 42. 1826; ex A. DC. in DC. Prodr. 10: 105. 1846, nomen nudum. Type: Ad pagum Ballin hinter dem Hofgarten, et prope Neobranden- burg am kleinen Ihlpol, C. F. Schultz s.n. (B). Icon.: Reichenb. Icon. Fl. Germ. 18: t. 120, fig. 1-5. 1856, under M. lingulata. Perennial, erect to suberect herb with fibrous, adventitious roots. Stems fleshy, usually solitary to sometimes many, simple to branched, to 60 cm. long, branches slender, ascending to spreading, sparsely covered with white, antrorsely appressed trichomes to 0.5 mm. long, with minute tuber- culate bases. Basal leaves oblong-ovate, entire, rounded to subobtuse, slightly attenuate towards broad base, 2-3 cm. long, 5—8 mm. broad, upper surface uniformly and sparsely covered with trichomes like those on the stem, lower surface glabrous or with some scattered trichomes; lower cauline leaves oblong-obovate to oblong-lanceolate, obtuse, usually broad- est at the middle of the upper half, gradually narrowed towards the base, 3-6 cm. long, 5-10 mm. broad; upper cauline leaves oblong-ovate to ovate, obtuse to subacute, gradually reduced upwards. Inflorescence very short, scorpioid in flower, later elongated, usually curved, bearing fruiting calyces ca. 10 mm. distant, to 30 cm. long in fruit, ebracteate. Pedicels very short in flower, slender, appressed hispid, + horizontal, always longer than the calyx, to 10 mm. long in fruit. Calyx divided + one half its length, cuneate at base, 1.5-2 mm. long in flower, to 3.5-4 mm. in fruit, lobes short, 7s obtuse to subacute, sparsely covered with appressed trichomes, ee mm. long. Corolla blue, 2-3 mm. long, tube equalling the calyx, sae narrowed at the throat, limb 1.5 mm. long, 3-4 mm, in age aan for nearly half its length, lobes ovate, spreading. Style much s the calyx. Nutlets broadly ovate, mature oneés dark brown, gre anes narrowly margined, with rounded apex, 1-1.5 mm. long, ca. | mm. ’ areola minute, 0.5 mm. broad. Distr1BuTION: Temperate and subtropical North America, and North Africa. 400 m., R. R. Stewart West Pakistan: GiLcIT AGENCY: ? e ’ebster 20419 (cui); Skardu, along irrigation canal near Indus River, 58 Meo 7s & Nasir 5778 (cu); Kangan to Wangat, el ease d Shahid Pani, Kazmi s.n. (cH), Hazara Dist.: Siran Valley, between see! ob (pes), 19. 7 ‘1963, R. R. (pes). Swat State: Kalam, ca. 2300 m., opt ae Ae River “near Kalam, Stewart s.n. (pm), A. Rahman 6 (BM); poe yo Aves 25 miles above Bahrain, 1800-2100 m., Rodin 03a, 8002, 12548 (GH), 21472 Kashmir: Pahlgam, 2100 m., 2. cceaueacuaiais regions of Europe, Asia, 682 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 (GH, US); vicinity of Pahlgam, on east Liddar Valley, 27 road-miles north of Islamabad, 2400-2900 m., Dickason 79 (micH); Srinagar, Aug. 1922, Barbous s.m. (BM); Shalimar Bagh (Srinagar), Dickason 78 (micH) ; Ganderbal, 1650 m., R. R. Stewart 7161 (k); Shapiyon, 2400 m., July 9, 1896, C. B. Clarke 28600a (pm); Tangmarg, 1800 m., Polunin 380 (pm); Anchar Lake, 1560 m., Ludlow & Sherrif 8074 (sm, cH). Miscellaneous: without locality, Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (BM, K). 2. M. palustris (L.) Nath, FI. Monspel. 11. 1756; M. Pop. Fl. URSS 19: 363. 1953; Riedl in Rechinger, Fl. Iranica 48: 257. 1967. M. scorpioides L. var. palustris L. Sp. Pl. 131. 1753. Type: Habitat in Europae aquosis scaturiginosis. Icon.: Hegi, Fl. Mitt.-Eur. 5(3): ¢. 221. fig. 4. 1927 under M. scor- pioides subsp. palustris; Reichenb. Icon. Fl. Germ. 18: ¢. 109. 1856. This species is very similar to M yosotis caespitosa from which it is dis- tinguished by having the style equal to the calyx, calyx divided only 1/3 of its length, 3 mm. long in flower, elongated to 4-5 mm. in fruit, calyx lobes more or less triangular, corolla limb 5-8 mm. in diameter. DIstRIBUTION: Temperate regions of Eurasia and North America. West Pakistan: Swat STATE: inter Madyan et Kolalai, Rechinger 30684 (w); Kalam, 2200 m., Rechinger 19390 (Ww); Utror, 2500 m., Rechinger 19580 (w). I was unable to see the Rechinger collections cited above. All the speci- mens from our area examined by me have the styles always shorter than the calyx, the calyces not exceeding 4 mm. in length, and divided for near- ly half their length. This material is of Myosotis caespitosa. 3. MM. stricta Link in Roem. & Schultes, Syst. Veg. 4: 104. 1819; C. B. Clarke in Hook. f. Fl, Brit. India 4: 174. 1883; Riedl in Rechinger, Fl. Iranica 48: 264, 1967 M. micrantha auct. plur., non Pall, ex Lehm. Neue Schr. Naturf. Ges. Halle 3(2): 24. 1817, M. arenaria Schrad. in C. F. Schultz, Prodr, FI. Stargard. Suppl. 1: 12, 1819. Type: “In arvis,” without citation of collector’s name (?). Annual herb. Stems usually many, branched, to 40 cm. long, 2-3 mm. thick white, crisped, spreading, unequal trichom longer trichomes usually arising from tu oblong to oblong-spathulate, entire, obtuse the broad or narrow base, to 4 cm. long surfaces with spreading trichomes to | ' erect to ascendent, profusely at base, covered densely with to rounded, attenuated toward 12 mm. broad, covered on both + mm. long, usually tuberculate at » ovate-lanceolate, obtuse, covered densely on the upper surface with subappressed, thinner and longer tri- 1971] KAZMI, BORAGINACEAE 683 chomes (usually lacking basal tubercles), less densely on the lower surface with shorter, spreading tuberculate trichomes. Inflorescence short in flower, elongated to 20 cm. in fruit, bearing fruits 3-5 mm. apart, ebrac- teate. Pedicels hardly conspicuous in flower, hairy, suberect, usually parallel to the inflorescence axis, rarely subhorizontal, never recurved or reflexed, always much shorter than the calyx, 1-2.5 mm. long in fruit. Calyx rounded at base, divided to nearly half its length, + 2 mm. long in flower, base slightly tapering towards the pedicel, densely covered with thin long trichomes, antrorsely appressed at least at the lobes, intermixed with stout, thick, curved to uncinate trichomes usually on the connate part, rarely above, elongated to 3-4 mm. in fruit. Corolla light blue, tubular-campanulate, obviously longer than the calyx, limb short, 2-3 mm. in diameter, lobes spreading. Nutlets dark brown, ovate, subacute, round- ed at base, smooth, shining, 1.3-1.5 mm. long, 0.8-0.9 mm. broad, areola minute. DistripuTIon: North Africa, Europe, Turkey, Syria, Iran, Afghanistan, West Pakistan, Kashmir, northwest India. Kashmir: Muzzaffarabad, Handu Hills, ca. 2100 m., Kazmi 225b (pes); Tang- marg, 2200 m., R. R. Stewart 12547 (cH). Miscellaneous: Himal. bor. occ. Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (K); without locality, Herb. Late East India Co., Falconer s.n. (GH). Myosotis stricta may be expected in West Pakistan. 4. M. refracta Boiss. Voy. Bot. Espagne 2: 433. 1839; Riedl in Rech- inger, Fl. Iranica 48: 264. 1967. Type: In monte Sierra de la Nieve en el Pilar de Tolox, Prolongo s.n. (c) Annual decumbent to erect herb. Stems usually solitary, sometimes few, much branched near the base, to 25 cm. long and 1.5 aa aes : base, densely covered with white, + soft, retrorsely suberect tric pasion the lower part, with spreading ones to 0.7 mm. long on . ee usually lacking basal tubercles. Basal leaves ovate to ob sais Pa tire, obtuse to roundish, narrowed at base to a short ae sk aa petioles) 1-2 cm. long, densely covered on the upper sur oe sp i SO on the lower surface with white, crisped trichomes oie vat : - arising from tuberculate bases; cauline leaves sessile, ia dean ae tuse to subacute, gradually reduced in size upwards. In Soe minal, simple, short in flower, later elongated to 12 aeons aie mm. apart, ebracteate. Pedicels scarcely 0.5 mm. long 1n . oie ite to 1.5 mm. in fruit, horizontal, slightly recurved to sound : a ; ao “eer Cat ie ee ae di whe lanceolate, long in flower, enlarged to 4 mm. in fruit, inclu ak oe att acute, erect lobes (2 mm. long), connate part 1.5-2 ao : | ose rowed towards the base, covered all over (more densely 684 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 area) with thick, sometimes uncinate or recurved stout trichomes, inter- mixed with + straight, shorter, thinner ones. Corolla blue, tubular-cam- panulate, 1.5-2 mm. long, tube + equalling the calyx, lobes ovate to roundish, suberect to patent, + 0.5 mm. long, nutlets ovate to ovate- oblong, glabrous, shining, 1-2 mm. long, areola small. DIsTRIBUTION OF SPECIES: Southern Europe, Turkey, Iraq, Iran, Af- ghanistan, Turkestan, Pamir Alaj, Tien Shan, West Pakistan, Kashmir, northwest India. 4a. Subsp. refracta. Calyces strongly reflexed in fruit, connate part somewhat longer than the calyx lobes. Nutlets ca. twice as long as broad, broadest at the mid- dle, gradually narrowed at both ends, 1.5-2 mm. long. DISTRIBUTION OF sUBSP.: Southern Europe, Turkey, Iraq, Iran. 4b. Subsp. chitralica Kazmi, subsp. nov. Type: West Pakistan: Chitral, Chitral-Mastuj tract, Suni, 6500 1 ae on stony ground, flowers blue, 16 May, 1958, J. D. A. Stainton 2458 (BM- holotype). Calyces subhorizontales, horizontales vel subreflexi, raro perreflexi; tu- bus calycis plus minusve lobos calycis aequans. Nuculae 1.5 longiores quam latae, supra basim rotundatam latissimae, apex obtusus, 1-1.5 mm. longae, 0.7-0.9 mm. latae. DISTRIBUTION or sussP.: Afghanistan, West Pakistan, Kashmir, north- west India, Turkestan, Pamir Alaj, Tien Shan eg? West Pakistan: CHITRAL STATE: Chitral-Mastuj tract, Suni, 1950 m., Stain- ton 2458 (BM). hmir: Srinagar, 1700 m., R. R. Stewart 11089 (GH). , by the stout, curved to uncinate but are subhorizontal to reflexed. The typical subspecies, which is distributed es, all over southern Europe and the Middle East differs markedly in the shape and size of its nutlets (as described above) from the plants of the species distributed in Afghan- = West Pakistan, Kashmir, and the adjoining parts of USSR and se sgt igi pee Ppa a geographical isolation from the plant 0 the typical su ies, I i r panel a6 peci have recognised them here as sub 5. M. alpestris F. W. Schmidt, Fl, Boém. 3: 26. 1794; M. Pop. Fl. 1971] KAZMI, BORAGINACEAE 685 URSS 19: 377. 1953; Riedl in Rechinger, Fl. Iranica 48: 260, 1967. M. silvatica Hoffm. subsp. alpestris Koch, Syn. Fl. Germ. 505. 1837. Type: Habitat locis paludosis montium Iserae majoris fluvii, Sudetor- um, sylvae Bohemicae, F. W. Schmidt s.n. Icon.: Reichenb. Icon. Fl. Germ. 18: t. 121. 1856. Perennial, caespitose herb. Stems usually many, erect to sometimes decumbent, simple to branched, to 30 cm. long, covered densely with white, crisped trichomes subappressed to appressed on the upper parts, spreading below, to 1 mm. long, usually tuberculate at base. Basal leaves petiolate, narrowly lanceolate, oblong or oblong-ovate, entire, obtuse, (in- cluding petiole) to 8 cm. long and 15 mm. broad, petioles usually longer, sometimes equalling the lamina, both surfaces covered sparsely or dense- ly with slender, spreading trichomes to 1 mm. long, arising from tuber- culate bases; cauline leaves sessile, linear, linear-oblong or linear-lanceo- late, obtuse to subacute, to 2.5 cm. long, 5(-8) mm. broad. Inflorescence short, scorpioid in flower, later elongated, simple or branched, ebracteate. Pedicel short in flower, elongated in fruit to 5 mm., longer than the fruit- ing calyx. Calyx obviously narrowed towards the pedicel at base, divided to or below the middle, 1.5-2.5 mm. long in flower, to 3 mm. in fruit, lobes lanceolate, densely ciliate at margins, cilia long and acute, sparsely covered throughout with short, white, + appressed trichomes, saranda with long, stout, straight or (towards the base of the calyx) curved (an the lowermost sometimes uncinate) spreading trichomes. Corolla ae tube + equalling the calyx, limb 5-9 mm. in diameter, 28% br ng spreading. Nutlets dark brown, smooth, shining, ue nie ae oe ntly apex, apically and laterally margined in the upper half, ey Wee rc keeled near the apex, areola basal, horizontally compressed. DistripuTion: Europe, Asia, North America. species of wide distribution. Ves- 939) recognized many subspecies t geographical areas. Myosotis alpestris is a very variable tergren (Arkiv Bot. 29A(8): 15-22. 199: = and varieties of M. alpestris represented . differen In our area the following subspecies occurs. sont. : ; Arkiv Subsp. asiatica Vestergren in Hultén, Fi. Kamtchatka 4: 80. 1930; Bot. 29a(8): 21. 1939. Type: “Typus in Tomsk.” ear i ecies by ; pg ae the re ca greet ding petioles) 3.5-5.5 ooh LO ei broad; pedicels to 5 mm. long in fruit; calyx i, inti sero stout and long trichomes on the calyx usually s . st os aie or rarely, with few uncinate ones on the lower po Subspecies asiatica i following characters: stems to to rarely oblong, obtuse, (inclu 686 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 DISTRIBUTION OF SUBSPECIES: Arctic Europe and Asia, central Asia, Iran, Afghanistan, West Pakistan, Kashmir, northwestern India. West Pakistan: CHITRaL STATE: Chitral Gol, west of Chitral, 3450 m., Stain- ton 2693 (BM); Owir, 3450 m., Bowes Lyon 886 (Bm); Ziarat, 3000 m., Harriss R 3900 m., Russell 1024 (Bm): Hispar glacier, right bank, E. of Sekarambarisi glacier, 3900-4500 m., Russell 1523 (8M); Biafo Glacier, 4150 m., Russell 1806 in glacier, 3580 m., Lloyd & Megan 72 (pm): Gharesa glacier, junction with Alcock 17744 (pm, kK). Hazara Dist. Kazmi 2387b (PEs) ; Gittidas, Shaukat Ali 124 (BM, RAW); Kaghan Valley, Has- & A. Rahman 24964 (BM); without locality, R. R. Stewart s.n. (GH), Kashmir: Burzil Pass, 3900-4200 m., R. R. Stewart 22057, 22071a, 19864 (GH); Sonamarg, 3500 m., R. R. Stewart 12550 (GH); Pir Panjal, 3300 m., July 6, 1876, C. B. Clarke 287204 (BM), Aug. 10, 1901, Znayat (x); Allyabad, Pir Panjal, Inayat 25704 (x); Zanskar, Pensi La, 4950 m., Koelz 5903 (GH, MICH, Us), 5832, 5854 (GH, US); Rangdum, Zanskar, 4500 m., Koelz 29214 (cu, us), 4500 shed s.n. (pm); Hudan Valley, 3300 m., Bellis in Gambl 1364 (x): B Sind Valley, 2700-3000 m., Duthie 11590 (BM); above Trakbal, 3000 m., Koelz 9199 (GH); Nuna Koi, 2450 m., Pinfold 287 (pm): Pahlgam, 3000 m., R. R. Sept. 1913, Evershed 5.n. (BM); gan Pass, 3300 m., R. R. Stewart On east Lidder Valley, 27 road-miles north ) Ahbad, C. B. Clarke 28388¢ (BM); Rajdhian 22552a (GH); vicinity of Pahlgam, of Islamabad, Dickason 77 (MICH) From the typical subspecies the plants belonging to subspecies asiatica differ in being tall, lax, and havin plants within the subspec; quency of branching, form of leaves, and the short trichomes on the calyx. The long and sti vary from straight to i Most of the collections from West Pakistan any hooked trichomes, I have found hardly any specimens from the Middle East or Afghanistan lacking uncinate trichomes, The very constant and distinct characters, which distinguish Myosotis 1971] KAZMI, BORAGINACEAE 687 alpestris subspecies asiatica from the other species of our area are the form of the calyx and of the nutlet. The calyx in subspecies asiatica is narrowed towards the base with a long pedicel, and the nutlets are mar- gined only on the upper half and apex, and are slightly keeled ventrally, The other species in our area with calyces attenuate at the base are Myosotis stricta and M. refracta which are distinguished by always having pedicels shorter than the calyx. M. albicans has nutlet more or less like subspecies asiatica but has white corollas and the calyx rounded at the base. 6. M. albicans Riedl in Rechinger, Fl. Iranica 48: 261. 1967. Type: West Pakistan: Swat, Utror, 2200 m., 6 June, 1965, K. H. Rech- inger 30826 (w-holotype). Icon.: Ried], 1. c. t. 43. 1967. Perennial (to biennial ?) subcaespitose herb. Stems solitary to many, erect, simple to branched, 5-25 cm. long, covered sparsely with spreading trichomes. Basal leaves petiolate, oblong-spathulate to oblong-ovate with rounded apex, petioles usually obviously shorter than lamina, a etee petioles) 3-4 cm. long, 5-12 mm. broad, covered above with i a and below with spreading, thick trichomes, arising from minute = ous tubercles; lower cauline leaves subsessile to short petiolate, - ns the basal leaves; upper cauline leaves sessile, linear-lanceolate, om long and 3—7 mm. broad. Inflorescence terminal, ed ps oe a young, becoming elongate, densely strigose with grayis , app erage chomes. Pedicels erect, to 2 mm. long in flower, later jit ea cee & ofa as long as the fruiting calyx, to 6 mm. long. Calyx ite forges 8 ox middle, to (1.5—)2.5 mm. long in flower, slightly elonga ' naees® mm., rounded at base, wide open above, lobes sai eipte oy on ae ail low with large, often spreading, uncinate tric agit Siete kage vane sppreserd {0 subeDp ae hite, tube slightly shorter over with short ones. Corolla white to bluish white, X, Ca. 2 m . . y short, stigma capitate. Nutlets (only imm reola small, sometimes narrowly margined, obtuse, ca. 1.5 mm. long, a slightly laterally drawn. DIstrIBUTION: West Pakistan. : i ., Rechinger 30787 ~ West Pakistan: Swat STATE: Kalam, in cedretis, 2000 m., Ae (w); Utror, 2200 m., Rechinger 30826 (w-type). - DC. Prodr. hi. Fl. 1: 61. 1791; ; i i h. ex Hoffm. Deutsc aig Balk f on olny Ghar pete vars.; Ledeb. Fl. Sead : ae Aged Riedl Fl re 4: "237. 1875; M. Pop. Fl. UR : in Rechinger, Fl. Iranica 48: 258. 1967. 688 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Type: In silvaticis umbrosis, F. Ehrhart 31 (?). Icon.: Hegi, Fl. Mitt.-Eur. 5(3): fig. 3125. 1927. Biennial to perennial erect to suberect herb. Stems solitary to few, usually branched to sometimes simple, to 50 cm. long, covered sparsely or densely with white, crisped trichomes, to 1.5 mm. long, arising from minute tuberculate bases. Basal leaves petiolate, oblong-elliptic to oblong- lanceolate, entire, rounded or broadly obtuse at apex, gradually nar- rowed towards the base into petioles, to 6 cm. long, ca. 35 mm. broad, covered on both sides with trichomes, more or less resembling those on the stem; middle cauline leaves sessile, to 12 cm. long, 35 mm. broad; upper cauline leaves narrower and shorter. Inflorescence short in flower, later elongated to 15 cm. and bearing the fruiting calyces to 15 mm. apart. Pedicels short in flower, equalling or longer than the calyx in fruit. Calyx 1.5-2.5 mm. long in flower, 3-5 mm. long in fruit, divided 2/3 to 3/4 of its length, bases obviously rounded, lobes lanceolate, acute, spreading in fruit, densely covered with straight, curved or uncinate trichomes. Corol- la blue, purple, or bluish white, to 10 mm. long, lobes roundish, spreading; faucal appendages well developed. Style equalling or shorter than the calyx, Nutlets blackish brown, ovoid, slightly and narrowly margined all around, acute, glabrous, smooth, shining, 1.5-2 mm. long. DISTRIBUTION OF sPEcIES: North Africa, Europe, Turkey, Iran, USSR, West Pakistan, Kashmir, northwest India. Subsp. tivularis Vestergren, Arkiv Bot. 29A(8): 12. 1938; Riedl in Rechinger, Fl. Iranica 48: 258. 1967. : Type: Taures cilicicus: Bulghar Dagh. j hears g gh, in valle Su Nedere, Kotschy 330 Differing from the typical subs erect, flaccid, more hetero cm. long;. basal leaves li pecies by the following characters: Stems phyllous than in the typical subspecies, to 35 ngulate, cauline leaves lingulate, broadly linear * * The lectotype is selected by J. Grau (Oesterr. Bot. Zeitschr, 110: 521. 1963). 1971] KAZMI, BORAGINACEAE 689 to subovate, apiculate to obtuse; calyx divided nearly to the middle, to 4 mm. long in fruit, lobes narrow, acute, covered with numerous spread- ing to reflexed trichomes, those on the lower part uncinate; corolla 5-8 mm. in diameter; nutlets small, acute, obviously marginate, areola small, DISTRIBUTION OF SUBSPECIES: Turkey, Armenia, Tran, West Pakistan, Kashmir, northwestern India. West Pakistan: Gincit AGENCY: Baltistan, environs of Skardu, 2000-2250 m., Schlagintweit 835 (BM, GH). Hazara Dist,: Kaghan Valley, between Shogran and Sari Rest House, ca. 2500 m., Kazmi 2138 (pes); Naran, stony ground, Jafri & Ali 3421 (pn). Kashmir: Gulmarg, 2400-2700 m., Duthie 13038 (BM), 2700-3300 m., Bar- bour sm. (BM); vicinity of Sonamarg, on Sind River, 50 road-miles north- east of Srinagar, Dickason 81 (micu); Sonamarg, R. R, Stewart 6483 (K), 3300 m., KR. R. Stewart 12549 (Gu), Aug. 11, 1921, R. R. Stewart s.n. (kK); Aru Nafran, upper Lidder, 2700 m., R. R. Stewart 12551 (cu); Fras Nag, Pir Panjal Range, 2700 m., R. R. Stewart 23190 (cu, us); Rattan Pir, 2400 m., 4 July, 1876, C. B. Clarke 28274A (pm); Khelenmarg, 3240 m., Polunin 56/264 (BM); 3000 m., Polunin 56/135 (Bm); Nowboog Valley, Ludlow 29 (BM); Baltal in Sind Val- ley, ca. 3000 m., 26 June 1892, Duthie 11590 (BM); Pir Panjal, 3250 m., 6 July 1876, C. B. Clarke 28720A (pm); Ahbad, ca. 3100 m., 6 July 1876, C. B, Clarke 28888 (BM); vicinity of Pahlgam, on east Lidder River, 27 road-miles north of Islamabad, Dickason 82 (micn); without locality, 1915, Prescott-Decie s.n. (BM). 8. M. arvensis Hill, Veg. Syst. 7: 55. 1764. Type: Pl. 52, fig. 2, a,b,c, Short leav’d Mouse Ear. Annual or biennial herbs. Stems erect, usually many, to 35 “a a branched, branches slender, short to long, covered sparsely pa Paes with crisped, spreading trichomes to 1 mm. long, arising sa an sis bases. Basal leaves petiolate, petioles shorter or mei mea ogee th lamina ovate, obtuse, 1-2 cm. long, to 10 mm. broad, co base; cauline sides with spreading to subappressed ST aide eee 5 ae . e . i cick oo wah bifircate, lax. Pedicels short in below the middle, to 2 mm. long slots lobes more or less incurved, holding the nutlets, Corolla blue to dark with short, stiff, uncinate, spreading trichomes. i i i —- m . . ca] scending to spreading, 3 Lage ae N ie ie narrowly margined all around, black, shining, minute : i istan, India. DistrisuTion: Europe, N. E. Africa, Kashmir, W. Pakistan, | D. Stewart 19018 (us); Kashmir: Burzil Past, 3300-3900 ee eee Mar Pas be ., R. R. Stewar ; ler 37 (x); Baltal, ony eda Kishtwar, 2700-3300 m., 8 Aug. 1918, Fu 690 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Sept. 1880, Young s.n, (BM); Tragbal, R. R. & J. D. Stewart 17958 (cu); Vicinity of Pahlgam, on east Lidder Valley, 27 road-miles north of Islamabad, Dickason 82 (micu); Kishanganga Valley, Keran, 1500-2700 m., R. R. & I. D. Stewart 17520 (cH): Khel to Taubat, 1950 m., R. R. Stewart 17851 (GH); near Bhogtaur, 2280 m., R. R. Stewart 17933 (GH). Miscellaneous: without lo- cality, Herb. Late East India Co., Falconer s.n. (GH); Himal. bor. occ., regio temp., 6000 ped., Herb. Ind. Or. Hook. f. & Thoms., Thomson s.n. (BM, GH) arvensis have slender stems more profusely branched than those of M. silvatica., Pi SER. LABORATORIES P.O. PESHAWAR UNIVERSITY PAKISTAN TWO NEW SPECIES OF LEGUMINOSAE G. P. YAKovLEv IN CONTINUING sTuDIEs on the Leguminosae tribe Sophoreae, material of two previously undescribed species in the genera Ormosia and Diplotropis came to light. Both these taxa seem clearly distinct from their presumed close relatives. My purpose in describing them now is to bring them to the attention of collectors in the hope that new and more complete material will become available through future field activity. I am glad to express my appreciation to the directors and curators of several institutions who were most generous in sending their material on loan for my studies. These include the herbarium of the Arnold Arboretum of Harvard University, Cambridge, Massachusetts (a); the herbarium of the Indian Botanic Garden, Howrah, Calcutta (cat); the herbarium of the Department of Botany, Field Museum of Natural History, Chicago (F); the herbarium of the Department of Systematics and Plant Geography of the Botanical Institute of the Academy of Sciences of the USSR, Leningrad (1x); the herbarium of the Diviséo de Botanica do Museo Nacional, Rio de Janeiro (R); the herbarium of the Department of Botany, University of California, Berkeley (uc); the herbarium, National Museum of Natural History, Smithsonian Institution, Washington, D.C. (us). Ormosia assamica Yakovlev, sp. nov. Arbor 25-30 m. alta; folia 7-foliolata, 3.2-3.7 dm. longa, petiolis circa 5 cm. longis, sparsim subtili- terque pubescentibus, foliolis oblongo-ellipticis vel late lanceolatis, aequabiliter acuminatis, basi truncatis, 10-15 cm. longis, Hs = ee utrinque subglabris, nervis secondariis 10-12, prominulis, er : aes 28 cm. longis. Inflorescentiae appresse ferrugineo-pubescentes a er g : villosae; flores 1.2-1.3 cm. longi, bracteolis jugatis notes La pia pedunculis ca. 0.5 cm. longis, calycibus 0.9-1 cm. longis, dentil ie na a longis, corolla alba, petalis interioribus uniauriculatis. Legumen 1g ‘ Ficure 1. Typr: Assam. Akha Hills, Assam Valley, Dec. 1890, Badul Khan 10 (CAL, holotype). .. This new species is probably most closely oe ee ak) ant (Lour.) Merr, and O. robusta (Roxb.) eat a ‘ . 5 rebate from only specimen I have seen was determined Se : Eas poottern which it differs significantly, as it also does from 0 P and southeastern Asia. : : its closest When fruit of Ormosia assamica is known 1t a pone legumes, relationship will prove to be with those species VOL: (52 PLE Set ps Thess '? Five uf 4 # A ame F . fi ff 4. Be /0 ae 3. f haghs Mee oy Bice 1g t+, BF fle Meh “44 alle Oe Z Te96 Ficure 1, Ormosia assamica Yakoyl ev. Holotype (car), Badul Khan 10. which as a grou Pp I have recently cheskij Zhurnal included in a distinct new genus ( Botani- , In press), 693 YAKOVLEV, LEGUMINOSAE 1971] ie ‘68e uuDmYyNny (1067 ‘OU 4) adAjopoyy “Ysury qsu H . C 3 WSU) € AMINO “pO ayonq ‘(v) adAjeivg ‘AQAOYRA woxyonp sidonoydiqy “(4ja]) Z ANNOY 2 (auyaoyy) esoursze: sidonordiq oN wad oboperD own! ao Ont svnbisy SEN BE SEN 694 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Diplotropis duckei Yakovlev, sp. nov. Diplotropis racemosa (Hoehne) Amsh. var. parvifolia Ducke, Legum. Amaz. Brasil. 104. 1939. Arbor; folia 15-17-foliolata, 7-8 cm. longa: foliolis oblongis basi oblique obtusis, supra cee subnitidis, subtus sparse appresse pubescentibus, 0.7—2.3 cm. ‘longis, 9 cm. latis, petiolulis ca. 1 mm. longis. Inflorescentiae axibus appresse pubescentibus: flores 1-1.2 cm. longi, pedunculis 1-1.5 mm. longis, calyci- bus ca. 0.5 cm. longis. Fructus 1—2-sperma, planus, membranaceus, lanceolatus, 6-10 cm. longus, 1.4-2 cm. latus; semina oblonga, exalbumina- ta, testa tenui.— Ficures 2 and 3 Type: Brazil. Manaos, Estr. do Aleizo, Ducke 1461 (1x, holotype; R, UC, US, isotypes); Manaos, Ducke 64 (a, F, paratypes). The new species, Diplotropis duckei, is most closely related to D. race- mosa Hoehne from which it differs in the number and size of its leaflets and in the character as well as the size of many other parts, as shown in the table below. DIPLOTROPIS RACEMOSA DIPLOTROPIS DUCKEI Lear 11-13(—15)-foliolate 15-17-foliolate LEAFLETS (1.4) 1.7-3.8(4.5) & (0.7)1-1.7(2.3) X 1.4-1.9(2.3) cm. 0.4—-0.9 cm. PETIOLULE 1.5 mm. 3-4 mm INFLORESCENCE 9-12 cm. 4-6 cm PEDICEL 3-4 mm. 1-1.5 mm CALYy 0.6-0.7 cm. 0.5 cm DEPARTMENT OF BOTANY CHEMICAL-PHARMACEUTICAL INSTITUTE Pror. Popov Str. 14, LENINGRAD, P—22 USSR 1971] PERRY, LEONARD J. BRASS 695 LEONARD J. BRASS (1900-1971), AN APPRECIATION Lity M. Perry On Aucust 29, 1971, with the passing of Dr. Leonard J. Brass in Cairns, Queensland, the scientific world lost one of its most kindly and unselfishly co-operative members. He first visited the Arnold Arboretum after the Nederlandsch Indische-Amerikaansche Expeditie (Third Archbold Ex- pedition, 1938-39) to West Irian (then Netherlands New Guinea). This was a co-operative expedition “carried out with the assistance and col- laboration of the Netherlands Indies government, which provided a mili- tary covering party and part of the scientific personnel.” As far as results were concerned, perhaps this was the largest of the Archbold Ex- peditions. Mr. Brass made 5,532 numbered collections on it, amounting to almost 34,000 sheets of specimens. This was the first time I had met him, although I had processed his Fly River Expedition specimens. I handled his collections from four Expeditions in all and kept in contact with him. It seems appropriate therefore, to mention a few of my memories. It was a joy both to process and to study Mr. Brass’s collections. The packages were so trim that one would think he had sheared the edges be- fore wrapping them; the specimens were well chosen, ample, and beauti- fully prepared; his collection data were both extensive and meticulous. He took great pains with his field notes, and he wanted to see them used. The report on his Fly River palms was published without the long, de- tailed (and I may add, necessary) field notes. “Didn’t you send my oe notes with the study set?” “Yes, Mr. Brass.” “Since they werent sd lished, will you please see that the data are sent out a full with ie duplicate set?” Another time a German worker published See Daviumino for Lake Daviumbu (I had seen the proof of the article but tes not od mitted to change the German spelling), so Mr. Brass called ey iat “a to my attention thinking that a mistake had been made obey at aw tion of the label. I simply said, “I will show a the master label, the sent was a carbon copy [which was correct: He seemed to be be : remember vividly every plant he had collected ; ed upon a single small plant and the place he got it. One day I happened up + caw ewnkio- with a beautiful blue flower lying near the edge of a sheet 0 oe tone umbellifer. He was working at the next table in the ia eet ; all the plants were strange to me, I wasn't trying oe , : ‘no precious, so I said, “How had seen, but I didn’t want to discard oe gel He immediately could you mix two plants like these under one ne came to look at the plants, drew 4 small re a ye es ered pocket, looked over a couple of pages, and gen ae “ad bene It belonged there, but had slipped out of order. 696 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 place he had found it, and as he described the surroundings it seemed as if he could see every plant in place. He was an excellent photographer. When we talked about this part of his work he said, “The secret is to take your pictures early in the morning when everything is still.” One anecdote he told on himself was that one morning about 9 o’clock he found a tree which he wanted to photograph but the wind was already beginning to rise. After trying several angles for his camera, the gusting wind spoiling each view, he became annoyed; “Then,” he said, “my boys [helpers] hid behind the trees and laughed at me.” Another time I asked him if it was difficult to get along with the natives. His reply was: “Only once I was in trouble. After emerging from the forest I saw a row of Pandanus trees — one had a cluster of fruits hang- ing. I shot off the cluster. A native came running, and he was very angry, so I offered him the fruits and told him I didn’t know I shouldn’t. Of course a crowd gathered, and I learned that Pandanus was sacred to them and this row had been planted. However, after much talk we parted and I was more careful afterward.” Then I asked about the head-hunters. From his pocket he drew a chain with a charm attached; it was small, made of tiny feathers, a bit of fur and other things. He said that it was a charm for protection. “How did you get that,” I asked. “When you go into the wilds where you have to be your own doctor, you must know how to take care of your- self if you should become ill. I exchanged some of my know-how with the witch doctor.” Incidentally, in connection with cannibalism he told the wife of a staff member who had invited him to dinner, “Some of the Papuans would love you.”’ She was very plump. He had the keenest interest in his collections and the work which was being done on them. If any identifications came in they were to be re- ported to him promptly. Just as he had worked hard to get the speci- mens into the hands of botanists, he expected them to do as much to put the results into his, but he helped in every way that he could. Many times when I had questions about habitats, I received a prompt and clear explanation of what I needed to know. One time he told me, “I was too late to get help from Corner’s monkeys and a bit early for access to the field by airplane,” —so he had had to m i i ake coll ; coping with many difficulties. e collections the hard way report on his plants collected on t feeling apologetic for the delay. His reply was, “In this matter of getting 1971] PERRY, LEONARD J. BRASS 697 Saree aed od iv ) 2 May, 1953. Photo. Leonarp J. Brass at Baniara (Milne Bay Div., Papua R. D. Hoogland. 698 JOURNAL OF THE ARNOLD ARBORETUM [VoL.. 52 things done, I used to think any persons interested in their job ought to be able to work eight days a week. I still do mostly six days, plus some night time. But ...I have to admit that I can’t get as much done as I used to . . . even though conscious weight of years (almost 61 of them) is something I have had no cause to bother about so far.” In 1965, Mr. Brass sent me (and Dr. van Steenis) a summary of the numbers, general localities, time, and place of deposit of his collections, approaching 35,000 in number. But he was not only a collector, he was also a naturalist and a brave explorer. His consideration of others per- haps may be illustrated by the following: after he returned from one ex- pedition on which another member of the staff (a zoologist) had suffered a stroke, he took over his collections and processed them for him at the American Museum of Natural History. I said, “How difficult when he couldn’t talk.” His reply was, “I watched his eyes; I knew him well enough to read approval or disapproval in the expression of his eyes.” When the last expedition went to New Guinea (while he was still living in this country but unable to go along) he took care of making all the ar- Tangements and of hiring whenever possible Papuan helpers he had al- ready trained on earlier trips. His mind was never very far from New Guinea. We remember his charming smile, the interest with which he greeted us, his readiness to help, his remarkable capacity for work and his complete devotion to it. Those of us who knew him and the scientific world are the poorer by his going. ARNOLD ARBORETUM 22 Divinity AVENUE CAMBRIDGE, MAssacHUsETTS 02138 1971] INDEX INDEX Acanthopanax giraldii, 210 227 Acrotrema appendiculatum, 321 — gracilistylus, 210, 227 — arnottianum, 3 — henryi, 210, 227 — costatum, 310-328, 327, 328 — innovans, 210, 227 — dissectum, — koreanus, 210, 227 — gardneri, 321 — leucorrhizus, 210, 227 — lanceolatum, 321 — senticosus, 210, 227 — lyratum, — sessiliflorus, 210, 227 — thwaitesii, 321 — setchuanensis, 210 Actinocarya, 353, 355 — spinosus, 210, 227 — acaulis, 354, 355 — trifoliatus, 210, 227 — tibetica, 353, 355 br Shoot Growth and Heterophylly in, Adelanthaceae, 450 del Acer, Wee Adelocaryum flexuosum, 352 sect. yon 258, 260 — schlagintweitii, 352 — sect. Arguta, 258, 260 Adhatoda, — sect. Campestria, 258, 259 — betonica, 636 — sect. saree ay 259, 265 — carthaginensis, 641 — sect. Palm origanoides, 641 — sect. Patani, 258, 265 , 640 — sect. Rubra — zeylanica, 637, 640 — sect. tia 258 Adicea, 55 — sect. Spicata, 262 Adike, — barbinerve, 2 Aeginetia, 408 — buergerianum, 261, 262 Nea a 408 — campestre, 259 gava — sicdiinex 9 59 Alchornea yr 437 — caudatum, 258 Alectra, 614, ee — creticum, 241 — br. iliensis, se — ginnala, 261, 262 — fluminensis, 6 is — grosseri, 259 — melampyroides, — hyrcanum, 259 Alloplectus ambiguus, 591 = miacrophyllum 256 Alobiella egy? 449 ’ — miyabei, 258 Alobiellaceae Abe site St 10h — monspessulanum, 259 Alobiellopsis omiice : ; — opalus, 2 Alvaradoa jam: coe ; _ _ 241, 259 as Bee ee — palmatum, 264 — egy — pensylvanicum, 241-248, 250, 254, 255, pee nig a st $39, 850, 587, 559 — cubenis, 574, — platanoides, 246, 253, 258, 264 Cc a — pseudoplatanus, 240, "246, 253, 261-264 ae si, — pycnanthum, 258 ae % 574, 581 237, —hottlei, 574, rubrum, 241, 242, 244, 249, 252, Amnpasscopinell po , 259 Amphora a a hdl a, 9. aa saccharinum, 258, 264 sage ala, 96 a a ie 258, 262 ne ciate, 102, 106 ial ay — tataricum, 261, 262 Bi es 96 — tegmentosum, 25 __ oscillarioides, 06 Achanthes —o 93 Anacandiacen®, 300 bes — micr hala, Ananthacorus angustifolius, ep Acrotrema, 319, pon 322-327, 329, 330 700 Anchusa, 668 — asperrima, 490 — petiolata, 350 a elaet 490 — zeylanica, 344 Aster echiodes, 637 Andropogon bicornis, 588 — squamosa, 36 Annonaceae, 12, 15, 32, 34-39, 285, 286 366 Anomoeoneis cos tata,99 Anoplolejeunea conferta, 451 Anthacanthus, 645 , 581 Arlhanocaps biformis, 95, 102 ee revillet ‘Apbanadieies'sa a 95, 102 — microscopica — Saxicola, Apiopetalum, 234 — spinosa, 210 Araliaceae ?, What is th Structure of, 205 Archilejeunea Viridissima, 437,451 Ardisia — oS 590 Arnebia aioe — € Primitive Floral rrima, = benthami, 360, 361, 363 — bungei, 496 — cornuta, 498, 500 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Arnebia decumbens, 360, 498, 499 a hispidissima, 360, ‘489, 491 — inconspicua, 3 — ce 360, ‘497 lani, 498 — waziristanica, 360, 501 Arnebiola migiurtina, 490 yocenartslgarg acutangulum, 646 Artabotrys, ne — odo ads Arthrophyllum thernianum, 211,'228 — diversifolium, 228 Arthrostylidium sarmentosum, 588 “oi hom oraqes e, 80 Asia in Cultivation, The Genus Neil- 137 rocumbens, 118 Astrotricha Pockoéa, 211, 228 Aulosira fritschii ii, — shrochieg he = 102, 103, 106 Aureolaria Pasineeseh- tes 3, 285, 2 Austrobaileyaceae, 3, ms ere 36, 285, 290 serena aig oid 92, 93, 102, 104-106 Balan Bazzania bidens, 446 — breuteliana, 446 — cubensis, 441, 447 1971] Bazzania eggersiana, 438, 447 — gracilis, 441, 447 , 447 — schwaneckiana, 438, 447 — stolonifera, 441, 447 Beech, 168 Beechdrops, 414, 416, 419 Beech Family, 159 Betulaceae, 160, 306 Beureria Apap 622 . venosa, 622 N. N. Embryology of the Magnoliales and Comments on their sitasiecatati, 1-39 ; 285-304 Biddulphia alternans, 102, 106 — favus Bignoniaceae, 591 Bladdernut, 199 Bladdernut Family, 196 Blechnum occidentale, 438 Boehmeria, 59, 62 — subg. Duretia, 60, 62 sie a aala novo-guineense, Pianist 211, Sale, 211, 226 % Boranaceae Me 614, 619, 626 Bor oe West Pakistan and as ck Revision of, 110-136, 334- 363, 405-524, 614-690 Boraginella, 515 — spinulosa, 516 INDEX 701 Borrago verrucosa, 5 Borreria bach 438 Bourreria, 619-6 — arborea foliis oat alternis, 622 — baccata, 621, 6 — ovata, 621 — rigida, 621, 622 — succulenta, 621, 622 — tomentosa, 621 Brachionidium ciliolatum, 588 Brass, em ms (1900-1971), An Ap- heer rassaiopsis sepia 211, 227 Bredemeyera, 26 Porabeihtay 6 be 588 rse Burseraceae, 366 Byrsonima spicata; 439 Caccinia, 666 — russellii, Calliandra, 69, 77, 79, 80 — boliviana, 83, 8 — a re 69, 70, 74, 75, 77, 78, 84 — nmatocepba History, Morphology, Taxonomy, 69 — penis 69, 70, 72, 74-77, 80, — guttata, Calothrix epiphytica, 96 weberi, Calycanthaceae, 93,32507 Calycanthus, 23, 25, — occidentalis, 2 Re Calycogonium squamulosum, C viet caespitosa, 438, 44 __ cellulosa, 443, 447 439, 590 — Jaxa, 448 ey . weer 442, 447 she ste as Caco Myr Calyptranthes es acutissima, 366 — krugii, a 591 702 JOURNAL OF THE ARNOLD ARBORETUM Camptylonema indicum, 97 Campylodiscus echeneis, 102 Carex polys stachy a ” 588 Ptr 5, ere Be 38, 291 — fili — glabella = lig Castanea, 160, 161,177 — sect. Balan mocastanon, 175, 176 — sect, Castanea, 1 si uarinales, 3 ceca bal um, ralis, 574 — ara cone lig 581, 584 — caudata a aan 574 — — var. texana, 581 — — var. crass if — pallida, ‘S74, 581 — paniculata, 574 — philippinensis, 574-581 ila, 584 occidentalis, 574, 581, 584 olia, 581 —_ i i 581, 584 — aaa | Ht — schippii, 574, 581 — Sinensis, 574, 581 Celtis soyauxii, 574 — spinosa, 581 — swartzii, 574 — tala, 574 — tenuifolia, 574, 581 — tournefortii, 584 — triflora, 581 — trinervia, 574, 581 — ugandensis, 574 — wightii, 574 — yunnanensis, 574 — zenk Centrales, 93, be 102, 104, 106 ephal Céphalearos oad, 104 Cephalozia caribbeania, 448 — crassifolia, 448 — subforficata, 444, 448 Cephaloziaceae, 448 Ceratocystis fagacearum, 188 Ceratolejeunea brevinervis, 443, 452 5 — rubiginosa, 438, 452 — spinosa, 452 — valida, 440 Cercidiphyllaceae, 31, 33-39, 291-293 Cercidiphyllum, 31, 32, 241, 291, 292 Be cu. n- emma 32, 263 gnificum, 31 Cerothamnus ceriferus, 310 — inodoru Chactachme, 523-543, 554-558, 565 — aris ta, 575, 581 rocarpa, 575, 581 Chashophoraica, 97, 104 ie 524, 534-538, 547, 556, 557, cies seo 581 um, stipitatum, Pk decidua, 438, 452 — polyan Cheirodendron kauaiense, 211, 227 27 Chestnut nes 177 25 ie, pos Chloranthaceae, 293, 589 [voL. 52 1971] Chlorella oo 104 — vulgaris Miscbccales, 97, 104, 106 Chlorococcum humic sls, 98 Chlorophyta, 87, 93, 97, 102-106 — turgidus, m5; Ace 103, 106 — varius, 95, itr, ues 93 Chrysolepis, 160, 16 Circacocarpus 479, 480 Clearweed, 55 ~ a albopunctata, 589 Clibadium erosum, 438 Clusia grisebachiana, a 590 oelosphaeriu um dubium, 96 Cohnia, 461 Colura clavigera, 452 Composi a on Comptonia, oped 315-317 — aspleniif olia, un iam a7 na, 3 Dacha, Ay 406, 413, 420-423 Convolvulaceae , 591 Opper Oey 169 any vs 614, 620, 630 ‘ree 629 — adnata, 631 ~~ Var. globosa, 632 —~ Var, humilj a bun _ 631 amaicensic 627 Sng 627, 631 INDEX 703 Cordia polycephala, 629, 630 — ulmifolia, 629 ere 459-461, 463, 465-471, 473, — australis, 461, 463, 466-471, 473, 474 — banksii, 461, 463, 466-471, 474 — dracaenoi des, 46 — indivisa, 461, 467, st iy 473, 474 — kaspar, 467, 470, 4 — pumilio, 461, 466, phi pcs 474, 476 — stricta, 461, 467, 471, 473, 474 — terminalis, 461, 463, 466, 467, 471, 474, 476 Corylaceae, 160 ory. 2 , 42 Crematomia attenuata, 621 — elongata, 621, 622 — molliuscula, 621, 622 — velutina, 621, 622 — venosa, 621 CRITCHFIELD, Wi1Ltt1AM B. Shoot Growth and Heterophylly in Acer, 240 Crossotolejeunea bermudiana, 438, 452 — boryana, 438, 452 CRANDALL, BARBARA, RAYMOND STOTLER, and MAarGARET FuLrorp. The Ecology of an Elfin Forest in Puerto Rico, 15. A Study of the Leafy Hepatic Flora of Ctenolophonaceae, 680 CULLEN, J. The Genus Neillia (Rosaceae) in Mainland Asia and in Cultivation, 13 Cussonia paniculata, 211, 228 Cuphocarpus, 235 Cyanophyta, 86-88, 91, 93, sig 102-106 Cyathea sete: 89, 90, 94— Cyclobalanopsis, Cyclolejeunea ee 440, 452 Cyclotella nee 104 — glomerata, 93, prego phere 98, 104 risitnteauisiadas: 98, 104 Cylindrocystis brebissonii var. minor, 99 Cylindrospermum catenatum, 102 Cymbella ventricosa, 102 704 JOURNAL OF THE ARNOLD ARBORETUM [voE; (52 Cymbianthopsis sintenisii, 590, 591 42 — hervosum, 343, 347-349, 351 — — Var. nervosum, 350 — — var. petiolatum, 350 — officinale, 342 — ot 350 Bi er ddraannh + — racemosum, — shane mae — stamineu — stewartii, eg 350, 352 — stylosum 337 2: wallichii, ie 345, 346 ystodinum cornifax, 105 Cystolejeunea lineata, 440, 453 Dacryodes excelsa, 437 Dactylococcopsis fascicularis, 103 — smithii, 9 et americana, 439 Decalepidant nis 677 — sericophyll s, 677 Dezeneraea, 9 11, 32, 34-39, 285, 286 287 —— — viti Pap ame lauterbachi, 211, 228 ta, 642 — foliis lanceolatis, 642 pecto: Diatoma hiemale, 99, 104 — — var. mesodon, 104 — tenue var. elongatum, 104 — vulgare, 93, 101, 104 Diatomella balfouriana, 99 CKIson, Wutt1am .C. Comparative Morphological Studies in Dillenia- pei Aer Additional Notes on Acro- “ia Diclidan ther 268 Dicliptera chinensis, 637 — sexangularis Didymopanax attenuatus, 211,:227 — morototoni, 211, 227 Dillenia, 329, 330 Dilleniaceae, Comparative Morphological Studies “tg lap Additional Notes on Acrotem iiicalaccsr oe Dillenioideae, 328 Dilomilis montana, 588 Dinococcales, 105 Dioclea hispidissima, 490 Dioscorea, 476 Dioscoreac eae, 588 ae eae 366, 369, 393, 396, 397, 400, — eal Cupulifera, 379 — sect. Forsteria, 370 — sect. Rhipidostigma, 370, 396 — subgen. Diospyros, 370, 397 — subgen. Maba, 369, 370, 395 — andersonii, 390, 3 — christophersonl, 371, 383, 394, 395 — discolor —_— io ll 310-32, 374, 376, 377, 379- 383, 393, 3 — — Var. crotial 373 Oo - — — var. savaiiensis, a. 379, 383 — palpcigares 373 — ferrea, pesh ot 380, 382, 383, 387, 389, 90, 393 — — Var. gillespiei, 384, 385, 388, 389, 392 — — var. glabrescens, 383 — — var. lateriflora, 390, 3 4, 395 — — var. nandarivatensis, pe 386-388 —— — Var. Samoensis, 394 — — var. savaiiensis, 379, 389 — Var. subimpressa, 384, 385 _ =i Silken, 378, 379 — globosa, 397, 402 1971] Diospyros lateriflora, 390, 394 — longisepala, 397, 40: — lotus, 397 — major, 371, 383, 390, 395 — parviflora, 379 — rufa, 379 — samoensis, it 398, 400, 401 — — var. longisepala, 401, 402 i — 307, 38, fi 401 — — var. vitien: Diplasiole} piceoe parm 453 Diplopanax sacipaanal 211, 228 Diplotropis, 691 — duckei, 694 — racemosa, 694 — — var. seas 694 Ditta myricoides, 439 Dizygotheca elegantissima, 211, 22 7 Dracaena, 459, 460, 467, 468 — tomentulosa, 644 Drepanclejeunea anoplantha, 453 — crucianella, 453 — evansii, 453 — fragilis, 438, 453 — inchoata, 438, ane? — lichenicola, 438, 4 ee 6, 2 54.35; ae 287, 289 sect. Tasmania, 9 — lance 9 eri, re 9, 36, 38 Divediylioks 188 Ebenaceae, 366, 369 Ebenus andersoni, 390 — elliptica, 373 — fasciculosa, 395 — foliosa, 378 Ecbolium carthaginense, 641 de, 637 Echiochilon anton 490 os asperri 90 — bunge INDEX 205 Echioides griffithii, 492 and * Interpretation of its Season- ality, pongito pean M. Solanaceae, 634 es, 305 i gic S., Lorin NEVLING, jr., and, Calliandra haema ts ala: pr Morphology, and axonomy, tory, Morphology soe Embryology of the Comments on their eutlonnhlbs: 1- Eremopanax angustata, 211, 228 Ericaceae, 590 Eritrichium acaule, 354 Ervatamia, 61 kere 00S 650 er, 403 Eugenia 4 borinquensis, 441, 590 _— perpusilla, 1 — praerupta, 93, 102 706 Eunotia soleirolii, 104 — tenella, 93, 100, 102, 104 Eupatorium odoratum, 437, 438 Euosmolejeunea clausa, 439, 453 — rigidula, 453 — trifaria, 437, 453 Euphorbiaceae, 36 Bupomatiaceae, 32, 286, 288 Euptelea, 285 Eupteleaceae, 288, 289, 292 196 and : is pies Primitive Floral Structure of Araliaceae ?, 205 Fagaceae in the Southeastern United States, The Genera of, 159 Fagaceae, 159, 160, 162, 166 — subfam. Castaneoideae, 161, 162, 173 — subfam. Fagoideae, 161, 162, 166 Sider agen Quercoideae, 162, 163, 179 Fagales, Fagus, iso, 168, 170 False nett atsia j rieeian 211,227 Ferreola, 370 —_ si ai oe 373 Ficus pe FIsHER, ig ee and P. B. Tomutnson. Morphological Studies ordyline Cor (Agavaceae) I. Introduction and Gen- eral Morphology, 459 FOERSTER, JOHN W The Flacourtiaceae, 366 Flax Family, 649 Floral Structure of Araliaceae, What is the Primitive ?, 205 Forestiera, 616 — rhamnifolia, 614 ——= Var. pil 615 iar: appre 614 Pea ponl Friedirichethalia { incana, 518 Frullania atrata, 440, 442, 451 os brasiliensis, 451 Frustulia ogress var. capitata, 91, 100, 102, 1 105, 1 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 “apse rhomboides var. Soe 91, 93 —_—— viridula, 91, 93, 102 an, Marcaret, Ba ARBARA CRANDALL, he Hepatic Hepatic Flora of the Luquillo i aeniiiica: 435 Galacea Gamblea ciliata, 211, 227 Garryales, 306 Gastonia horkdiiing: 211, 226 26 — hispida, 678, 679 Geminella interrupta, 104 Gendarussa, 63 — hyssopifolia, 637 Genera of Fagaceae in the Southeastern United States, The, 159 Genera of Myricaceae in the Southeastern United States, The, 30 Genera of Orob aiichaceat in the South- eastern U 404 ® B 5 3 7 BE a S -O in Mainland Asia and in Cultivation, The, 137 , 409, 591 Ginkgo biloba, 240, 241, 248, 2 Gironniera, 524, 534, 542, 553-554, 558, 559, 565 — celtidifolia, 575, 581 — cuspidata, 575, 581 oe linearis var. composita, 95 , 10 Glocotrichia echinalata, 96 lata, 96 1s, 590 cet cea spicata, 438 1971] Gramineae, 588 Guapira fragrans, rey 367 Guzmania berteroniana, 588 Gymnotheca, 479, 480 Haenianthus waite 7 442, 591 Hamamelidaceae Hamamelidales, ee Hantzschia amin, 105 — welwitschii, 93, 97, 103 H. aculeatus 228 — harmsii, Harpalejeunea rae 443, 454 54 Hedycarya angustifolia, Hedyosum arborescens, aa 439, 589 Hemiptelea, 524, 534, 541, 551, 556, 559- 566 — davidii, 575, 583, 584 Hepaticae, 436, 438 Herberta divergens, 438, 446 — ae 44 uniperoidea, 446 is, 446 Herbertaceae, 446 Hernandiaceae, 33, 290 — nsi ales Heteropanax eae 212, 228 — fragrans, 212, 228 Heterophylly and Shoot Growth in Acer, 240 Heteroscyphus elliottii, 449 Heterotrichum cymosum, 437 Hibbertia, 325, 330 Hibiscus tiliaceus var. purpurascens, 384 Himantandraceae, 32, 286 Hirtella rugosa, 439 History, Mor , and Taxonomy of, Calliandra ‘heematocenalat , 6 po maar by 534, 543, 557 grandis, — ssc he 575, 583 INDEX Hornemannia, 440 Hyalotheca undulata, 99 Hygrolejeunea, 454 Hyobanche Hypoestes purpurea, 637 ee a 588 Idenburgi Tlex sine oy 589 Inga 707 Ua theeas be 32, 34, 35, 37, 38, 287-289 ici 287 ? ’ — anisatum, 18, 20 — floridanum, 18, 20 — parviflorum, 18, 20 Indorouchera, 651 Indokingia crassa, 212, 226 a Isachne soa 588 Isthmia enervis, 102 Isotachidaceae, 446 Ixonanthaceae, 650 Txora ferrea, 437 Jacobinia jamaicensis, 64 amaica, A New pak of Reynosia (Rhamna naceae) from, Jasminum periclymeni folio, fiore albo, 25 6 Jounston, Marswatt C. A New Species of Reynosia (Rhamnaceae) from Ja- maica, 364 Jun Justicia, 614, 636-659 — adhat toda, 637, 639 ae — alai — beto = oot) 639, 640 __ brandegeana, 639, 640 m __ maestrensis, 643, 644 __ martinsoniana, 591 708 Justicia mirabilioides, 643, 644 as — sexan — tomentosula, 643, 644 Kadsura, 20, 21, 288, 289 Kaernbachia, 197 Spain pictus, 212, 227 Kazi, S. M. A. hi of the ae i of Wes akistan and Kashmir, 110-136, Spa: A 486-522, 14-690 Kibara Kissodendron australianum, 212, 227 Kopsiopsis, 405 Krameriaceae, 268 Lactoridaceae, 36, 289 87 Lauraceae, 25, 26, 28, 33, 35-39, 290, 291 589 Laurelia novae-zelandiae, 23 Laurus, 25 New Species of, 691 oe sinks, Sophoreae, 691 Leiocarya, spore fg Lepidoziaceae, is t € macrostachya, 361 Leptolejeunea — 438, 454 — ulata, 45 Leptoscyphus ebbaus 444, 449 Bi one age hae a, 642 pice ete xanthocarpa, 439, 454 ee Portoricense, 439, 449 Liliacea 459 ceils in the Southeastern United States, The, 649 JOURNAL OF THE ARNOLD ARBORETUM Linaceae tribe Eulineae, 650 — anchusoides, 334, 336, 337 — benthamii, 337, 3 — longiflora, 334, 337-339, 341 335 Gan 649, 6 — sect. Baim, 655 — sect. Cliococca, 658 m, ine — speciosum, 3 — mers, 27 Lizard’s- i hand 479 8 or Bens, 438 is, 591 Lophocole ¢ euadeloupensi 438 _ ati, 575, 582 Lycopodi 2 438 — meridionale, 1971] INDEX 709 Lycopsis arvensis, 669 sehen brachiata, 441, 455 — caspica, 667 Mastoglo — micrantha, 669 Mattia ia: 128 — orientalis, 669 — staminea, — picta, 669 Mattiastrum, 126 — pulla, 676 — sagt yy — taurica, 669 — bungei, 13 Lyngbya limnetica, 96 _ settee oy 128, 130 — honigbergeri, 1 Maba, 3 — howardii, 129, 130 — sect. iced 395 — andersonii, 390, 391 — tibeticum, 131-133 — buxifolia, 37 — elliptica, 370-3 73, 380, 381, 383, 388 Megalopanax rex, 212, 22 — — var. glabrescens, 383, 384, 388 Melasma, 635, 636 Meliaceae, 589 Melosira nee 99, 102, 104 — major, ; — nandaravatensis, 386 _ granulata, 10. — samoensis, __ varians, 93, 102, 105, 106 — savaiiensis, + 2 380 Menispermaceae, Macrobalan Merismopedia aerugina, 95 Mackinlaya eas 212, 228 — elegans, 99; — macrosciadea, 228 lauca, 9 ‘is — subsimplex, 438, 454 — eit 5 Macropanax undulatus, 212,227. = 212, 227 Macrotomia benthamii, 361 Merrilliopanax es — endochroma, 486 i siociaiti Miche — euch 486 ae oe eae 486 apie cuit, 442, 590 <6 - ees 488 — pachyph: ylla, 89, 90, 9 94-100, eae 7 2 ee 4 sein aie: pis — racemosa, 438 i ; on isii, 590 — onosmoides, 486 oe a rie "437 M pa a ie 38 Micrasterias ws denticulat,108 agnolia, 4, 6, 36, Microchaete uberrima, . tien 3 eee aenano = tee — ela — soulangeana, 6 © ents, "95 — splendens, 437 a sic ca, 103 — stellata, 4 a, 106 a — virginiana, 35 eee acutifolia, 4° Masnoliacee, 4, 32, 34-39, 285, 286, 288, a. 8 __ Jaetevirens, 455 tee and Comments on their Saget _—— monoica, 439 sjlaris, 447 ra Embryology of the, Microlepidozia ca mgt ~304 pholis £4 447 Magnolias a a be Microptery sium carinatum, 444, pighia glabra _— exalatum Malpighiaceae, 268, 367, 650 Mikania mes 591 Malvales, 43 Miliusa, 12 Manglietia, 6 dig ara? 35 Margarospermum arvense, 356 Milkwort, 27 “ad — officinale, 358 Milkwort Fai 2 The Genera of the Marcgravia ‘sintensii, 440, 589 Mutter, Nor Marcgraviaceae, 589 710 Urticaceae in the Southeastern United tates, 40 Miter, Norton G. The Polygalaceae in the Southeastern United States, 267 Mirandaceltis, 525, 534, 542, 553, 558, 565, 566 — monoica, 575, 582 Mitropsidium sintenisii, 442 Moltkia libanotica, 135 Mollinedia, 21, Monimiaceae, 21, 24, 32, 35-38, 285, 289, 29 Montjolya brownei, 627 Moraceae, 43, 366, 589 — subfam. Conce ocephaloidea 42 py shale Studies in Cogs tee (Aga- aceae) I. Introduction and General 459 , 486 Pe mri racemosum, 215 Muraltia, 26 Musanga, a yodocarpus involucratus, 212, 228 Myosotis, 679 — albicans, 680, 687 a, Bere srt 337 — uligin Myrein ¢ mes a: Myrica, 306, 308, 309, 313, 317 — sect. Cero phora, 309 — subg. Myricn, 306, 311, 312 — subg. Morella, 306, 309, 310 Myricaceae, 305-308 —— aceae in the Southeastern United States, The Genera of, 305 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 giclee ' ipoaceen 14, 34-39, 288, 290 ceae, 590 e, 310 Se spectabilis, 103 Navicula, 103 — carniolensis, 103 — Poe apy 103 Re age Arie “ae 439, 446 Neillia, 137, 139, 157 — affinis, he tae 147, 149 — — var. affinis, 148, 154, 155 — — var. ge 149, 155 — fallax — sandal, 140, 151, 152 — gracilis, i nea 140, 153 — serratisepala, 138, 139, 146 — sinensis var. caudata, 140, 148, 152, 153 — — var. hypomalaca, 140, 148, 153 — — var. ribesioides, 138, 148, 156 Fa Val roses. 149 — — Var. sinensis, 0 — 154, 156 — — var. villosa, 14. — Sparsiflora, 139, sa — thibetica, 138-142, 148, 150, 152-156 —~ — var. caudata, 151 — thyrsiflora, 138, 139, 144, 146, 149, 154, 155 — tunkinensis, 145 — — var. bibracteolata, 145 — uekii, 149, 150, 154, 156 osa, — virgata, 144 Nepsera aquatica, 439 Neraudia, 40 Netrium digitus, 99 0 5 EVLING, Lorin L., Jr. The Eco! — of Ch 1971] INDEX NEvLING, Lorin I., and THomas S. ELtas. Calliandr. — dissipata, 100, 103 — hantzschiana, 93, 103 _ eta 100 — tyblionell 105 p. zygomorpha, 673 — edgeworthii, 671, 674 — kandaharensis, 671, 674 sp. rudbarensis, 677 — turcomanica, 671, 675 Nonnea diffusa, 671 Nowellia caribbeania, 440, 448 minicensis, 440, 448 Nyctaginaceae, 367 Oak wilt disease, 188 Seer 589 Ochran Octotheca, 234 ontolejeunea lunulata, 438, 455 Odontoschismaceae, 448 Odontoschisma denudatum, 448 8 Oleaceae, 591, 614 Omphalodes, 119, 131 — rugulosa, 121 History, al, 197 ene com 89, 90, 94-100, 439, 589 711 Onosm — sie m, 508 ne tine, 503, 511, 512 atralicuna, 503, 504 _ emer ine S03 512 atum, — khyberianum, 507, 509 macrocephala, 511 — setosum var. dicroanthum, 512 — thomsonii, 503, 510 Gonystis borgei, 98 y riformi Orchidaceae, 80, 3 patina foribundus, 212, 226 ae ee 2522 — pec a a6 226 e€ 3, 226 404-409, 413, 425 theastern United — amnae, __ foreaui, 93, 106 712 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Oscillatoria geitleriana, 103 Phormidium tenue, 96, — limnetica, 96 Phyllostylon, 524, 534, = 557, 559 — subbrevis, 103 — brasiliensis, 575, 582 Osmoxylon, 234 — rhamnoides, 582 Oxalidaceae, 650 Physocarpus, 137, 157 Pilea, 55,57 Pachylarnax pleiocarpa, 6 — sect. Dentatae, 56 Pachynema, 319 — sect. Heterophyllae, 56 Palmae, 366, 588 — sect. Pilea Palmervandenbroekia, 234 — krugii, 589 Panax quinquefolius, 213, 227 — yunquensis, 589 — trifolius, 213, 227 Pinnularia sudetica, 93, 103 aracaryum, 120 — viridis var. mi — asperum, 127 Pinus clausa, 075 — bungei Piper hispidum, 437 — bornmulleri, 132 Piperaceae, 293, 589 — calathicarpum, 122 Piperales, 293 — heliocarpum, 33 Piptocalyx, 285, 290 — himalayense, 128 Pitcairnia angustifolia, 439 — intermedium, 123, 125 Pithophora oedogonia, 104 — longiflorum, 338 Plagiochila abrupta, 439, 450 — micranthum, 123 — amoena, 438, — papillosum, 125 — bicornis, 443, 450 — platycalyx, 122 — bidens, 440, osum, 120-122 — breuteliana, 450 — rubri 131 — bursata, 440, 450 — tibeticum, 131, 132 — chinantlana, 441, 450 — — var. schlagintweii 131 — confundens, 440, 450 — thomsonii, 13 — distinctifolia, 443, 450 — stellatum, 12 — divaricata, 4. 5 Paracromastigaceae — dominicensis, 44, 450 Parasponia, 524, 334, py 552, 565 — jamaicensis, 4 582 — rutilans, 438, 86 _ 575 — simplex, 444, 450 Parietaria, 6 — subbidentata, 450 — subgen. Freirea, 66 = i ag 438, 443, 450 — floridana, 40 — tenuis, 450 Parthenocissus, 241 Plagiataiic ac 50 Planera, 524, mae 544, 556-560, 566 a m, — aquatica, 575, Pedicularis melampyroides, 636 Plectonem : ms Sorting spectabilis, 213, 226 Plerandra insolita, 213, 226 Pellito — solomonensis, 213, ohciien teiiities, 95 — vitiensis, 213, 226 Pennales, 93, 99, 102, 104, 106 Plerandropsis, 234 Pentapanax castanopsidicola, 213, 227 Pleurocapsales, 97, 103 — subcordatus, 213, 227 Peperomia emarginela, 589 — he diifolia, ae ; Plumbag ericlymenum rec = 31 Podo i —e planctonia, 99 aad. ia > ¥ M. Leonard Brass (1900- — africana, 519 ‘ i 0, An Appeedeiieg., 695 — indica, 516 ersea Pp — Heenan yr pe eek, aes — bo 14 Peumus, 21, 23, 3 Polygala, 268, 269, 271, 278, 279 au : Seet, Semeiocardiu um, 2 oldus, 21, 23, en 38 — sect. Pseudosemeiocardium, 273 1971] INDEX 713 Polygala subg. Chamaebuxus, 272, 274, Quercus subg. Cyclobalanopsis, 18 280 ubg. Ery chro 180, 8 183, — subg. Hebeclada, 272, 274 184-186, 188, 1 — subg. Polygala, 275, 276 — subg. Quercus, 8-18 185, 186, 189 — subg. Orthopolyeals, 275 — sect. Leucobalanus — paniculata, 438 — sect. Pri Polygalaceae, 268, 279, 650 — alba, Polygalaceae in the Southeastern United _ baruensis, 562 tates, The, 2 — candicans, 562 Polygalales, — coccinea, 562 Polypodiaceae, 366, 367 — copeyensis, 562 Polyscias grandifolia, 213, 227 — falcata, 562 — multijuga, 213, 227 — laurifolia, 562 odosa, 213, 227 — oleoides, 562 Polystichum trapezoides, 367 — phellos, 562 Populus trichocarpa, 240, 248, 262 — virginiana, 562 Porospermum michieanum, 213, 228 uiina jamaicensis, 366 Povder-puff ie, 69 apace: ? Prestoea montana, 588 ‘ Prionolejeunea aemula, 440, 455 Raclathris, 110 — aequitexta, 438, 455 wea Ot 650 451 * congener 444, 455 — — helleri. 455 : Radula ener, 444, 451 Protococcus viridus, 97, 104, 106 — inflex: vs Psacadocalymma comatum, 642 ae el — Pseudanabaena catenata, 96 = ee — constricta, 96 es saccatilobs 0, wi Pseudopanax arboreus, 213, 227 ajania, — crassifolius, 213, 227 om stone > A — davidii, 213, 227 sh het Ranunculales, 35, 3 Pseudosciadium, 235 Red maple, 241, ie 744-246, 249, 251, Pseudowintera, 6, 9, 34, 287 253, 254, 258, 61-263 — axillaris, 6, 34, 3 — colorata, 6, 7, 9, 34, 38 Rae ee ee Psilanthele, 646 — phyllobola: — jamaicensis, 646 Rese ee — minor, 646 Reinwaraua, 538 Psychotria balbisiana, 366 esapeeies Janae raginaceae f West — berteriana, 591 Revision ot Kashmit, A, 110-136, 39+ — gua nsis, 591 akistan an Pteroceltis, 524, 539, 550, 557 363, 486-522, ee from Jamaica, — tatarinowii, 575, 58 Reynosia (Rhamn " 364 Ptychocoleus ped toc 437, 45 New ae 9 uerto Rico, Th of an Elf Reger SA, 50s ae Forest in, 14. The Aew of Pico del g. Neoreynosia, Oeste, 86; 15. A Study of the Leafy Bi censis, 364- Hepatic Flora of the Luquillo Mout _ jerugli, 366 76 tains, 435; 16. The Flowering Cycle __ pleiosperma, 213, 226 an Interpretation of its Seasonality, megs ie ’ nail angolensis, 16 Rhamnaceae, ie Pycnolejeunea innovata, 455 Rhapidospora, pee Pyrrophyta, 87, 105 Rhapis one poe 637 Rhinacanthus Pp : Qua — chodatii, 98 Rhinanthoide2®, | me — lacustris, 98 een : Quereas, 160, 161, 177, 180, 136-188, 421, Rhiz rysidales, 8 ? 714 Rhopalodia oe 93, 103 weed, 55 RoBERTSON, KENN The Linaceae in the Southeastern Sore States, 649 Rochelia, 1 a “aso ae 115, 116 — macrocalyx, 112, 118 = oe 113, 118 — rectipes, 113 — sacchara — stellulata, 111, 114, 115 — stylaris, 115 — zeylanica, 344 R 2, — Sceptrum-marianum, 646 Rungia repens, 637 oe Sapindales, 197, 268, 269 Sapium laurocerasus, 437 Sapotac ae, 590 Sarcandra Stange 293 assatras, 25, 2 Saururaceae, po 479 Saururaceae in the Southeastern United e, 479 8 — Sect. Spathium, 484 Sauvagesia erecta, 438, 539 Scarpania portoricensis, 451 cenedesmus arcuatus var. platydisca Schefflera, 214, 226 fae de — bangii, 213, 226 JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Schefflera wey 213, 226 — digitata, khasiana, 213, ge — quinduensis, 213, 226 Schisandraceae, 19, 21, 22, 34-38, 288, 289 mayen 20, 21, 285, 288, 289 —- —— 0, 38 — ts a, 20, sala Schizothrix ao. 96 Sciadodendron sii, 214, 226 34 Scrophularia fluminensis, 635 Scrophulariaceae, 408, 614, 635 Scyphostegia, 290 Scytonema amplum, 96 — burmanicum, 96 02 — iets a o — stuposu nt cabo art 96 Scytonematopsis kashyapii, 96 Securi 269 a, 268, Seemannaralia, 23 Selaginella hispida, 366 re 56 3 66 Shoot Growth and Heterophylly in Acer, e ida carpinifolia, 438 set He Sinop tt tscdenes, 214 cscogghg 235.37, 38 — eggersii, ae - MITH, ALB C. Studies of Pacific Island mand XXIII. The Genus Di- Ospyros (Ebenaceae) in Fiji, Samoa, and Tonga, 36 — Tumicifolius, 134 1971] igen stamineus, 134-136 uneatifo wee 134 -- trict, “i tournefor “aherteraovate Unite States, The Genera of Fagaceae in the Southeastern Seder ron gh The Genera of Myricaceae in the, Southeastern United ie The Genera of Orobanchaceae in the, 404 Southeastern United States, The Genera of the Urticaceae in the, 4 cepa United States, The Linaceae 649 ern chee States, The Poly- galaceae in the, Southeastern sed States, The Saurura- ceae in 479 Southeastern "United States, The Staphy- leaceae in the, Sphaerocystis tis schroeteri, 98 Sphaeroplea —— 98 Sphaeropleal — glaucus, 518 Spirulina subtilissima, 96 SPoNGBERG, STEPHEN. The Staphyleaceae in the Southeastern United States, 196 Squaw root, 420 Staphylea, 196, 199-202 Staphyleaceae, 196, 197 — subfam. Staphyleoideae, 196 — subfam. Tapiscioideae, 19 Staphyleaceae in ig Southeastern United tates, The, 19 Stauroneis mabe 100 STEARN, WILLIAM Taxonomic and Nomenciatural notes on Jamaic Gamopetalous Plants, 614 a perenne, 486 Stephan 137 Ecthonie coma, 642 — pectoralis, pean abel s, 93, 98 Stictolejeunea squamata, 438, 456 “ssa polymorphum, 104 ti — aerugineum, 97, 103 — dendroideum, 102 — hormoides, 97, 102, 103, 106 e, 97 m, 9 Stigonematales, 93, 97, 102, 103, 106 Stilbocarpa lyalli, 214, 228 — polaris, 214, 227 INDEX 715 SToTLeR, RAYMOND, MarcARET FULFoRD, and BARBARA CRANDALL. The Ecology of an Elfin Forest in Puerto Rico, 15. : Study of the Leafy rae Flora of e Luquillo Mountains, 435 Pr Pe 515 Strepsilejeunea involuta, 456 Striped maple, 241, 242, 246, 251, 253, 254, 256, 258, 259, eo Strobila hispidissima ; Studies of Pacific toa Plants, XXIII. The Genus Diospyros (Ebenaceae) in Fiji, Samoa, and Tonga, 3 Sugar maple, 246, 25 Surirella iis 103, 105 Sweet gale, 3 pehndier Ewa M. Comparative Anat- of Ul ae, 523 Synibiexidiuss sartadineeinl 456 — Jaceratum, 439, 4 nsversale, 45 e, 591 ve Symplocos micrantha, 442, pba age 95, 102, 103 Synedrella Syayeell pattotiata, ne 449 rubrocaulis, 444, 449 Tabebuia rigida, 89-100, 439, 591 Taber 615, 616 a, 616 clatural Notes oP ic ‘and Nomen ieee? mn Telaranea woes 447 Temnogame m transeaul, 1 ‘olla gigas, 99 Tetragoniella 1845, 214, 228 Tetrapanax get rpa, 214, 228 hawaie 214, _— kaalae, 214, 228 — Janaens 4, 228 __ micrantha, 215, 228 716 semeneenin. a 215, 228 pkeensis, 215 W. The Genera of Oro- banchaceae in the Southeastern United States, 404 Thrinax parviflora, 366 650 Tolypothrix distorta, 97 Tomunson, P. B., and J. B. FIsHER, Morphological Studies in Cordyline (Agavaceae) I. Introduction and Gen- er orphology, 459 Toralbasia sett obi, 442,589 Touchardia, 40 Peirnctortin 614, 632 — astrotricha, 632, 633 — — Var, astrotricha, 632-634 lomonas ele 100 Trachylejeunea inflexa, , 456 Trema, 524, 534, 541, 559, 565 —am 7 — lamarckiana, 578, 582, a — micrantha, 578, 582 — orientalis, 578, 582. — vieillardii, 578 — virgata, 578 remandraceae, 268 parucni-sapers, 87, 107 rea, 97, 104, 105 — Var, tenuior, 104, 106 — BSP sod 97 —_ torulosa, 104 — umbrina, 97 Trevesia, 2 a 22 Tribonema hppa, 98, 104 oo 8, 104 ratium secre 99, 104 — a a Trichilia apa Trichocolea eve JOURNAL OF THE ARNOLD ARBORETUM [voL. 52 Trichocolea elliottii, 438, 441, 446 46 — — var. amplexicaule, 518 — — var. indicum, 51 — longipedicellatum, 516, 521, 522 oe — 518 — ar. strictum, 518 a oa um, 516 — stocksii, 516, 520-522 ata, 104 Troc ges ahaa 29, 30, 33, 35-37, 39, 291, Tostindendzo, A 285, 291 — aralioides Cartes C., and Ricnarp H. is the Primitive Floral Structure of Araliaceae?, 205 orgie calyptratus, s 219, 226 Turpinia, 196, 197 Two New Species of Leguminosae, 691 Ulmaceae, Comparative Anatomy of, 523 Ulmaceae, 43, von 543, 556, 559-565 — tribe Celte — tribe ines Ulmus, 524, amy 538, 544, 557-566 — alata, 578, 583 — americana, 578, 583, 584 57 oo Rd we she. 578, 583, — parvifolia, 579, 583 — plotii, 579 — procera, 579, 584 1971] Ulmus pumila, 579, 583, 584 — rubra, 579, 583, 584 — scabra, 579 — serotina, 579, 583 — sieboldii, 579 —_ ai ae — suberosa, 579 — thomasi, 579, 583, 584 Ulothrix gai de — variabilis heey 25, 28 — subtribe Euboehmeriineae, 59 Urticaceae in the Southeastern United States, Genera of the, 40 Urticales, 42, 43 Urticastrum, 53 Utricularia, 439 Uvaria, 36 — kirkii, 36 Varronia, 630 Vochysiaceae, 268 Vriesea sintenisii, 100, 101, 588 INDEX 717 Wallenia yunquensis, 590 ata, 441 West Pakistan and Kashmir, A Revision of the Boraginaceae of, 110-136, 334- 363, 486-522, 66 66-690 Wintera, 287 ee 6, 8, 32, 34, 286-289 Woop, CARRO ., JR. The Saururaceae in ‘be Southeastern United States, 479 Woodburnia, 234 Wood-nettle, 53 Xanthophyllae, 268 Xenococcus kerneri, 97 oe cata, 366 ek nitida, 366 Yaxovtev, G. P. Two New Species of Leguminosae, 691 Zelkova, 524; 534, 539, 559-566 — carpinifolia, 5 ms al 99; Gk 104 7 zamena 99; 102 Zygnemops!s Zaaophyllaceae, 268, 680