24ColorCard

zr *

NEW YORI

k ; **s*\ ii ^ /♦ * BOTANICA

( ■**' . . @AJU o«S

Memoirs

i

of

The Torrey Botanical Club

Vol. 21, No. 1 August 29, 1958

Editor

Murray F. Buell

MONOGRAPH OF THE GENUS ARGEMONE FOR NORTH AMERICA AND THE WEST INDIES

By

GERALD B. OWNBEY

Department of Botany, University of Minnesota Minneapolis, Minnesota

t

Published for the Club by The Seeman Printery Durham, N. C.

N )

,1s vz

lAU-t , 3*1

Memoirs of the Torrey Botanical Club

Volume 21 Contents

No. 1 Monograph of the genus Argemone for North America

and the West Indies . \ Gerald B. Ownbey

No. 2 A study of Russula types . L. R. Hosier

No. 3 Studies on Secotiaceous fungi IX the Astrogastraceous

series . Rolf Singer and Alexander II. Smith

No. 4 A monographic study of the fern genus

Cystopteris . Robert F. Blasdell

No. 5 The origin and evolution of ferns :

Introduction . Theodore Delevoryas

The Psilophytes and their relationship to the

origin of ferns . Francis M. Hueber

Putative Devonian ferns . Harlan P. Banks

The woody, fern-like trees of the Devonian . Charles B. Beck

The question of the phylogenetic position of the

Coenopteridales . Donald A. Eggert

Mesozoic and Tertiary fern evolution and

distribution . Chester A. Arnold

Correlation of fern tracheid morphology and evolutionary

divergence . Richard A. White

Evolution in the leaf of living ferns . Rolla M. Tryon

The evolutionary patterns of living ferns

Warren II. Wagner, Jr.

THE TORREY BOTANICAL CLUB

Officers for 1961

President :

Vice-President: James E. Gunckel C or responding Secretary :

Annette IIervey

Recording Secretary :

Matiiilde 1'. Weingartner

Treasurer :

Grace A. Dietz

Editor :

Murray F. Buell

Associate Editor, Torreya:

John A. Small

Frank Lier

Manager of Publications :

Virginia Weadock

Bibliographer :

Clark T. Rogerson Historian :

Harold W. Rickett Member, Hoard of Managers The New York Botanical Garden:

Lawrence P. Miller Chairman, Local Flora Committee : J. Harry' Leiir

Representatives on the Council of the AAAS Peter K. Nelson Pierre Dansereau

Ex-President, 1962: Lindsey S. Olive Ex-President, 1963: Bassett Maguire

Elected Members of the Council 1964-1966

Pierre Dansereau Edwin B. Matzke

Regina Duffy^ Clark T. Rogerson

1963-1965

William Crotty James Gunckel

1962-1964

Ralph H. Cheney Rita McMahon

David E. Fairbrotiiers Donald D. Ritchie

Charles A. Berger Lawrence ( ’rockett

Field Committee

Harold IF. Clum, Chairman Barbara B. Cook David E. Fairrrotiiers Elizabeth C. Hall

( jtEORGE IVALM BAG 1 1 ER

.). Harry Lehr James K. McGrath Robert C. Meyer, Jr. Matiiilde P. Weingartner Eleanor Yarrow

MONOGRAPH OF THE GENUS ARGEMONE FOR NORTH AMERICA AND THE WEST INDIES’

GERALD B. OWNBEY

Department of Botany, University of Minnesota Mi nneapolis, Min n esota

Contents

PAGE

Introduction . 1

Historical Review of the Taxonomic Literature . 2

History of Cultivation . 5

Medicinal Use . 6

Common Names . 6

Geographical Distribution . . . . 7

Generic and Intrageneric Relationships . 8

Species Concept . 9

Speciation . 9

Hybridization . 11

Chromosome Numbers . 13

General Morphology . . . . . . 13

Other Features of Taxonomic Value 19

Acknowledgments . . 20

Herbarium Abbreviations . 20

Systematic Treatment . 21

Key to the Species . 22

1. A. fruticosa Thurb. ex Gray . 27

2. A. mexicana L. . 29

2a. A. mexicana f. •leiocarpa (Greene) Ownb. . 37

3. A. ochroleuca Sweet . 39

3a. A. ochroleuca subsp. ochroleuca . 40

3b. A. ochroleuca subsp. stenopetala (Prain) Ownb. . . 46

4. A. superba Ownb. . 48

5. A. aenea Ownb . 50

6. A. aurantiaca Ownb. . . . . 53

7. A. echinata Ownb. . 56

8. A. squarrosa Greene . 59

8a. A. squarrosa subsp. squarrosa . 59

8b. A. squarrosa subsp. glabrata Ownb . 62

9. A. hispida Gray . 65

10. A. munita Dur. & Hilg. 69

10a. A. munita subsp. robusta Ownb. 73

10b. A. munita subsp. munita 74

10c. A munita subsp. rotundata (Rydb.) Ownb. . . . 77

lOd. A. munita subsp. munita X subsp. rotundata . 82

lOe. A. munita subsp. argon tea Ownb. . 83

1 Publication costs of this paper were met by the Torrey Botanical Club and the

Guggenheim Foundation.

Historical Review of the Taxonomic Literature

The first species of Argemone known to science was A. mexicana which was grown by John Gerard from seeds brought to him from St, Johns Island, West Indies, in 1592. Caspar Bauhin (1596, p. 311) was the first to publish a description of the species, under the name Papaver spmosum. His description was followed a year later by that of Geiaid (159/, p. 993) who illustrated and discussed the species under the name Carduus chry- santhemus peruanus. Tournefort (1694, 1700) established the generic name Argemone in its modern sense. In it he placed only one spe< ies, Bauhin ’s Papaver spinosum which he referred to as Argemone mexicana. Species previously treated under Argemone were transferred by Tourne¬ fort to the genus Papaver.

Linnaeus (1753) included Argemone in his Polyandria Monogynia. He listed three species, viz., A. mexicana , ,4. armeniaca and A. pyrenaica. The genus, as a consequence, was unnatural in the Linnaean sense, since the latter two species according to all subsequent authorities belong to the genus Papaver. Lamarck (1783) re-establish Argemone in the Tourne- fortian sense by referring the alien elements to Papaver (see Poiret, 1804, p. 118). Lamarck (1783) was also responsible for first recording a white flowered “variety” of A. mexicana. This could be none other than the plant subsequently referred to by Lestiboudois (1799) as Argemone alba, a nomen nudum. The same white-flowered species was validly published

as A. albi flora by Hornemann (1815).

The next Argemone to be described was A. alba James (1823), a later homonym of A. alba Rafinesque (1817). A. alba James falls into synonomv

under A. polyanthemos in the present paper.

The next species of Argemone known to science were introduced into Europe by means of seeds from Mexico. These species, all described from cultivated plants, were A. platyceras Link & Otto (1830), A. ochroleuca Sweet (1829), A. grandiflom Sweet (1829), and A. intermedia Sweet (1830). All of these are clearly defined species except A. intermedia which was not illustrated as were the others ; nor were specimens preserved in-so-far as I can determine. As a result, the correct application of this

name has never been properly established.

The next species of Argemone to be described was .4. hispid a Gtay (1849), followed by A. mu nit a Dur. & Ililg. (1854). After this date no new discoveries were made in the genus until Greene described -4. coiym- bosa (1886).

The first comprehensive revision of Argemone is that prepat ed b\ Hi. David Brain based upon specimens in the herbaria of London, I aris and Geneva, published in 1895. This revision is notable primarily lot its

MEMOIRS OF THE TORREY BOTANICAL CLUB

3

exhaustive coverage of the literature up to that time. All later students of the genus should refer to P rain’s monograph in compiling bibliographic data, especially with regard to A. mexicana. Prain’s discussion of the pre-Linnean and later literature is also of great value. Since he was a student of the family Papaveraceae his ideas of generic inter-relationships merit special attention. With regard to Argemone itself, Prain was extremely reluctant to describe new forms. He recognized a total of six species and five varieties, not counting the typical varieties, for a total of eleven named taxa for the world. He did not consider his treatment of the species of Argemone to be definitive in the least, stating (l.c., p. 177- 178) that “the ‘species’ of this paper are in reality aggregations of forms that probably most botanists would recognize as specifically distinct.” Two of the taxa treated by Prain are not considered in the present treat¬ ment, viz., A. alba var. glauca, really a distinct species endemic to the Hawaiian Islands, and A. platyceras var. chilensis , also a distinct species with at least two distinguishable subgroups. Still another distinct South American species with possibly two sub-groups was submerged in A. mexi¬ cana var. ochroleuca by Prain. Of the remaining nine taxa recognized by Prain, only one was new. This was A. intermedia var. stenopetala, cer¬ tainly distinguishable, but more properly treated as a subspecies of A. ochroleuca. A. fruticosa and A. grandiflora were justifiably upheld by Prain, as was A. mexicana which, with the removal of var. ochroleuca, is a homogenous species. A. mexicana var. ochroleuca of Prain, when re¬ stricted to Mexico and Central America and when raised to full species status is probably biologically sound. A. intermedia var. typica Prain was not adequately typified and included a series of forms better treated as distinct species. A. platyceras var. typica Prain included not only A. platyceras L. & 0., but several other taxa better treated as distinct species. A. platyceras var. hispida as circumscribed by Prain includes two com¬ ponents, true A. hispida and A. mu nit a subspp. Finally, A. alba var. typica Prain, exclusive of Texas specimens, is A. alba Lestib., a name which should be replaced by A. albiflora Hornem. From this brief summary, it is evident that, with a few notable exceptions, Prain’s understanding of the species of Argemone is reasonably valid today. Several of the taxa herein described as new were briefly mentioned by Prain, but in the light of his very limited knowledge of these groups he understandably refrained from formally proposing them as new.

During his lifetime, Dr. Edward L. Greene described several species of Argemone. A. corymbosa was described in 1886, A. bipinnatifida, A. gradient a and A. leiocarpa in 1898, A. sanguinea and A. squarrosa in 1899, and A. pleiacantha in 1908. Of these, A. sanguinea is a new name for A. platyceras var. rosea Coulter; A. leiocarpa is the glabrous fruited form

4

GENUS ARGEMONE FOR NOKTH AMERICA AND THE WEST INDIES

of A. mexicana and seems to have been prompted by its mention in Prain’s monograph, and A. bipinnatifida is clearly a synonym of A. hispida Gray. The remaining names were applied to good species previously not distin¬ guished by taxonomists. Greene’s published writings about Argemone are not extensive. His ability to recognize new forms does, however, indi¬ cate a wide familiarity with the species in nature.1

J. N. Rose (1903) published a brief but excellent account of the Mexi¬ can species of the Argemone with which he was familiar. He recognized eleven species for the area, one of which was newly described, viz., A. arida. The specimens referred to A. hispida by Rose are more correctly placed in A. echinata as herein described; those referred to A. rnunita actually belong to A. pleiacantha which was not distinguished until 1908. A. ste- nopehala (Prain) Rose is best left a subspecies of A. ochroleuca. The remaining names taken up by Rose agree in their application with that set forth in the present paper. It is notable that Rose did not apply the name A. intermedia to any Mexican species although this species was reputed to be a native of Mexico at the time of its original cultivation in England.

The most recent revision of Argemone for the world is that of Friedrich Fedde published in 1909 in Das Pflanzenreich. This revision is based very largely upon Prain’s earlier publication, but embodies several advances especially with regard to the incorporation of several of E. L. Green’s proposals. A. corymbosa is given recognition, but as a variety of A. inter¬ media. Similarly A. gracilenta and A. pleiacantha are recognized, but only as varieties of A. platyceras. A. squarrosa is given full species standing. On the other hand, several proposals seem to the present writer to be with¬ out merit. For example, Greene’s A. leiocarpa is given full species stand¬ ing and is divided into two varieties. Probably A. leiocarpa is no more than a form of A. mexicana A. ochroleuca Sweet is left as a variety of A. mexicana regardless of differences which are of specific weight in Argemone. A. purpurea Rose is provided with a description for the first time without recognizing that it is synonymous with A. sanguinea ; A. sanguinea is submerged in A. platyceras. A. platyceras E. & O. is not properly delimited and includes a melange of forms more properly treated as distinct species. A. intermedia Sweet is allowed to stand without ade¬ quate typifieation and includes several distinguishable taxa. Never-tlie-

1 In Greene’s herbarium at Notre Dame University there is a handwritten manuscript entitled “Observations on Argemone,” by Greene. This in actuality is a commentary on Prain’s monograph and evidently was never completed or published. After some critical remarks about the lack of field knowledge of Argemone by Europeans and most Americans the manuscript ends with the words: “Meanwhile let me offer, as an original contribution to the knowledge of the genus, some geographic facts.” No new informa¬ tion of value is included in the completed section of the manuscript.

MEMOIRS OF THE TORREY BOTANICAL CLUB

5

less, Fedde’s contribution to our understanding of Argemone is con¬ siderable and has the merit of being generally conservative. Its deficiencies may be attributed primarily to a lack of adequate study material upon which to base a sounder circumscription of species and varieties. Also, Fedde did not have a first hand knowledge of the distributional aspects of the species of Argemone and much unnecessary lumping of taxa and in¬ correct disposition of individual specimens may no doubt be traced to this factor.

History of Cultivation

The first Argemone to be introduced into Europe was A. mexicana first cultivated by Gerard. On the authority of Aiton (1811), Sweet (1830) and Loudon (1830), all of whom evidently refer back to Clusius, the date of introduction was 1592. The first mention of the species in literature was by Bauhin (1596) and by Gerard (1597), the latter having received seeds of the species from St. Johns Island in the AVest Indies. It has been in more or less continuous cultivation in Europe since its introduction. The next species to be introduced into cultivation appears to have been A. albi- flora , first mentioned by Lamarck (1783) as the white-flowered variety of A. mexicana. The actual date of introduction was prior to 1783, but I know of no account of exactly when or under what circumstances it was introduced into Europe. It was introduced to the gardens at Copenhagen in 1812 according to Hornemann (1815), and into England in 1820 (Sweet, 1830) or 1821 (Loudon, 1830). Tt evidently enjoyed its greatest popu¬ larity in gardens before the introduction of other species between 1827 and 1830. A. ochroleuca was introduced into Europe before 1790 according to Brain (1895, pp. 208, 328) but was not continued in cultivation. Prain’s basis for this statement was a specimen cultivated at Paris in the eighteenth century and preserved in the A. L. Jussieu herbarium. The species was again introduced, into England, in 1827 according to Sweet (1830) and has since remained in cultivation. A. grandiflora , the best of the white- flowered species for gardens, was also introduced into England in 1827 according to Sweet (1830), Loudon (1830) and Prain (1895, p. 333). It has been in cultivation since that date. Another species, A. intermedia , was introduced into England from Mexico in 1828, according to Sweet (1830, 1839). I have been unable to establish the identity of this species, but it almost certainly is not the same as the species of the Great Plains to which the name has been applied for the past sixty years (A. polyanthe- mos of this paper). At any rate, A. intermedia Sweet did not continue in cultivation and seems to be listed nowhere else in the literature of the period. The white-flowered species of the Great Plains, A. polyanthemos, was introduced into England from Colorado in 1877 by J. D. Hooker and

6

GENUS ARGEMONE FOR NORTFI AMERICA AND THE WEST INDIES

was figured by Hooker (1878) under the name A. hispida. it was not continued in cultivation. A. platyceras was introduced into Europe in

1829 according to Link & Otto (1830) and flowered at Berlin in 1829 and

1830 without, however, bearing seeds. Prain (1895, p. 365) states that A. intermedia var. stenopetala (i.e., A. ochroleuca subsp. stenopetala of this paper) was cultivated by Cosson from seeds taken from specimens sent to him by Pringle. Specimens of the garden plants are preserved, according to Prain, in the Cosson herbarium at Paris. The date of cultivation was not earlier than 1885 since this is the year Pringle first collected the subspecies.

The period of greatest popularity of Argemone in the gardens of Europe was 1827-1840 according to Prain (1895, p. 129). In the United States, none of the species of this genus has ever achieved any great popu¬ larity in gardens. A. mexicana and A. grandiflora have been cultivated more or less continuously but never widely. A. platyceras , A. sanguinea, A. munita subsp. munita, A. polyanthemos (as A. intermedia ) and prob¬ ably other species, have also been cultivated occasionally. Species which merit further attention as subjects for gardens are A. sanguinea, A. poly¬ anthemos and A. aenea.

Medicinal Use

The use of Argemone for medical purposes is recorded by numerous authors. In Puerto Rico, Cook and Collins (1903) report that the seeds of A. mexicana are employed as a purgative, as a substitute for ipecacuanha, and that the juice is used in the treatment of ophthalmia. This species and others are similarly utilized throughout Latin America and also for a wide variety of other ailments. The active principles in Argemone appear to be alkaloidal in nature, but very little scientific work on their identification and medicinal uses has yet been done.

Common Names

A. mexicana was referred to by Gerard (1597) as “The golden Thistle of Peru.” The seeds received by him carried the Spanish name of the day “Figue del Inferno.” In Britton & Brown (1913) A. mexicana is termed the “Mexican Prickly or Thorn Poppy,” also “Bird-in-the-bush, Devil’s fig, Yellow, Flowering or Jamaica thistle.” In Jamaica, according to Fawcett & Rendle (1914), A. mexicana is called “Prickly Poppy,” “Gam¬ boge Thistle,” or “Mexican Thistle.” In the Bahama Islands, according to Britton & Millspaugh (1920), A. mexicana is termed the “Donkey this¬ tle” or “Mexican Poppy.” In Bermuda, Britton (1918) states that the names “Mexican Poppy,” “Prickly Poppy,” “Stinging Thistle,” and

MEMOIRS OE THE TURKEY BOTANICAL CLUB

7

“Queen Thistle” are applied to A. mexicana. In Mexico, according to Martinez (1937) the name “Cardo” is applied to A. mexicana as well as to other species; “Cardo Santo” is used for A. mexicana and A. fruticosa ; A. grandiflora is called “Chacalote” or “Chicalota Grande”; “Chicalote” is used for A. mexicana, A. ochroleuca and also Bocconia sp. ; and “Chi- calotl” is used exclusively for A. mexicana and A. ochroleuca. Standley & Steyermark (1946), for Guatemala, list the following names for A. mexicana: “Chicalote,” Cardosanto, “Cajhuoc, “lxmucur, “Kixa- tucan, “Sajouix, and “Cahhouc. In North America north of Mexico the common name “Prickly Poppy” is used for all species of Argemone; when reference is made to one of the yellow-flowered species, usually A. mexicana, it is usually termed “Mexican Poppy.”

Geographical Distribution

Argemone is an exclusively American genus with the exception of one species endemic to the Hawaiian Islands. In North America there are twenty-three species recognized in the present paper. These are found along the coastal areas of southeastern and southern United States to Texas, thence westward to the Pacific and northward across the Great Plains to western South Dakota and southern Montana; in the far west Argemone is found throughout the desert regions and northward across the Great Basin to northern California and southeastern Oregon. Argemone is present throughout Mexico and Central America except in the higher mountains, and throughout the islands of the West Indies. In South America there are four or five recognizable taxa found in Chile, Argentina, Uruguay, Paraguay, Bolivia and northward along both coasts to the Isthmus of Panama. Argemone mexicana is introduced into all tropical and subtropical regions of the world; A. ochroleuca is introduced into Australia. In North America the species of this genus occur almost exclusively in regions of low rainfall at elevations of near sea level to 8,000 feet, rarely more; when they occur in areas of moderately heavy rainfall they are found only in soils of low water-holding capacity. Many species spread rapidly into newly disturbed areas; many are incipient weeds in fields and about dwellings within their natural distributional areas and in adjoining regions. For this reason the absolute distributional boundaries of many species before the introduction of modern agricultural practices is difficult if not impossible to establish.

Man has played an important role in the dispersal of the seeds of Argemone in the past. This role has probably been passive in most in¬ stances; either the seeds have been carried along as contaminants in the seed of crop plants of a similar size or the seeds have been mixed with

8

GENUS ARGEMONE FOR NORTH AMERICA AND THE WEST INDIES

other materials transported from place to place ; in the case of A. mexicana the seeds probably have been carried from port to port in ship’s ballast. In the southwest and in Mexico Argemone is often closely associated with cultivated fields and dwellings of man. It is avoided by all livestock, and because of the prickliness of its stems and leaves survives successfully where most other plants would be eliminated. Man has undoubtedly played an active role in the dispersal of A. mexicana and A. ochroleuca because of the supposed medicinal properties of these plants previously mentioned.

The usual natural forces of wind and water also are effective in the dispersal of the seeds of Argemone. Certain features of the seed seem to the writer to be especially important in their dispersal by water. First, the seeds are covered by a thin coat of wax which inhibits the absorption of water; then, the shallow pits in the seed coat appear to act as air traps when the seeds are placed in water ; finally, in the seeds examined, there was a small air pocket under the seed coat beneath the micropyle. The first of the features prevents the immediate water-logging of the seed when it comes into contact with water ; the latter two features add to its buoy¬ ancy. Taken together they would appear to enhance the chances of water transport of the seeds under favorable conditions. When the seeds of Argemone are placed in water in the laboratory, they float on the surface for one to several days.

Generic and Intrageneric Relationships

Argemone is set apart from other genera of the Papaveraceae by several important features including tin1 basically trimerous arrangement of the sepals and petals, the presence of sepal horns, the united styles and the mode of dehiscence of the capsules. None of these features taken indi¬ vidually is peculiar to Argemone but their presence in combination in Argemone forms the basis for the taxonomic delimitation of the genus. For example, trimery and horned sepals are found in species of Papaver ; capsules of Meconopsis and Romneya dehisce much as they do in Argemone; united styles are found in Meconopsis. Argemone is placed in the tribe Papavereae by Fedde (1909) along witli Meconopsis, Papaver , Canbga, Cathcartia, Roemeria and Glaucium. The genus appears to be most closely related to Papaver and Meconopsis.

Some effort has been made without success to discover groupings of species in Argemone which might form the basis for establishing sections within the genus. There is no evidence for the existence of groupings of this magnitude. Some poorly definable alliances among species may, how¬ ever, he pointed out, but there are many species which cannot In* assigned with any assurance to any alliance, nor do the alliances themselves appear

MEMOIRS OF THE TORREY BOTANICAL CLUB

9

to have any high degree of reality. For example, A. mexicana and A. ochroleuca and possibly also A. aenea and A. superba form one alliance; A. cchinata, A. squarrosa, A. aurantiaca, A. pleiacantha, A. hispida and A. munita, in part, and possibly A. platyceras form another alliance; A. albiflora, A. polyanthemos, A. grandiflora and A. gracilenta form still another alliance. Several species such as A. fruticosa, A. corymbosa and A. sub intergri folia seem to have no very close relatives; the remaining species do not fit well into any of the above alliances, yet are close to one or more of them. It is probably unwise to attempt any taxonomic sub¬ division of Argemone above the rank of species.

Species Concept

In delimiting the species of Argemone, I have been guided by the follow¬ ing generalizations. When two taxa are sympatric over a part of their ranges and evidence indicates that they hybridize readily in this area thev are generally treated as subspecies of the same species. If there is no visual evidence of hybridization between the two taxa they are treated as distinct species. When the ranges of two taxa are allopatric but experience with greenhouse cultures indicates that they may be crossed readily and produce rather highly fertile offspring they are also generally treated as subspecies of the same species. If two taxa occupy allopatric ranges and consequently are effectively isolated spatially and, further, if their crossability is un¬ known, then they are treated as distinct species.

Differences in ploidy level between individuals of different breeding populations or of those within a single population have not been considered to be of nomenclatorial relevance unless accompanied by morphological differentiation. Such differences in ploidy level may obviously be of great biological significance in their effects on the breeding behavior and evolu¬ tion of the species. It happens, however, that in Argemone the ploidy level is constant in most species insofar as has been determined ; conversely, differences in ploidy level are generally accompanied by morphological differentiation of sufficient magnitude to warrant a degree of taxonomic recognition.

Speciation

The present study indicates that the most important factor leading to speciation in Argemone is geographic isolation followed by the accumula¬ tion of genetic differences in the isolated populations. One of the most interesting features of the genus is the manner in which one species is replaced by another geographically. In the United States, A. polyanthemos is replaced to the west and southwest by A. hispida and A. squarrosa subspp.

10

GENUS ARGEMONE FOR NORTH AMERICA AND T1IE WEST INDIES

and these are replaced by A. gradient a, A. corymbosa subspp. and A. munita subspp. still farther west. To the south and southeast, A. poly- anthemos is replaced by A. chisosensis, A. albiflora subsp. t ex ana followed by A. sanguinea and A. aenea. Other species occupy smaller but discrete ranges within this larger area. There is, of course, some overlapping of distributional ranges, but these are more apparent than real. This is be¬ cause the two partially sympatric species occupy different ecological posi¬ tions. In Mexico, similar conditions to those in the United States prevail. All of the Mexican species have partially or wholly discrete geographical ranges.

Many diploid species such as A. albiflora and A. poly ant hemos appear to maintain their identities mainly because of geographic isolation. Breed¬ ing experiments indicate that these two species will cross readily under experimental conditions. A. gradient a , a diploid species mainly of the Sonoran Desert, is perfectly distinct over most of its range, but appears to hybridize with A. pleiacantha subsp. pleiacantha where the latter has been introduced into its range.

Another factor which has played a role of some significance in Arge- mone speciation is polyploidy accompanied by reproductive isolation. It is postulated that A. ochroleuca may have arisen as an autotetraploid of A. mexican a. These species are very much alike in general characteristics and reproductive behavior. Plants of both species are highly self-com¬ patible, and both species exhibit a high degree of crossability with other species. A. mexicana now occupies a very different geographical range from A. ochroleuca (see Pig. 5), and the occurrence of the former in the range of the latter may be attributed entirely to introductions. Diploid and tetraploid races herein included under A. .sanguinea occupy slightly different though adjacent geographical areas and exhibit some inconstant morphological differences. This type of variation might easily lead to speciation. Two species which are effectively isolated from all others are A. aurantiaca , a hexaploid, and A. squarrosa subsp. squarrosa , probably an octaploid. These two species may be crossed with diploid and tetraploid species under experimental conditions, but the resultant highly unbalanced genomes in the Pi ’s lead to complete sterility of these hybrids.

The presence of sterility barriers between individual species pairs in Argemone are not uncommon, although the exact nature of these barriers has not been investigated. For example, A. platyceras has been success¬ fully crossed with only one other species to date, namely, A. mexicana . but attempts to cross it with its nearest relative, A. arida , have not been made. Once sterility barriers have formed between segments of a population, even though they are not complete, the potentiality for genetic drift, either morphological or in the direction of different ecological amplitudes (i.e..

MEMOIRS OF THE TORREY BOTANICAL CLUB

11

differing habitat optima), or more likely both, may lead to speciation. Probably this factor is involved in speciation within the genus Argemone. It is unlikely, however, that sterility barriers have played the primary role.

Hybridization

'Well authenticated instances of natural hybridization in Argemone are not numerous. This is partly due to lack of opportunity for hybridization since over most of the geographical range of the genus the species are spatially or ecologically isolated from each other. .Well authenticated instances of putative natural hybridization, however, do occur, the best example between A. munita subsp. munita and A. mu nit a subsp. rot undata in southern California in a small area of sympatric distribution where hybrid swarms of the two are common. Curiously enough, plants of inter¬ mediate character occur far up the coastal ranges where neither subspecies is found in typical state. This is the main basis for placing these two taxa in the same species, the two being otherwise sufficiently distinct and sepa¬ rated geographically to justify treating them as separate species. In Sonora, single hybrid plants of A. pleiacantha subsp. pleiacantha X A. ochroleuca subsp. ochroleuca are not uncommon, but these always appear to be Fj hybrids which do not reproduce. The latter species is almost certainly introduced into the area where the hybrids occur. .4. gracilenta appears to hybridize with A. pleiacantha subsp. pleiacantha and A. pleia¬ cantha subsp. ambigua in southern and western Arizona in and near the areas of distributional overlap. A. polyanthemos and A. albiflora subsp. texana seem to hybridize freely in north-central Texas. In southern Texas several introduced and native argemones are found in more or less close proximity. The variation patterns of such species as A. sanguinea (diploid and tetraploid) and A. albiflora subsp. texana (diploid) in this area are not understood; part of this variation may be due to genetic interchange between these and other species. Further studies on the southern Texas argemones are much needed.

The classical work on controlled interspecific hybridization in Argemone is that of Vilmorin (1912). Vilmorin succeeded in crossing A. mexicana and A. platyceras and in growing first and second generation progeny. From the distribution given for A. platyceras, i.e., Rocky Mountains: Wyoming, Nebraska, Texas, one might suspect that Vilmorin ’s A. platy¬ ceras was actually A. polyanthemos. For reasons clarified below, 1 am of the opinion that he was working with true A. platyceras from Mexico. Vilmorin reported segregation in the F2 for flower size and color, stem color, latex color, stature, abundance of foliage, color of the nerves, etc. Monstrosities resulting from malformations of the floral organs were also

12

GENUS ARGEMONE FOR NORTH AMERICA AND THE WEST INDIES

reported, e.g., stamens transformed into carpels. Most important, he reported that some F2 plants had flowers with petals of varying shades of buff. Since one parent had bright yellow petals and the other white, recombinations giving a new color are of extraordinary interest.

I have repeated Vilmorin’s experiments exactly with parental stocks of known origin and have obtained identical results. I have, unfortunately, been unable to grow large populations of F2 plants, having succeeded in rearing to maturity a total of only 36 plants. Only one plant had “buff” (actually bronze-colored) petals and, interestingly, this plant was identical to A. aenea of Texas and northeastern Mexico. This experiment should be repeated for absolute confirmation with large populations, and, if A. aenea- like progeny are recovered, a further step should be taken, that is, selection in succeeding generations with a view to reconstituting .4. aenea. I am not at all sure that this cross is the one actually involved in the evolution of A. aenea, although A. mexicana is almost certainly one of the parents. The other parent is more likely a lavender-flowered, diploid race of A. san- guinea. The bronze color of the petals in /l. aenea is possibly resultant from the presence of both purple and yellow pigments. Although lavender filaments are generally present in A. platyceras, the coloration does not carry over to the petals in-so-far as is known.

Numerous other attempts at controlled crossing of Argemone species have been made by the writer in the greenhouses at the University of Minnesota. The species most often involved in these crosses have been A. mexicana, A. ochroleuca subsp. ochroleuca, A. platyceras, A. polyanthemos, A. albiflora subspp., A. aenea, A. grand i flora subspp., and A. sanguinea. Publication of details of these experiments must be deferred to a later date, but some generalities may be mentioned here. Experimental hybridi¬ zation between many species of Argemone is possible, and no species seems to be completely isolated from all other species by sterility barriers. A. platyceras is most nearly isolated ; crosses of this species were successful only with A. mexicana- among several species tried. A. mexicana and A. ochroleuca can be successfully hybridized with most if not all other species. Interspecific crosses can be carried beyond the first generation only if the ploidy levels of the parental species are the same. Numerous first and second hybrid generation populations were grown in the experimental gar¬ dens of the University of Minnesota and the Rancho Santa Ana Botanic Gardens. In general, a marked loss in vigor and fertility in F2 plants was evident. This indicates the prevalence of incompatible genomes in Arge¬ mone species even among those which are erossable under experimental conditions.

MEMOIRS OF THE TORREY BOTANICAL CLT7B

13

Chromosome Numbers

Chromosome numbers have been determined for the species listed in the table below. Root tip sections have been employed in making most of these counts. In A. corymbosa, A. gracilenta, A. mu nit a and A. pleia- cantha the counts have been made from PMC squashes.

It is evident from the table that most species of Argemone have a somatic chromosome number of 28. The polyploids form a nearly regular series with known somatic chromosome number of 56 (4 species), 84 (1 species) and ca. 112 (1 species). A meiotic number of 14 is recorded for two putative hybrids between A. ochroleuca and A. pleicantha. This is at variance with the expected situation in an Fi cross between these two species, the former a tetraploid, the latter a diploid. Herein lies indirect evidence that A. ochroleuca may sometimes have a somatic chromosome number of 28. Pairing of the chromosomes at Metaphase-I was nearly regular in the hybrids.

The chromosome counts listed below have all been made by myself over a period of several years. The only other counts for this genus known to me are those reported by Sugiura. Sugiura (1936, p. 570) reports the meiotic number for A. grandiflora as 28 and that for A. mexicana as 14. These counts are in agreement with my own studies. Sugiura (1940, p. 564) further reports the meiotic number for A. barclayana as 14. I do not know which species is referred to by this name, since in my opinion the binomial falls into synonomy under A. ochroleuca xubsp. ochroleuca, a tetraploid with 56 somatic chromosomes over most of its range, with the possible reservation noted above. I agree with Sugiura (1940) that the basic chromosome number in the genus may be 7, but it now appears doubtful if any species with a meiotic number of 7 survive.

General Morphology

Roots: Seedlings of Argemone initially produce a strong tap root in all species observed, hi certain species such as A. polyanthemos, when grow¬ ing in sandy soil, the tap root may penetrate to a depth of two or more feet. In other species large lateral roots soon develop and the root system is consequently more diversified.

Stems: In all of the herbaceous species of Argemone the young plants first produce a basal rosette of leaves. A main stem then arises from the apex and often one or more additional stems arise concurrently from axil¬ lary buds near the base of the plant. Often, however, the axillary buds do not become active until after growth of the central stem has ceased and may even be delayed until the advent of the next growing season.

14

GENUS ARGEMONE FOR NORTH AMERICA AND THE WEST INDIES

Table I

Collector and

Locality where

Garden

Species

number

collected

number

n

2n

A. aenea

G. B. Ownbey F.

17.5 mi. n.w. of

A -50- 2

28

Ownbey 1 409

Montemorelos,

Nuevo Leon, Mexico

A -50- 5

F. G. Meyer A D. J .

54 mi. s. of Mata-

A-12- 1

28

Rogers 2479

moros, Tamaulipas, Mexico

A. albiflora subsp.

G. B. Ownbey A- F.

Savannah, Chatham

A -35- 5

28

albiflora

Ownbey 1350

Co., Georgia

G. B. Ownbey & F.

Carrabelle, Franklin

A