/981
BIOSYSTEMATIC STUDY OF A DESMODIUM COMPLEX
THESIS
Presented to the Graduate Council of the
North Texas State University in Partial
Fulfillment of the Requirements
For the Degree of
MASTER OF ARTS
By
John G. Williams, B.A.
Denton, Texas
December, 1977 Williams, John G., Biosystematic Study of a Desmodium
Complex. Master of Arts (Biology), December, 1977, 70 pp.,
13 illustrations, 7 tables, 5 appendices, bibliography, 49 titles.
An examination of the Desmodium canescens complex
(D. canescens; D. tweedyi; D. illinoense) has resulted in the delimitation of a previously unreported alliance between
D. canescens and D. tweedyi. The following points support this view: (a) morphological data taken from herbarium and garden specimens indicate that for many characters, the mean values of D. canescens and D. tweedy are not significantly different (b) breeding experiments have shown that artificial interspecific hybridization is possible between D. canescens and D. tweedyi (c) cytological studies have shown that D. canescens and D. tweedyi have a base number of x = 11, while
D. illinoense has a base number of x = 10. A new combination is suggested: Desmodium canescens var. tweedyi (Britt.) Williams. TABLE OF CONTENTS
Page LIST OF TABLES...... iv
LIST OF ILLUSTRATIONS...... v
Chapter
I. INTRODUCTION...... 1
II. MATERIALS AND METHODS...... 11
Morphological Evidence Herbarium Studies Field Garden Studies Reproductive Biology Preliminary Studies of Breeding Behavior Hybridization Studies Flowering Behavior Cytological Evidence Chromosome Counts Palynological Evidence Pollen Morphology and Size
III. RESULTS AND DISCUSSION...... 25
Morphological Evidence Herbarium and Field Garden Studies Genetic Evidence and Repro- ductive Biology Cytological Evidence Palynological Evidence
IV. SUMMARY AND CONCLUSIONS ...... 54
iii LIST OF TABLES
Table Page I. The Number of and Geographical Distribution Pa of Herbarium Specimens Examined . . . 15
II. Herbarium Specimens from Which Seeds and Flowers Were Removed for Study . . . 17
III. A Comparison of Nine Qualitative Characters . 24
IV. Summary of Preliminary Studies of Reproductive Patterns .*...... 4)4
V. Fruit Set and Seed Set Following Intraspecific and Interspecific Hybridizations . . . . 45
VI. Number of Days from Planting to First Visible Flowers...... 47
VII. Flowering Behavior of the Desmodium Species ...... 48
iv LIST OF ILLUSTRATIONS
Figure Page
1. Distribution of Desmodium canescens Herbarium Specimens Used in Morphological Analysis. . 12
2. Distribution of Desmodium tWeedyi Herbarium Specimens Used in Morphological Analysis. . 13
3. Distribution of Desmodium illinoernse Herbarium Specimens Used in Morphological Analysis...... 14
4. Ranges of the Desmodium canescens Complex Species ...... *...26
5. Leaflet Markings on a D. tweedyi Leaflet . . . . 28
6. Mean, Range and Standard Deviation for Selected Characters of Desmodium canescens Complex Series...... 33
7. Mean, Range and Standard Devision for Selected Characters of Desmodium canescens Complex Species...... 35
8. Mean, Range and Standard Deviation For Selected Characters of Desmodium canescens Complex Series...... 36
9. Mean, Range and Standard Deviation for Selected Characters of Desmodium canescens Complex Series ...... 38
10. Mean, Range and Standard Deviation for Selected Characters of Desmodium canescens Complex Series ...... 39
11. Analysis of Variation Within the Desmodium canescens Complex by Means of a Scatter Diagram ...... 41
12. Analyses of Variation within the Desmodium canescens Complex by Means of a Scatter Diagram ...... 42 v LIST OF ILLUSTRATIONS --'Continued
Figure Page 13. Camera Lucida Drawings of Meiotic Chromosomes Figures...... 50
vi CHAPTER I
INTRODUCTION
Desmodium Desv. is a polymorphous genus that includes more than two hundred species of annual or perennial herbs,
shrubs, and small trees. The genus, which is a member of
the Leguminosae, is widely distributed over the temperate
and tropical regions of both hemispheres, with the excep-
tion of the following areas: western United States (Paci-
fic slope), continental Europe., and New Zealand (20).
As in the case with most large genera, the taxonomic
unity of Desmodium has been subjected to question. Although
Desmodium Desv. was conserved over the older Meibomia Heist. by the International Botanical Congress of Vienna in 1905,
the German authority Schindler (25) has maintained the use
of both names by segregating several genera from Desmodium proper, thus dividing the genus into Desmodium and Meibomia.
Following this interpretation, all of the North American species of Desmodium would belong to the genus Meibomia, except Desmodium glutinosum (Willd.) Wood, D. nudiflorum (L.) DC., and D. pauciflorum (Nutt.) DC. (21). However, subsequent authors have not followed Schindler, but rather have adhered to the more traditional classification--allow- ing Desmodium to circumscribe all the species (2,10).
1 2
To avoid confusion, it should be noted that Schindler's usage of Meibomia differs from that of American authors who,
following the American code of nomenclature, did not recog-
nize the action of the Vienna Congress and therefore con-
tinued to publish the name Meibomia instead of the conserved name Desmodium( 24,7).
All of the American species of Desmodium are herbaceous
perennials. Most of them possess thick, woody crowns and
a ligneous, perennial root system. Most of them grow in light, sandy soil and are usually found in open woodlands or
at the edge of more densely shaded areas. Like many other
native legumes they may act as pioneers in the revegetation of denuded, burned over, or eroded areas, gradually disap- pearing as the vegetational cover, particularly that of
shade plants, increases (8).
In the United States the common names beggar's-tick,
stick-tight's, beggar-weed and tick-trefoil are variously applied to members of the genus Desmodium. All of these names take their origin from the tendency of the mature loment segments to adhere to clothing or to the hair of animals.
In a phylogenetic sense, two groups may be distinguished among the American species of Desmodium (21). Except for three species, the eastern and central representatives of the genus are allied with southwestern and Mexican species.
The center of distribution of the complex may be tropical 3
America, where there is perhaps the largest number of spe-
cies with the most diverse development of plant parts (22).
The other phylogenetic group, composed of Desmodium gluti-
nosum (Willd.) Wood, Desmodium nudiflorum (L.) DC., and Des- modium pauciflorum (Nutt.)DC. is of Asiatic affinities (10,
25). These species differ from their American congeners
in a number of morphological relationships, including the
nature of the calyx, presence of monadelphous stamens, obso-
lescence of stipules, and characters of the laments and seeds.
Schubert (21,22) has grouped the American species of
Desmodium into 'Series' for the purpose of making clear the presumed taxonomic relationships of the associated taxa.
Schubert believes that these Series represent "perfectly natural species-groups" (21,p. 135). One such Series, Sti- pulata Schub., is characterized by species which have con- spicuous ovate-attenuate stipules which are chordate or semi-chordate at the base. The related species in this group include Desmodium rotundifolium DC.., Desmodium ochro- leucum M. A. Curtiss, and species of the Desmodium canescens complex: D. illinoense A. Gray, D. tweedyi Britt. and D. canescens (L.) DC.
Within the series Stipulata, Desmodium rotundifolium and D. ochroleucum are morphologically distinct and taxono- mically discrete. The interrelationships of the species in the D. canescens complex, however, are vague. The 4
taxonomic status of D. tweedyi and its relationships with
D. illinoense and D. canescens has never been clearly es-
tablished. In 1950 Schubert, the foremost student of the
genus Desmodium, reduced D. tweedyi to synonomy with D.
illinoense, thus reflecting the opinion that there are no
real differences between the two taxa (2).
Isely (10), the preeminent student of the American
leguminosae, indicates (in a reference to D. tweedyi) that
"a Desmodium which appears associated with D. illinoense
occurs in east-central Texas." Isely does not recognize
the specific status of D. tweedyi, nor does he indicate
his endorsement of Schubert's reduction of the taxon to
synonomy with D. illinoense. He concludes his remarks on
these Desmodium species by saying that their relationships
deserve further study.
Regional floras covering the ranges of the species
in the D. canescens complex reflect the vagueness of the taxonomic status of D. tweedyi (2, 7, 12, 20, 24, 26, 27,
28, 31). Jones and Fuller (12) extend the range of D.
illinoense into Texas by following Schubert and consider-
ing D. illinoense and D. tweedyi as synonyms. Shinners (26) however, recognizes D. tweedyi at the specific level,
though noting that "it has been referred to the more nor- thern D. illinoense. " Shinners does not include any jus- tification for his taxonomic treatment. Turner (31) also 5 recognizes D. tweedy at the specific level. With refer- ence to the tweedyi-illinoense confusion, Turner (31, p.
214) writes that "....the two taxa appear distinct, both morphologically and geographically." Steyermark (28) in disagreement with Turner, again treats D. tweedyi as a synonym of D. illinoense.
In the most recent publication dealing with these spe- cies, Schubert (20) rejects Steyermark's treatment (and indeed her own 1950 treatment) and lists D. tweedyi as one of the twenty-six Desmodium species which occur in
Texas. Schubert notes, however, that "D. tweedyi seems to bridge the gap between D. canescens and D. illinoense, and is doubtfully distinct from the latter species" (20, p.
862). On clarification of this point, Schubert has writ- ten that ". .D. illinoense and D. tweedyi are very close
perhaps D. tweedyi is not a good species" (23, perso- nal communication). In a reappraisal of his 1955 observa- tions, Isely (11) views the relationships of canescens- tweedyi-illinoense entirely differently. He recognizes the similarities between the three, but concludes that D. tweedyi most closely resembles D. canescens. He does not however, make any decision as to where D. tweedyi should be placed taxonomically.
Since the relationships of the species of the Desmo- dium canescens complex have been subjected to several in- terpretations in recent literature, it seems desirable to 6
obtain data elaborating the nature of the morphological variability of this group. An improved understanding of this complex might also be obtained through a biosystema- tic study utilizing data from several disciplines. In the past knowledge of the presumed relationships among the spe- cies of Desmodium has been based entirely on comparative morphological studies (3, 8, 9, 13, 21, 22, 29). No bio- systematic studies have been made in the genus. Studies dealing with the embryology (16) and anatomy (14) of Des- modium have been made, but not from the standpoint of dis- cerning phyletic relationships. The genus has also been almost totally ignored in biochemical studies, even though it is a member of a family noted for its chemotaxonomic literature (32). The only limited chemical studies of the genus have been made in efforts to extract medicinally im- portant compounds (4, 5, 6, 30). Information concerning the cytology, reproduction biology, and palynology of the genus is also very incomplete, although as the source of a potentially valuable forage crop, several Desmodium species have been studied genetically (1, 15, 17, 18, 19). The following will be sources of data for this biosystematic study of the Desmodium canescens complex: (a) Morphological Evidence - based on herbarium spe- cimens from regional herbaria and on plants grown
under uniform conditions 7
(b) Genetic Evidence and Reproductive Biology - inter-
specific and intraspecific hybridizations and ob-
servations of the flowering behavior for each spe-
cies.
(c) Cytological Evidence - meiotic chromosome counts
for each species
(d) Palynological Evidence - morphological comparisons
of the pollen of each species
From these sources, the following objectives will be met:
(1) To accurately determine the range distributions
of each species in the D. canescens complex
(2) To examine the morphological variation in the
complex
(3) To provide the first reported chromosome counts
for species of this complex
(4) To provide the first reported description of pol-
len morphology for the genus Desmodium
(5) To provide the first reported study of the repro-
ductive biology and flowering behavior for any
North American Desmodium species
(6) To correlate biosystematic data and taxonomically
delimit the elements of the Desmodium canescens
complex. CHAPTER BIBLIOGRAPHY
1. Crowder, L.V., Hybridization of Desmodium canum (Gmel.) Schin, and Desmodium uncinatum (Jacq.) DC., Plant Breeding PaperNo.T3l, Cornell University, Ithaca, New York, 1972.
2. Fernald, M.L., Gray's Manual of Botany, 8th ed., New York, American Book Company, 1950.
3. Fosberg, F.R., "Critical Notes on Pacific Island Plants," Micronesica, 4(1968), 255-259.
4. Ghosal, S., "Beta-phenethylamine, tetrahydroisoquinoline and Indole Alkaloids of Desmodium tiliaefolium," Phytochemistry, 12(1973), 193-197.
5. Ghosal, S.., "Chemical and Pharmacological Evaluation of D. pulchellum," Planta Medica, 21(1972), 398-409.
6. Ghosal, S., "Desmodium Alkaloids. II. Chemical and Phar- macological Evaluation of D. gangeticum," Planta Medica, 22(1972), 434-440.
7. Gleason, H.A., The New Britton and Brown Illustrated Flora of the NortheasteinUnited States and Adjacent Canada, Vol. 2: 424-432, Lancaster, Pennsylvania, Lancaster Press, 1952.
8. Isley, Duane, "Desmodium paniculatum (L.) DC. and D. verdiflorum TL.) DC., American Midland NaturalisF, 49 71953), 920-933.
9. Isley, Duane, "Desmodium: Section Podocarpium Benth.," 7(1951), 185-247
10. Isley, Duane, "The Leguminosae of the North-Central United States. II. Hedysareae," Iowa State College Journal of Science, 30(1955), 33-118.~
11. Isley, D., unpublished manuscript, Papilionoid Legumes of the United States, Department of Botany, Iowa StEate University, Ames, Iowa, 1975.
12. Jones, G.N. and G.D. Fuller, Vascular Plants of Illinois, Urbana, University of Illinois Press~,~~1955.
8 9
13. Knapp-Van Meeuwen, M.S., "Preliminary Revisions of Some Genera of Malaysian Papilionaceae, V.A. Census of the Genus Desmodium," Reinwardita 6(1962), 239-276.
14. Moore, J.A., "The Vascular Anatomy of the Flower in the Papilionaceous Leguminosae," American Journal of Bo- tany, 23(1936), 279-290.
15. Park, Soon Jai, "Genetic Studies in Spanish Clover, D. sandwicense. I. Inheritance of Flower Color, Stem Color and Leaflet Marking," Crop Sciences, 8(1968), 467-470.
16. Rau, M. Anataswamy, "The Development of the Embryo Cya- mopsis, Desmodium and Lespedeza with a Discussion on the Position of the Papilionaceae in the System of Embryogenic Classification," Phytomorphology, 4(1954), 418-430.
17. Rotar, P., Crossing and Flowering Behavior in Spanish Clover, D. sandwicense E. Mey. Technical Progress Report No. 164, Hawaii Agricultural Experiment Sta- tion, University of Hawaii, November, 1967.
18. Rotar, P., Morphological Variation and Interspecific Hy- bridization Among D. intortum and D. uncinatum, Tech- nical Bulletin No.~82, Hawaii Agricultural Experiment Station, University of Hawaii, January, 1971.
19. Rotar, Peter, Some Agronomic Observations in Desmodium Species: Seed Weights,Tchnical Progress Report No. 147, Hawaii Agricultural Experiment Station, Uni- versity of Hawaii, June, 1966.
20. Rydberg, P.A., Flora of the Prairies and Plains of Cen- tral North America, New York, New York Botanical Garden, 1932.
21. Schubert, Bernice G., in Manual of the Vascular Plants of Texas, by Donovan S. Correll and Marshall C. John- ston, Renner, Texas, Texas Research Foundation, 1970.
22. Schubert, B.G., "Desmodium: Preliminary Studies. III," Rhodora, 52(1950T~135-155.
23. Schubert, B.G., "Desmodium: Preliminary Studies. IV." Journal of the Arnold Arboretum, 44(1963), 284-297.
24. Schubert, B.G., personal communication, Arnold Arboretum, Harvard University, Cambridge, Massachusetts, October 4, 1973. 10
25. Schindler, A.K., In: Fedde, Repert' Spec. Nov. Beihefte. Bd. 49. Die Desmodiinen in der botanischenLiteratur nach Linne, i928 26. Shinners, L.H., Spring Flora of the Dallas and Ft. Worth Area, DallasPrestige Press, 1-5.-
27. Small, J.K., Flora of the Southeastern United States, Lan- caster, Pennsylvania, New Era Printing Co., 1903.
28. Steyermark, J.A., Flora of Missouri, Ames: Iowa State University Press,~1963.
29. Taubert, P., in Die Naturlichen Pflanzenfamilien, by A. Engler and K. Prantl, III - 3: 70-396, 1894.
30. Tiwari, R.D., "Physcion 1-glycosyl Rhamnoside From Seeds of D. pulchellum," Phytochemistry, 10(1972), 1922- 1929.
31. Turner, B.L., The Legumes of Texas, Austin, The Univer- sity of Texas Press, 1959.
32. , Chemotaxonomy of the Leguminosae, New York, Acade- mic Press, 1971.
33. Young, J.O., "Cytological Investigations in Desmodium and Lespedeza," Botanical Gazette, 101(1947T,~33~ 850. CHAPTER II
MATERIALS AND METHODS
Morphological Evidence
Examinations of morphological characters and charac- ter variations in the Desmodium canescens complex (D. canescens, D. tweedy, and D. illinoense) were based on evidence obtained from both herbarium and field garden spe- cimens. Herbarium specimens provided a basis for studying characters from a large number of plants collected from across the north-central Texas and central United States ranges of the taxa (see Figures 1,2,3, pages 12-14). Plants grown from seed and maintained under the uniform environment of a field garden provided a basis for more precisely exa- mining the diagnostic characters used in the identification
of the species of the canescens complex and in addition were utilized as experimental material in investigations of
their reproductive biology.
Herbarium Studies
Herbarium specimens of each species in the Desmodium
canescens complex were examined at the following herbaria:
SMU, UT, TAES, and OU. A total of 198 specimens were stu-
died (See Table I, page 15). A comprehensive character
analysis sheet was completed for each specimen (Appendix I,
11 12
000
000
0.0 0 0 .0
0 0
0 0
Figure 1. Distribution of Desmodium canescens herbarium specimens used for morphological analysis. '3
0 0
Figure 2. Distribution of Desmodium tweedy herbarium specimens used for morphological analysis. 14
Figure 3. Distribution of Desmodium illinoense' herbarium specimens used for morphological analysis. 15 to
Sp l o 4-) 0 0 o co H c --I- t- H HEO
4O -H mq o (>-o HsO
H c ) LC\ H 0 HI
-) to OH HH\ 0 Cr
o C 0 -3 0
toE -0 H H E HD -C EH H E-1 S0 0 C cj .H , H H CM H
0
p H, 0-4 2l -r N 0 H H H Cm) H (D (D
H0 o H C Z CM -z1
Z (-'3 ro
Ca 0 o *Ho 'H C) O -H 0 O XO 0) S Cd4-D9 0 U) to C SU)- C 0 H o cd 3 H O p -H dH E O HD * . - 16 example form). Quantitative characters measured include
length and width of stipules, length of petioles, leaflet
length and width, length of stipels, length and width of the primary and secondary bracts, length of the calyx lobes,
length of the stipe, number of articles, length of articles,
and width of isthmi. Small characters were measured by means of a dissecting microscope reticle. Qualitative cha-
racters examined include pubescence of: the inflorescence
rachis, primary and secondary bracts, pedicels, loments,
stems, stipules, petioles, and leaflets; shape of stipules,
leaflets, primary and secondary bracts, and articles; and
persistence of stipules and primary and secondary bracts.
All of these characters have been utilized in various keys
in the separation of Desmodium canescens, D. tweedyi, and
D. illinoense (4, 5, 8, 10, 12).
Field Garden Studies
Under the advice of Dr. Bernice G. Schubert to "grow
these taxa /D. canescens complex7 and see what their real
characters are and how important they are in distinguish-
ing species" (9), several loments were obtained from recent
herbarium accessions of D. canescens, D. tweedyi, and D.
illinoense.(Table II). All seeds were scarified by prick-
ing the seed coat with a needle. Plants were grown from
seed in plastic pots containing a mixture of sand and hu-
mus potting soil. Each pot was labelled and numbered so
that the seed source of each plant was known. Th.e pots 17
(Y
O H C\J (Y-) 0 0-I H H 0> H E 0-H0 H -H HH) H H UP 0 0> U) U)O CdH 0 0 0Qo C3 H o U) 0 H C 0 CXI
O 0O 0 *H0 0 0u0 U0u0bf) 000 0 4-)-H 0 . -0 r . - Cd 0 0 . Cd. Cd -P A,0 -HrH U)m U ) 0 Q H O H 0 4-)(H d -H to x HO QH H CO HH P OH E- Cd z cE- dH O 0 0D0 4 Cd ~Cd
ZH
O0 0 H oo ) OH Cd Cd 0) o0U) H CO 0 H rr GH OHOQ) U Cd 0> 0O(f1) 4-)-HQ q\10 *HCO ) C") : O 0 0 - C0 0 0> - c 0>J o o, D\Cd (010 0CM *HH * aC\J0 -Cf-C E H H -_ 3 UH\0 * Z09, H o 00 0
2 fm
SH H4> H0 r :Z Cd E- E- E 9-1a 0 - E1H HC- OCO H M
0 U) Ul) 0 0c -HU) H CO Cd ri H 18
were held in a greenhouse until seedlings were well esta-
blished (three weeks). The pots were then transferred to
a lattice-covered field garden near the author's residence
in Denton, Texas. The pots were moved because the plants
were to be used in a reproductive biology study, and expe-
rimental studies with Desmodium uncinatum and D. intortum
(1,6) have shown that when plants are grown under uniform
field garden conditions, a high percentage of successful
interspecific hybridizations can be effected (as compared
to when the same plants are grown under greenhouse condi-
tions). A complete nutrient solution was applied to the plants every three weeks and water was supplied as needed.
A total of ten pots (one plant/pot) for each species were maintained.
Observations on growth form, flowering behavior, and reproductive biology were made on the basis of these field garden plants. The plants were also a source of flowers used for determining chromosome numbers, pollen morphology and the study of intraspecific and interspecific hybridi- zations. Several specimens were sacrificed to serve as vouchers and have been deposited at the Benjamin B. Harris
Herbarium (NTSU) and at the S. M. Tracy Herbarium (TAES). 19
Reproductive Biology
Preliminary Studies of Breeding Behavior As a preliminary study of the breeding behavior in Desmodium canescens, D. tweedyi, and D. illinoense, the following treatments (after Radford, 7) were performed on plants growing under uniform environment conditions. For each species:
(a) Twenty flowers were tagged but otherwise left alone as a test for cross-pollination and as a control to determine normal seed set.
(b) Twenty flowers were caged (using nylon netting) and left alone to test for self-pollination; normal seed set being an indication of self-pollination and self-fer- tility (or apomixis).
(c) Twenty flowers were emasculated to test for apomixis, seed set under these conditions would indicate apomixis.
As a test of relative pollen fertility samples of pol- len were stained for 24 hours by using a 1 per cent aqueous solution of anilin blue added to a lactophenol solution made up of 20 ml melted phenol, 40 ml glycerine, and 40 ml of water (7). Percentage of pollen fertility of each spe- cies was determined by counting at least 200 pollen grains from each of two flowers per plant. Pollen grains were classified as either fertile (full) and stained or sterile (shriveled) and unstained. 20
Hybridization Studies
Five plants of Desmodium canescens, numbered Cl, C2, C3, C4, C5; five plants of Desmodium tweedyi, numbered Ti, T2, T3, T4, T5; and five plants of Desmodium illinoense, numbered Il, 12, 13, I4, 15 were used for interspecific and intraspecific hybridization studies and also for obser- vations on flowering behavior. Crosses were made according to the method of Crowder (2). Flower buds that would open the next morning were hand emasculated by opening the ends of the keel petals with forceps and carefully removing the anthers from the closed buds. The racemes bearing the emas- culated buds were covered with fiber pollinating bags. Wet cotton was inserted at the base of the raceme inside the bags to increase the relative humidity around the emascula- ted flowers. Pollinations were completed between 8 A.M. and 10 A.M. on the following day. Pollinations were made by collecting pollen from the male parent with a sandpaper- tipped toothpick for transfer to the female parent. The following sets of interspecific crosses were made:
(A) D. canescens x D. tweedyi Cf (B) D. tweedyi 9 x D. canescens 5'
(C) D. illinoense g x D. canescens C5' (D) D. canescensj x D. illinoense &:r (E) D. tweedyi 9 x D. illinoense J (F) D. illinoense x D. tweedyi &' Intraspecific crosses for each species were also made. 21
Flowering Behavior
Thirty plants representing the three species were ob- served for flowering behavior during June and August 1975.
Observations on number of flowers per raceme, flowering period in days per raceme, and number of flowers opened per day were recorded. Also recorded were the events of flower maturation up to the time of anthesis. Brief obser- vations were also recorded concerning the pollinators of the Desmodium flowers.
Cytological Evidence
Chromosome Counts
Flowers for use in chromosome determinations were ta- ken from field garden plants. Since the plants were grown from seeds taken from annotated herbarium accessions (see
Table II, p.17), the chromosome counts for each species are based on plants whose parents were known to be members of different (usually widely separated) populations. Chro- mosome counts for each species are thus based on replica- ted, independent observations of several figures from the
same slide, from different slides, from different indivi- duals, and from different populations.
Flower buds were fixed in chloroform, 95% ethanol, and glacial acetic acid (6:3:1 v/v/v) for twenty-four hours.
The materials were washed and stored in 70 per cent ethanol,
and dissected in 70 percent ethanol or 45 percent acetic 22
acid (11). Anthers were mounted in Hoyer's medium (1).
Squashes were viewed with bright field optics. Chromosome
numbers were based on diakinesis or metaphase I figures.
Drawings of the chromosomes were made with the aid of a
camera lucida at a magnification of 1000 x. Voucher slides
for chromosome counts have been deposited at NTSU and TAES.
Palynological Evidence
Pollen Morphology and Size
Both herbarium specimens and field garden plants were
used as sources for pollen material. Herbarium specimens
which served as sources for seeds were also utilized as
sources of pollen material (see Table II, pageC'). The
acetolyis method of Erdtman was used for the preparation
of pollen grains for study (3). Pollen grains were then
transferred from 100 per cent alcohol, to benzene, and fi-
nally to silicone oil. Permanent pollen slides were made
by placing a drop of pollen-silicone oil mixture on a slide,
applying a cover slip, and then sealing the edges of the
cover slip with nail polish. A standard light microscope
equipped with a reticle was used to observe the morphology
of the pollen grains and to make size measurements. A to-
tal of ten slides were prepared for each of the species
being studied. Size values are based on measurements of
fifty grains on each slide. Voucher slides have been de- posited at NTSU and TAES. CHAPTER BIBLIOGRAPHY
1. Chow, Kuan Hon, "Hybridization of' D. cahum (Gmel. Schin and Thell. and D. uncinatum (Jacq.7]5 "'Crop Sciences, 12(1972)787 . .8Cr,
2. Crowder, L.F., Hybridization of Desmodium cahum' (Gmel.) Schin , and Desmodium uncinatum~ a cq~~ Plant BreedinE~Taper NO. 63l, Cornell~Uniersity, Ithaca, New York, 1972.
3. Erdtman, G., An Introduction to Pollen Analy'sis, Waltham, Massachusetts, Chronic a Botanin C company, 1943.
4. Fernald, M.L., Gray's Manual of Botany, 8th ed., New York, American Book Company, 1950,
5. Gleason, H.A.,' The New Britton' And Brown' T'lstrated Flora of the Northeastern UnitedStates and' Ajbent Canad, Vol.~~:24-4~~acasterFenns3y va Thn~aanc UsKe Press, 1952.
6. Park, Soon Jai, "Genetic Studies in Spanish Clover, D. sandwicense. I. Inheritance of Flower Color, ,Stem Color and Leaflet Marking," Crop Scienc'es, 8(1968), 467-470~
7. Radford, A.E. and W.C. Dickinson, (Preliminary Edition) Vascular Plant Systematics, New York, Harper and Row, Publishers, Inc., 1972
8. Schubert., Bernice G. , in' Manual of the Vascular \Plants'of 'Texas, by Donovan S., Correll aTnd- -r sTiT~7 T~fstjon Renner, Texas, Texas Research Foundation, 1970.
9. Schubert, B.G., personal communication, Arnold Arboretum, Harvard University, Cambridge, Massachusetts, Oct 4, 1972.
10. Small, J.K., Flora of the Southeastern United' States, L anc as t e r, Pen nsyIvan i a, ~77~FrTn Fifgo'-7T9 3.
11. Snow, R., "Alcoholic-Hydrochloric Acid-Carmine as a Stain for Chromosomes in Squash Preparations, "- Stain Tech- nology, 38(1963), 9-13.
23 24
12. Turner, B.L., The Legumes of Texas, Austin, The Unive sity of Texas Press, 1959utU CHAPTER III
RESULTS AND DISCUSSION
Morphological Evidence
Herbarium and Field Garden Studies
Range distributions. -- The distributions of the three species which comprise the Desmodium canescens complex are shown in Figure 4. Desmodium canescens is the most widely distributed of the species, ranging from the east-central
United States, east to New England, south to Florida, and west to portions of Texas. Desmodium illinoense also has a broad range, although it is somewhat more geographically restricted than D. canescens. Generally, D. illinoense extends across the north-central United States, but unlike D. canescens, is does not extend into the Southern or New England areas.
Desmodium tweedyi, in contrast, has a very restricted geographical range. This species in known only from north- central Texas (including the Edward's Plateau and the eastern part of the Hill country) and from a small section of southern
Oklahoma.
The southern and midwestern limits of D. canescens and
D. illinoense are clearly established for the first time by this study. The significant northern extension of D. tweedyi
2,5 26
0
(J00 0 0-H
0
to~
Cd
0
0 Q4CI'H
LO
o H
(H
cop
'Hz 0 'H 0 c D D comq5- 0
u 4--)
0 0 0 -H -
0 q
00 27
into southern O:klahoma is also documented for the first time.
The clarification of these distriubtions is noteworthy because previously Desmodium tweedyi has been considered "geo- graphically separated" from D. illinoense (15, p. 214). It is now clear that in the counties of southern OQklahoma this is not true. Previous floristic works have shown that D. illinoense and D. canescens are sympatric in portions of their ranges (6), and that D. canescens and D. tweedyi are sympatric in parts of north central Texas (12, 14) but the sympatry of the entire D. canescens complex in the area of southern Gklahoma has never been reported. D. tweedyi, for
example, is not reported in the 1972 edition of Waterfall's Keys to the Flora of Oklahoma (16). The area of sympatry is
very significant in light of the poorly defined relationships within the D. canescens complex. For example, the use of geographical data as a basis for separating D. illinoense and
D. tweedyi (15) can no longer be accepted. This area of sympatry also raises some interesting bio- logical questions. Whether or not, for example, hybrid swarms might exist in the area and what their parental stock might
be. Information of this type would be invaluable in the
estimation of relationships within the D. canescens complex.
Although the occurrence of hybridization between Desmodium species in nature has never been reported, Park (9) has shown that interspecific hybridizations can be made between 28
Desmodium species which are known to be closely related.
Analysis of morphological data. -- Table III presents
a summation of nine qualitative characters used in various
combinations by different authors (1, 5, 6, 12, 14, 15) in
the delimitation of D. illinoense, D. tweedyi, and D. canes- cens. (A qualitative character, as opposed to a quantitative measurement character, is one for which a choice between two
or more alternatives is made.) It is clear from these cha- racters that the taxonomic confusion in the D. canescens
complex centers on D. tweedyi because it seems to be morpho-
logically intermediate between the illinoense and canescens extremes,
One qualitative character, however, variously described as leaflet markings (12) or white patches along the mid- rib (15) has been used to distinguish D. ttWeedyi. This char- acter is illustrated in Figure 5. These leaflet markings are
Fig. 5--Leaflet markings on D. tweedyi terminal leaflet 29
TABLE III
A COMPARISON OF D. CANESCENS, D. TWEEDYI AND D. ILLINOENSE ON THE BASIS OF NINE QUALITATIVE CHARACTERS
Character D canescens D. tweedyi D. illinoense Racemose ~~~ Inflorescence to Branching Paniculate Slightly Racemose Paniculate. .
Leaflet Reticulation Weakly Slightly Strongly Evident Evident Evident
Stipule Absent Scattered Dense Pubescence
Inflore s cence Weakly Pubescence Pilose Spreading Pube rulent.
Leaflet Thickness Thin Interme- Thick diate
Stipules Reflexed Yes Variable No
Flower Color Pink to White White White
Shape of Articles Semi-rhombic Subrhombic Suborbicular (acute) (acute.) (rounded) Leaflet Markings Absent Present Absent 30 readily observable on all specimens which exhibit the other intermediate and less distinct tweedyi characteristics.
Turner(15) and Schubert (12) emphasize this character in their artificial keys to the Desmodium species of Texas. In both cases it is the diagnostic character used to distinguish
D. tweedyi from D. canescens. It is significant to note, however, that within the genus, the occurrence of leaflet markings is not restricted to D. tweedyi. The same pattern of whitened patches is also used in the descriptions of the following Hawaiian Island species: D. uncinatum (Jacq.) DC. and D. sandwicense E. May. Park (9), Chow (2), and Crowder
(13), have shown that inheritance of the leaflet markings character is controlled by a single pair of genes, with dom- inance for the markings and recessive for non-markings. These data are important because of the indication that even though the leaflet marking character is dominant, some plants of these species can be expected to exhibit plain, unmarked leaves. Among Texas species, D. psilophyllum Schlecht. is also described as occasionally exhibiting leaflet markings. A preliminary examination of D. psilophyllum specimens at TAES revealed that specimens collected along the west Texas limits of the species uniformly exhibit leaflet markings and those collected in south-central Texas do not. In light of these variations in the leaflet markings character, it does not seem appropriate to utilize it as a definitive character in the separation of D. canescens and D. tweedyi. 31
The qualitative comparisons seen in Table III ( page 29) are representative of the findings of this study, in which 198 herbarium specimens and thirty field garden plants w re examined. With the exception of leaflet marking (pr( piously discussed) and article shape, none of these character s could be used to consistently delimit D. tweedyi from eith( r D. canescens or D. illinoense. The shape of the lom: nt
articles did provide an accurate separation of D. twE edyi and D. canescens (with angular articles) from D.illinoenm e (iith distinctly rounded articles). This morphological chtracter is very constant and though difficult to define in qt antita- tive terms, it is still very recognizable. The artic les of D. tweedyi and D. illinoense are strongly dissimilar. In fact, the analysis of qualitative characters for the Desmodium canescens complex reveals an interesting pattern, i.e ., that (a) except for two, all of the character states of D. tweedyi are intermediate between D. canescens and D. illinoer se and do not show clear relationships in either direction; and (b) that between the two distinct characters, one has bee n shown to be inconclusive and the other reveals a close rela tionship not between D. tweedyi and D. illinoense, but rather between D. tweedyi and D. canescens.
Several quantitative characters which are used i n the description or delimitation of species in the D. cane scens complex were analyzed for morphological variability i a this study. Figures 6 - 10 are graphic representations of the 32 mean, range, and standard deviation values of data obtained by taking measurements of selected characters. (See Appendix
11,111, for supporting data.) Sources for these values were
(a) herbarium specimens, and (b) field garden plants.
Figure 6 (page 33) provides data on the interrelationhips of D. tweedyi, D. canescens, and D. illinoense as determined by measurements of stipules. Characters of the stipule are used by Schubert as the basis for grouping D. rotundifolium,
D. ochroleucum, and the D. canescens complex in the Series
Stipulata (5). Each species exhibits stipules which are ovate- attenuate and chordateto semichordate at the base. This character is used by most authors of regional floras as a means of identifying any or all of the D. canescens complex species (6, 12, 15). Measurements of stipule length based on both herbarium and garden materials show that there are no significant differences between the three taxa with ref- erence to this character. Stipule width data does show some separation of the species, with D. tweedyi being char- acterized by stipules which are more narrow than those in
D. illinoense and D. canescens. However, the most significant measurement, the length/width ratio (which gives a more accurate estimation of stipule form), shows that there are no meaningful differences between these species in regards to stipule shape. It should also be noted that again D. tweedyi data show it to occupy an intermediate position within the
D. canescens complex. 33
N F-4
z0 I r12 C4 i a) -IFI -J t 4- O ~4 - ct$ I C S N 0:) 4 0 - [AQ aml 3SN3ONI1,j p H0.)
(1H H0Ct S4-40rc 0)
E
mom 0o a 1- CQ O o 0 CL -P Ht
. E
A cc *403-j C10 SN3:ZS3NY:) 3-N3ONII -2
-H-Ho - 0 'V 4 V) c 'O E
N t drO ct 0oroco Zq1)b. - z F I F I F 0 ] + t AM Fzo
4T 0--: IAQka3 3AI 3SNIONI-1-1 I 34
Figure 7 (page 35) illustrates the morphological vari- ation in two other characters which have been used to delimit
D. canescens taxa. Turner (15) has used stipel length to distinguish between D. canescens and D. tweedyi in Texas.
According to Turner, D. tweedyi stipels are generally shorter than those of D. canescens. The data of the present study, based on specimens collected over a wide geographical range, indicate that stipel length is not a character which can be used to delimit taxa within the D. canescens complex. Like- wise, as shown in Figure 7, petiole length is not a diagnostic character either.
Quantitative characters relating to leaflet size and shape have also been used in regional floras to separate the
D. canescens complex from other Desmodium species (5, 6, 12,
14). In this case, however, leaflet size measurements have been used only to recognize the complex, and not the component species. Data from this study, however, indicatethat a size relationship does exist within the complex. (See Figure 8, page 36). The data show that although terminal leaflet length is not a good character for use in separating Desmodium canescens, D. tweedyi, and D. illinoense, measurements of terminal leaflet width and length/width ratios can be used.
Both herbarium and garden material show significant overlaps in the values for leaflet length. Also, the mean length values for all three species are very similar. Values for leaf- let width demonstrate a like phenomenon, although there is some 35
o <
Cd 0mcti C V~)
4.0 *1r I- a 0 71[ I t 0 0 tt 0. co 'Ck 0 ) -H 0 4- -$ < m < 0g 0o .H H t IAaa3AMJ 3SN3ONII 0- H 0
rd Q o > 0 Cd ro P, o Cr. r Cd r5 rd cr
Sro Cd C (1 bD Cd 0)
: - - E E * 'H I[ I I. C.0 -- ][ II I ch 0 Co -) Z - tor: 0.o
-H -
P q '00 < 0o0 c t4 0 0) .t SN3)S 3NV IA03AI3M 3OT~IMi 0) Cd 36
4i to
E
Cd
0) S N3:S 3NY:) IAG33m1 3SN30N111 I 4) r.
- z WA >. - 4m -i a)0 o CQ
.o r to H C II- + 0 -H 0
roCd oQ) C rd o SN3)S 3NY:) IAQ3MJ. 3SN30NI~I to.Hro (0 0
4-)
r. IM I Cd C Cto I, co (1) ) E Sto II -II z Cd -4 e -so oa +a-II- - IL
WA :3 rd rz 4 "H E - .4
- g 3SN0 -4 0 " SN3S)S3NYD 37 evidence that the values for D. canescens and D. tweedyi are more similar than are those of D. tweedyi and D. illinoense.
The data for terminal leaflet length/width ratios makes even clearer the similarity in leaf shape of D. canescens and D. tweedyi. The data show that the leaves of D. illinoense are significantly different from those of the other two species in the length/width proportion, i.e., D. illinoense leaves tend to be more lanceolate or lanceolate-ovate while those of
D. canescens and D. tweedyi tend to be broadly ovate.
Measurements of primary tracts (another frequently used keying character) illustrate even more clearly the species relationship (Figure 9, page 33). Both length and width data show that both D. tweedyi and D. canescens are significantly different from D. illinoense and that they are very similar to each other. The primary bract length/width ratios show dramatically what first became apparent from the terminal leaf- let data, i.e., that within the D. canescens complex, the clearest evidence of affinity is the previously unreported re- lationship of D. canescens and D. tweedyi.
Three additional quantitative characters are illustrated in Figure 10, (page 39). These are among the few floral-related characters used in the diagnosis of Desmodium species. The graphic comparisons of the measurements of morphological vari- ability again show the close affinity of D. canescens and D. tweedyi and their separateness from D. illinoense. 38
i I
'Q E + t 4) -a ci) Cd rcd lulz IL I (1 to
H :5
a) 0
SN3DS3NYD 1A033MJ 3SN30NI111 ci)O 0
-H O t -HH co Ix H
rOl Zto SCdr $ s 0 < w 4) QC) 3SN30NIIII roriC
C O -"
o0
CdciE o
CdE 4
Q) 04M 0 -IF I 'o" t -J ]I - .- bo 0:s00 t a" ci) -H r:: -- ][ I- tI2l 4 0o 4~ 0to SN3:S 3NV:) IA aa 3 j 3 S N:i 0-7ITfhI 39
* to J
to C
4m 0
4) 4 SN33S~ NY:) I A (3 3 3 I 3SN3ONITII
4-)t0 0
Cd0 r. 0 I I F to r S0 6 t to -nz EOH0 -I- mHCt
t.. o toC0
reN
S3SN 4 03 +S3 +
- ~C0 0 I
0 -H 0::-- + x o ::I I FT1 o I9, cc a) Cd *H 10 cm C A 0 94c SN3:S3Nv:) IA3gJ 3SN3ONII 2 40
In order to more clearly illustrate the relationships within the D. canescens complex, two scatter diagrams were prepared. The data for these diagrams is based on values ta- ken from twenty randomly selected character analysis forms for each species. (See Appendices IV and V).
Figure 11 (page 41) shows the diagnostic value of two quantitative characters (terminal leaflet length/width and lo- ment isthmus width) in segregating D. tweedyi and D. canescens from D. illinoense. The separation of the two clusters is apparent.
In Figure 12 (page 42) the correlation of stipe length and primary bract length/width again clearly indicates the affinity of D. tweedyi and D. canescens and their separation from D. illinoense.
Thus, analysis of morphological character variation in the D. canescens complex has revealed an unreported alliance between D. canescens and D. tweedyi. The data also shows that in most diagnostic characters, D. tweedyi is not separate and distinct from D. canescens, i.e., it does not seem to hav an identity at the specific level. Both herbarium and gas- den specimens reveal this relationship. The sampling of characters based on field garden specimens is relatively small, but the garden data seem to support the view that or- phological variation within the D. canescens complex is both environmentally and genetically dependent. The relatively high degree of variability with respect to such vegetative 41
2.5- F*
2- 0.
0 %o o 00
0@0 0. U.. 0 @00 0o w 0
-z tm- X I-
CANESCENS 0 TWEEDYI 0 ILLINOENSE $
1 1 I 1 2 3 5 T LOMENT ISTHMUS WIDTH mm
Figure 11. Analysis of variation within the Desmodium canescens complex. The variation of two morphological characters is compared by means of a scatter diagram. 42
6A
3-
o00C
I- .8. 0 @000.0 So0 o 000 0 lz 0 0* 40o
uw 1.5+ lz 0* C- .
I-
0.5+ CANESCENS 0 TWEEDY * ILLINOENSE 4
I I I I i 2 3 4 5 6 STIPE LENGTH mm
Figure 12. Analysis of variation within the Desmodium canescens complex. The variation of two morphological characters is compared by means of a scatter diagram. 43 characters as terminal leaflet length, stipel length and pe- tiole length among plants grown under uniform environment con- ditions indicates a significant level of genetic variability.
Likewise, the wide range values (based on herbarium specimens) shown for many characters in Figures 6-10 indicate a strong degree of ecotypic or environmental variation. It is possi- ble that these sources of variation have contributed signifi- cantly to what has been a misinterpretation of actual relation- ships within the D. canescens complex.
Genetic Evidence and Reproductive Biology
Preliminary studies. -- Initial investigations in the
Desmodium canescens complex demonstrated that there are no basic differences in the reproductive patterns of the compo- nent taxa. (See Table IV, p.44).
As illustrated by Treatment A, all three species normally have a high percentage of fruit set and are capable of being cross-pollinated. Flowers which were caged so as to exclude external pollen sources (Treatment B) also exhibited a high percentage of fruit set. This would generally indicate that
D. canescens, D. tweedyi, and D. illinoense can also be self- pollinated and that when this occurs, the plants are self-
fertile. It is possible that a high percentage of fruit set
under Treatment B conditions could indicate apomixis. The
results of Treatment C, however, indicate that apomixis does
not occur in this complex. Flowers which were emasculated and caged failed to produce any fruit. 414
TABLE IV
SUMMARY OF PRELIMINARY STUDIES OF REPRODUCTIVE PATTERNS IN THE DESMODIUM CANESCENS COMPLEX
-r 1 Treatment A Treatment B Treatment C
-4
ro -p 4-) -) 0 (D -) -) rO
0 4-p 4-)a) 'H ~ -) bH Species 9- -H 4) H bID H cr3 CO-4 0 0or .0 0 OH3 Fm-A H M.C) o> H 0 4- o o 4-p'H -p o ~rd CTHb0 H 0P- 0 0 -p 0 sI -o 0 L 9 '0 0 0 t;rl a
De smodium canescens 20 19 95 20 18 90 20 0 0 97.5
Desmodium tweedyi 20 20 100 20 17 85 20 0 0 95
Desmodium illinoense 20 19 95 20 17 85 20 0 0 98
Pollen fertility among these species was also shown to be high. Determinations made on multiple samples taken from each field-garden plant demonstrated little variability,
Hybridization studies. -- Results of attempts of intra- specific and interspecific hybridizations are shown in Table
V. Intraspecific crosses again indicate that the species in the D. canescens complex are easily cross and self-fertile, 45
cY ON H O\H- 0 C) 0 C- -r k0 - L(N 0d S94 ro0
H co C C 00 co 00H . 0 I H P-4 H0 0 H H 00192(y CO CX CO H\H 0 0 :7 Q H CO C CO C CO H * 0 :4 H 01 E U) Sr 0 E LA rzLC\ Ln Ln -\ Ln Lf) LnL-\ Ln LC) Ln L E>1 H 0)A J NJ (J OJ C'J Cu Cu ZZCuJ CZJ 0 Fm 0 1:2 Fm 0 o o E0 Ul) I U)I i:4 H 0 01 .H 10 o 01 -H I UI 0 H 010 E-) 0 C C -H -H r o H z N e H 0101 Ho 0 4 0w H H 0IU) -H C 1111 Fm 'H 'H 'H H p-PH 04 H 0 0 r H 4H'H U) I- Q)H -) 0 H H H Al N N N Fm 0 QI C H U) 04 C/U) U) U) Hd r- 10 So I 0 N 0 0 0 0 U)os- 0kn O' 'H S OH 0iI- 00 H 0eI (i ro roq-, 0 C S4-) O H 'H p-H H OH U) IOl *H 0 0 0 H GH 0 4H 4 ~4-) H OH U) OH H 'H J-'H 0 r O1r1m U) C1O1 H yr TI| I 0 The mechanical techniques involved in these hybridizations are very simple and the successful percentages of attempted with the artificial intraspecific crosses compare favorably in the percentages of natural pollinations as established preliminary studies. Results from attempted interspecific hybridizations sug- gest that D. canescens - D. tweedyi may be more closely re- lated genetically than are D. tweedyi and D. illinoense. The number of successful canescens-tweedyi crosses (three) is relatively high compared to the number of successful cros- ease ses (zero) between the other two taxa. Of course, while of hybridization may indicate a fairly close relationship between two species, the reverse is not necessarily true. Closely related species may still have evolved reproductive isolating mechanisms. It was hoped that numerical character analysis of herbarium specimens of all D. canescens complex species collected in southern Oklahoma (the area of sympatry for the complex) might have yielded more conclusive hybridi- zation evidence. Unfortunately, the number of collections in that area is too small to afford any meaningful data. F progeny of the successful D. tweedyi-D. canescens crosses made in this study have not yet been grown. Flowering behavior. -- Field garden plants of D. canes- cens, D. tweedyi, and D. illinoense were observed for flower ing behavior in July and August, 1974. All species are erect, 47 vigorous growers and, as shown in Table VI, all produced first functional flowers between 48 and 55 days following planting. Herbarium data also indicate that all three spe- cies generally flower during the same period (July-October, as early as May in the South). TABLE VI NUMBER OF DAYS FROM DATE OF PLANTING TO FIRST VISIBLE FLOWERS FOR THREE DESMODIUM SPECIES Days to Flowering *Number of plants observed to. flower/day from Planting Date 30/5/74 D. canescens D. tweedyi D illinoense 48 1-.-. 1 50 3 2 52 4 3 2 54 2 5 2 55 - --- 5 *Based on 10 field-garden plants for each species Desmodium species generally produce a prolific number of flowers and these species are no exception. Table VII illus- trates the flowering behavior of the D. canescens complex. D. canescens and D. tweedyi both exhibit panicle type inflorescences, although the inflorescence of D. tweedyi tends to be significantly less diffuse than that of D. canescens. 48 TABLE VII FLOWERING BEHAVIOR OF THREE DESMODIUM SPECIES Number of Number of Flowering Flowers Flowers Period Opened Opened Per Raceme, Per Raceme Per Raceme In Days Per Day Spe cie s Average~Range ,Average...Range~Aver.pnge D. canescens 40.2 38-43 12 10-13 3.3 3.2-3.8 D. tweedyi 20.3 18-24 10 9 -12 2.0 1.8-2.3 D. illinoense 147 138-154 D. illinoense exhibits a strongly racemose inflorescence with essentially no axillary branching. As a consequence of its branching pattern, no data besides number of flowers produced per raceme were determined for D. illinoense. The average flowering period/raceme in both D. canescens and D. tweedyi are very similar. D. canescens produces a more densely flo- wered raceme than does D. tweedyi and consequently its raceme generally exhibit more pollen receptive flowers/day than do those of D. tweedyi. Total number of flowers/inflorescence was not determined for D. canescens and D. tweedyi. General observations on flower maturation. The flowers of all three species are ephemeral, lasting only three to four hours depending on the environmental conditions. On warm days the flowers open shortly after sunrise and are wilted 49 by about 11 A.M. On cooler, moist days, they may last all day. Once a flower is tripped, it quickly wilts and the standard petal is folded over the stigma within an hour. As has been noted, only a few flowers per raceme open per day. Several days before the flowers open, the part of the stem containing that set of flowers elongates rapidly so that the flowers are separated from each other on the axis. Flowers are borne in pairs at the nodes. The day before the flowers open, the closed petals expand and project beyond the sepals. Truest flower color is present at this time (D. canescens, pinkish white; D. tweedyi, white; D. illino- ense, white). In the late afternoon of the day before the flowers open, the anthers are white (under magnification). By around 8 P.M. the anthers start to trun yellow--a sign of pollen maturation. The anthers dehisce about midnight and the pollen will fall from the anthers if distrubed. The stigma is extended beyond the anthers at this time and little if any pollen touches the stigma. By dawn, the petals have completely expanded but the flower is still closed. Shortly after daylight, the standard petal becomes erect and the flower is receptive to pollinating insect. If the pollen is tripped at this time, pollen is discharged in a cloud around the stigma. Pollinators of field garden plants were normally small bees, and occasionally, wasps (specimens collected but posi- tive identifications not determined). 50 CytologicalEvidence Chromosome counts. Results of chromosome counts for D. canescens, D. tweedyi, and D. illinoense are shown in Figure 13. The counts for each species were the same for D. canescens D. illinoense D. tweedyi (x 1000) Fig. 13. Camera lucida drawings of meiotic chromosomes. every replicate. Early reports (17) on chromosome numbers within the genus Desmodium indicated that the basic number is x=ll. Subsequent authors have shown, however, that the genus is not monobasic, with reports of 2n=20 being common (15). This study likewise has shown that the genus is di- basic, with base chromosome numbers of x=10 (D. illinoense) and x=ll (D_. canescens and D. tweedyi. Most of the chromo- somes have median to sub-median centromeres and there is considerable variation in their length. The difference in chromosome numbers within this complex supportive is of the conclusions drawn from the morphological and hybridization studies, i.e. that among the three species, the closest affinities lie between D. canescens and D. tweedy. 51 Palynological Evidence Pollen Morphology Details of pollen morphology, as determined on the basis of light microscopy, showed no differences between the pollen of any of the D. canescens complex species. The grains are monads (solitary) and exhibit polarity. They are triangular in equatorial view, typically appearing oblate (flattened). Apertures are present, the grains being colporate (exhibiting compound apertures). The grains lack any surface sculpturing and are therefore termed psilate. Measurements of grains from all species likewise revealed no size differences. The grains are in a medium size class, with a mean diameter (equatorial) of 26 pm and a range of 20 Pm - 36 Pm. Significant palynolo- gical differences were not detected within the D. canescens complex. CHAPTER BIBLIOGRAPHY 1. Coulter, John M. Botany of We'stern Texas:. 'C'ohtributions from the U.S. National HerbariTu-m. _' W II,, Washington D.C., Government Prf4nting Office, 1891, 2. Chow, Kuan Hon, "Hybridization of D. canum (Gmel.) Schin. and Thell. and D. uncinatum (Jac-,)T7 "Crop Sciences 12(1972), 784-7'95.- 3. Crowder, L.V., Hybridization of Desmodium 'cahum (Gmel) Schin, and Desmodium uncinaifm (JCacMq C., Pliant Breeding Paper No. 63i, CorneIlU esity, Ithaca, New York, 1972. 4. Erdtman, G., An Introduction to Pollen Analysis Waltham, Massachusetts,Chron'ica Bo5tEnriicET- ompin~f~I- 5. Fernald, M.L., GrayIs Manual of Botany, 8th edo New, York American Book Company, 19507 ~ 6. Gleason, H.A., The New Britton and Brown of the Northeaste Illustrated Flora UniUted~St es andj jan~~ 7-6l~~~2 :3-~ ,~Tan cas t .- Pe nnsy1'an'~T"'LE~c T6astr~ Press, 1952. 7. Isely, Duane, "The Leguminosae of the North-Central Uni- ted States. II. Hedysarease," Iowa State College Jour- nal of Science, 30(1955), 33-11U-~ 8. , unpublished manuscript. Papilionoid Legumes of the United States, Department of~Botain7Iowa~~Eate Uni iersity,~Ames,~~Iowa, 1975. 9. Park, Soon Jai, "Genetic Studies in Spanish Clover sandwicense. D. I. Inheritance of Flower Color, Stem Color and Leaflet Marking," Crop Sciences, 8(1968), 467-470.- 10. Radford, A.E. and W.C. Dickinson, (Preliminary Edition) Vascular Plant Systematics, New York, Harper and Row, Publisher s~,~IrsInc~77~T27 11. Rotar, P., Crossing and Flovering Behavior' in Spanish Clover , D. sandwc 'se~I .eyTe~Bn*17c6 l~7r rssi Re-p~ort N6. I 7 [ai-i Agricultural Experiment Station, University of Hawaii, November, 1967. 52 53 12. Schubert, Bernice G., in Manual of the Vascular Plants of Texas, by Donovan S. Correll and Marshall Johnston, C. - Renner, Texas, Texas Research Foundation, 1970. 13. Schubert, B. G., "Desmodium: Preliminary Studies. III," Rhodora, 52(1950), 135-155. 14. Small, J.K., Flora of the Southeastern United States Lancaster, Penn3sylvanTij, w n7TiTnnTF~Uo~~ 1903. 15. Turner, B.L., The Legumes of Texas, Austin, The Univer- sity of Texas Press, 1959. 16. Waterfall, U.T. Keys to the Flora of Oklahoma, 5th ed. Dallas, Prestige Pres, ~ 17. Young, J.O., "Cytological Investigations in Desmodium and Lespedeza,"i Botanical Ga01 850. 54 CHAPTER IV SUMMARY AND CONCLUSIONS This study represents the first biosystematic study of any Desmodium species complex cenetered in North America. It had defined the range and distribution of the D. canescens complex (Figure 4, p. ) and has provided the first report of chromosome numbers for the component species (D. canescens: n = 11; D. tweedy: n = 11; D. illinoense: n = 10). Palyno- logical studies have shown that the pollen of all three spe- cies are morphologically inseparable at the light microscope level and that the grains are (a) psilate, tricolporate, and triangular in equatorial view and (b) have a mean equatorial diameter of 26 pm. The Desmodium canescens complex species have also been shown to be equally fertile whether they are cross-pollinated or self-fertilized. Artificial crosses have shown successful interspecific hybridizations between D. canescens and D. tweedy and analysis of morphological data also has shown clear affinities between them. Studies of flowering behavior have shown that all three species are similar with reference to the sequence of flower maturation. This critical examination of the D., canescenscomplex has resulted in the delimitation of a previously unreported 55 alliance within the complex. Previous authors have variously granted D. tweedyi specific status (2,4) or reduced it to synonomy with D. illinoense (1,3). The present study however, has shown that D. tweedyi is actually closely allied with D. canescens and that on this basis the synonomy of D. tweedyi and D. i_llinoense should not be recognized. Further, the da- ta suggest that D. tweedyi is not a distinct species but rather that it is more likely a geographical variant (the southwestern extreme) of D. canescens. The following points support this view: (a) analysis of morphological data taken from herbarium and field garden specimens indicate that for many diagnostic characters, the mean values of' D. canescens and D. tweedyi are not significantly different; the same values for D. illi- noense, however, are distinctly different from both D. cane- scens and D. tweedyi. (b) breeding experiments have shown that artificial in- terspecific hybridization is possible between' D. cahescens and D. tweedy. (c) cytological studies have shown that the basic num- ber of both D. canescens and D. tweedyi (x=ll) differs from that of D. illinoense (x=10). Systematic Treatment -- A new combination is suggested: Desmodium cane'scens var. tweedyi (Britt.) Williams. CHAPTER BIBLIOGRAPHY 1. Jones, G.N. and G.D. Fuller, Vascular Plants of Illinois, Urbana: University of Illinois Press, 1955. 2. Schubert, Bernice G.,, in Manual of the Vas'cular Plants of Texas, by Donovan S. Correll and Marshall C. Johnston, Renner, Texas, Texas Research Foundation, 1970. 3. Steyermark, J.A., Flora of Missouri, Ames, Iowa State University Press, 1963. 4. Turner, B.L., The Legumes of Texas, Austin, The University of Texas Press, 1959. 56 APPENDIX I CHECKLIST FOR CHARACTER ANALYSIS DESMODIUM PROJECT Specimen # Collector Acc# Date Location Habitat ...... Vegetative CHXS ROOT LEAFLET STEM texture shape pubes mottled x-sect apex misc. base ciliate STIPULE upper surf. lower surf. shape quantitativee: pubes mid-rib length terminal: width length misc. width lateral: PETIOLE length width pubes x-section length misc LEAF rachis length misc. 57 APPENDIX I -- Continued Reproductive CHXS INFLORES. FLOWER type Calyx: rachis chx pubes # flws ciliate length (central tooth PRIMARY BRACTS of lower lobe) trichomes present shape length (lateral teeth) striate length (upper-bifid- lobe) puberulent ciliate Corolla: pubes color persistent length Loment: width stipitate length of stipe SECONDARY BRACTS # of articles shape of articles ciliate reticulate pubes pubes shape article length persistent article width length isthmus width width PEDICELS pubes. length width 58 APPENDIX II TABULATION OF MORPHOLOGICAL DATA BASED ON HERBARIUM MATERIAL (Desmodium twe'e'dyf) Standard Character Mean Deviation Range Stipule Length (mm).. . 9.8 .91' ' ,70-14.0 Stipule Width (mm) 3.0 .42 '2.0-4.0 Stipule Length/Width (mm) 2 . 20 .06 1. 9-2. 4 S-tipel ength (mm). 3.2 .62 1.0-4. 0 Petiole Length (cm) 5. 3 .40 4.o-6.o Terminal Length Leaflet (cm) 6.0 .92 4.0-1. 0 Terminal Leaflet Width 4.0 72 3.0-8.0 Terminal Leaflet Length/Width (cm) 1.6 '.4' '1.0-1.8 Primary Bract Length (mm) 4.2. 60 3.1-5.1 Primary Bract Width (mm) 3.2- 20 2.0-3,8 Primary Bract Length/Width (mm) 1.6 06' 1. 4-1. 8 Upper Calyx Lobe Length (mm) 4.2 29' '3.0-4.8 Loment Isthmus Width (mm) 5.1 .52 3.0-7.0 Stipe Length (mm) 4.2 .23' .3.0-4.8 59 APPENDIX II TABULATION OF MORPHOLOGICAL DATA BASED ON HERBARIUM MATERIAL (Desmodium illinoense) Standard Character Mean Deviation Range Stipule Length (mm) 10.0. .90 6.0-13.0 Stipule Width (mm) 5.0 .49 Stipule Length/Width (mm) 2. 0 . 06 1. 8-2.-3 Stipel Length (mm) 3. 8 .60 . 474. 7 Petiole Length (cm) 6.0. . 77 3.0-9.0 Terminal Leaflet Length (cm)' 7.0. . 89 2 .,' ,0 '.- 2.0-9.0 Terminal Leaf let Width (cm) 3.0 . 30 Terminal Leaflet Length/Width (cm) 2.4 .12 2. 0-2. 6 Primary Bract Length (mm) 6.3 *70 Primary Bract Width (mm) 2.0' .17 1.2-2.3 Primary Bract Width/Length (mm) 3.15 .07 2.8-3.2 Upper Calyx Lobe Length (mm) 3.0 .26 2.0-3 5 Loment Isthmus Width (mm) 2.0 .09, l5-2 5 Stipe Length (mm) 2.0 .20. 1.5-3 d1 60 APPENDIX II TABULATION OF MORPHOLOGICAL DATA BASED ON HERBARIUM MATERIAL (De sodium cane'sbcens) Standard Character Me.an Deviaion Range Stipule Length (mm 9.0 111 5.0-13.0 Stipule Width '(mm) 6.0 .61 5.0-8.0 Stipule Length/Width (mm) 2.25 .06 Stipel Length (m) 3.4 .64 1.0-5.0 Petiole Length (cm) 5.0 1.12.1-10.00 Terminal Leaflet Length (cm) 7.0 14 4.0-12.00 Terminal Leaflet 'Width(cm)' ' 5.0 '80 3.010.00 Terminal Leaflet Length/Width (cm) 1.5 15 11-1 Primary Bract Length (mm) 5.1 .65 3.0-6.0 Primary Bract Width (mm) 3.0 14 2.7-3.2 Primary Bract Length/Width (mm) 1.7 .07' .5.l Upper Calyx Lobe Length (mm) 4.5 .30 3.0-5.0 Loment Isthmus Width (mm) 5.0 .40 '. 3.0-6.0 Stipe Length 4.6 .25' 3.0 5.2 61 APPENDIX III TABULATION OF DATA BASED ON PLANTS GROWN IN UNIFORM CONDITIONS (Desmodium 'tw'e'edyi) Standard Character Mean Deviation Range Stipule Length (mm)' 8.0 .23 ' 7.0-10.0 Stipule Width (mm) 2.8 .. 17 2.0-4.0 Stipule Length/Width (mm)' 2.1 .'07 1.9-2.3 Stipel Length (mm) 3.0 .46- 1.0-3.6 Petiole Length (cm). *57' 4.1-5.2 Terminal Leaflet Length (cm) 5.6 .39' 5.0-10.0 Terminal Leaflet Width(cm) 4.2 2 . 27' 3 0-6.0 Terminal Leaflet Length/Width (cm) 1.2 12' 10-1.6 Primary Bract Length (mm) 4.0 .10. 3,8-4.5 Primary Bract Width (mm)' 3.0' .09 2.5-3.2 Primary Bract Length/Width (mm) 1.3 .04 1.2-1. 4 Upper Calyx Lobe Length (mm) 3.9 .09 3.6-4.2 Loment Isthmus Width (mm) 4.9 .09 4.3- 52 Stipe Length (mm)' 4.0 .05 3.8 4.1 62 APPENDIX III TABULATION OF DATA BASED ON PLANTS GROWN IN UNIFORM CONDITIONS '(Desmodium canescens) Standard Character Mean Deviation Range .0. ,1 ...... 1 --. 0,| Stipule Length (mm) 8.0 19 7.0-9.0 Stipule Width (mm) 4.8 .20 4,0-6.0 Stipule Length/Width (mm) 2.1 .06 1.9-2.2 Stipel Length (mm) 2.5 45 1.0-3.0 Petiole Length (cm) 4.8 '38 3.0-6.0 Terminal Leaflet Length (cm) 6.3 .28 5.0-8.0 Terminal Leaflet Width (cm) 4.8 .32 3.,0-7.0 Terminal Leaflet Length/Width (cm) 1. 4 .09 '1.2-1.5 Primary Bract Length (mm) 4.8 .12 4.0 .0 Primary Bract Width (mm) 2.8 .07 2.6-2.9 Primary Bract Length/Width (mm) 1.7 .06 1.4-1.6 Upper Calyx Lobe Length (mm) 4.3 09 3.7-45 Loment Isthmus Width (mm) 4.4 .12 4.0-5.1 Stipe Length (mm) 4.2 06 4.0-4.5 63 APPENDIX III TABULATION OF DATA BASED ON PLANTS GROWN IN UNIFORM CONDITIONS '(Desmodium illi'noense ) Standard Character Mean Deviation Range Stipule Length (mm)' 9. 4 .18 8o.-10.oo0 Stipule Width (mm) . 21 5.0-7.0 Sti pule Length/Width (mm)' 1.9 .01 1.9-2.0 Stipel Length (mm)' 3.8 .48' 0.9-4.0 Petiole Length (cm) 5.8 .41 4.0-7.0 Terminal Leaflet Length (cm)' 6.5 .40 2.0 8.0 Terminal Leaflet Width (cm) 3.2 .10 2.0-4.0 Terminal Leaflet Length/Width (cm) 2.1 .08' 2.0-2.3 Primary Bract Length (mm) 6.5 .13 5.5-7.0 Primary Bract Width (mm) 1.9 .06 1.8-2.0 Primary Bract Length/Width (mm) 2.9 .07 2.6-3.1 Upper Calyx Lobe Length (mm)' 2.7 .07 2.5-2.9 Loment Isthmus Width (mm) 2.1 .07 1.9-2.3 Stipe Length (mm) 1.9 .07 1.6-2.2 64 APPENDIX IV DATA SUPPORTING THE SCATTER DIAGRAM COMPARING LEAFLET LENGTH/WIDTH AND ISTHMUS WIDTH D. canescens D. tweedyi D. illinoense ordinate/abcissa ordinate/abcissa ordinate/abcissa 4.5/1.8 4.4/1.7 1.0/2.5 4,0/1.50 4.0/1.52, 1.0/2.4 4.8/1.50 4.3/1.6 1.8/2.3 5. 1/1.50 4.1/2.0 1.9/2.4 5.5/1.50 4.0/1.50 2.0/2,6 4. 3/1. 52 5.0/2.0 2.0/2.7 4.1/1.53 4.5/1.6 2.1/2.6 4.5/1.8 4.5/1.6 2.4/2,3 5.0/2.0 4.3/1.4 2.8/2,4 4.8/2.1 4.2/1.5 2,1/2.5 - 65 APPENDIX V DATA SUPPORTING THE SCATTER DIAGRAM COMPARING STIPE LENGTH AND BRACT LENGTH/WIDTH D. canescens D. tWeedyli D. illinoense ordinate/abcissa ordinate/abcissa ordinate/abcissa 5.1/1.5 4.8/1.4 1.5/2.9 5.4/1.6 5.0/1.5 1.5/2.6 5.5/1.8 5.1/1.8 1.9/2.9 5.5/2.1 5.2/1.9 2.1/2.8 5.9/2.2 5.6/2.2 2.4/2.7 6.0/1.7 5.9/2.1 2.7/2.6 6.2/1.6 6.0/1.9 2.8/2.5 6.1/1.9 6.2/1.4 2.8/2.5 5.8/2.0 6.1/1.4 3.0/2,5 5.2/2.0 5.9/1.6 3.1/2.4 66 BIBLIOGRAPHY Books Cory, V.L., and H.B. Parks, Catalogue of the Flora of Texas, Washington D.C., Government PrintingOffice, 1937. Coulter, John M., Botany of Western Texas: ' Contributions from the U.S. National Herbarium. Vol. II. Washington D.C., ioveFnment Printing Office, 1891. Erdtman, G., An Introduction to Pollen Analysis, Waltham, Massachusetts, Chronica Botanica Compan7y~1943. Fernald, M.L., Gray's Manual of Botany, 8th ed., New York, American Bo ok Comrpany~, 19I50 . Gleason, H.A., The New Britton and Brown Illustrated Flora of the Northeastern United States and Adjacent Canada, Vol. 2: 424-432, Lancaster, Pennsylvania, Lancaster Press, 1952. Hutchinson, J., The Genera of Flowering Plants, Oxford, The Clarendon Press, Vol. 1, 477-4784,3196T~ Jones, G.N. and G.D. Fuller, Vascular Plants of Illinois, Ur- bana., University of Illinois Press, 1955. Radford, A.E. and W.C. Dickinson, (Preliminary Edition) Vascular Plant Systematics, New York, Harper and Row, Publishers, Inc., 1972. Rydberg, P.A., Flora of the Prairies and Plains of Central North America, New York, New York Botanical7Garden, 1932. Schindler, A.K., In: Fedde, Repert. Spec. Nov. Beihefte. Bd. 49. Die Desmodiinen in der botanischenfLiteTaeture ach Linne. 192. Schubert, Bernice G., in Manual of the Vascular Plants of Texas., by Donovan S. Correll and Marshall C. Johnston, Renner, Texas, Texas Research Foundation, 19t0. Shinners, L.H. , Spring Flora of the Dallas and Ft. Worth Area, Dallas, Prestige Press, 195.~ ~~ 67 68 Small, J.K., Flora of the Southeastern United States, Lancas- ter, Pennsylvania, New Era Printing Co., 1903.' Steyermark, J.A., Flora of Missouri, Ames Iowa State Univer- sity Press, i1 7 Taubert, P., in Die Naturlichen Pflanzenfamii'ien, by A. Engler and K. Prantl, III - 3: 70-396,71894. Turner, B.L.,* The Legumes of Texas, Austin, The University of Texas Press, 1959. , Chemotaxonomy of the Leguminosae, New York, Academic Press, 1971. Van Brussen, Theodore, The Vascular Plants of South 'Dakota, Ames: Iowa State University Press,~1976. Waterfall, U.T., Keys to the Flora of Oklahoma, 8th ed., Dallas, Prestige Press, 1972. Articles Chow, Kuan Hon, "Hybridization of D. canum (Gmel.) Schin. and Thell. and D. uncinatum (JacqY) D 7C.7 Crop Sciences, 12 (1972), 784-785. Desai, NM.V., "Bacterial Leaf Spot Disease of D. rotundiflorum," Current Science, 29(1960), 65-66. Fosberg, F.R., "Critical Notes on Pacific Island Plants," Micronesica, 4(1968), 255-259. Ghosal, S.,, "Beta-phenethylamine, tetrahydroisoquinoline and Indole Alkaloids of Desmodium tiliaefolium," Phytochem- istry, 12(1973), 193-197. Ghosal, S.., "Chemical and Pharmacological Evaluation of D. pulchellum," Planta Medica, 21(1972), 398-409. Ghosal, S., "Desmodium Alkaloids. II. Chemical and Pharma- cological Evaluation of D. gangeticum," Planta Medica, 22(1972), 434-440. Isely, Duane, "Desmodium paniculatum (L.) DC. and D. veridi- florum (L. ) DC .-,," American Midland Naturalist 49 (1953), 920-933. Isely, Duane, "Desmodium: Section Podocarpium Benth.," 7 (1951), 185-224. 69 Isely, Duane, "The Leguminosae of the North-Central United States. II. Hedysareae," Iowa State Colle ge'Journal of Science, 30(1955), 33-117.g Knapp-Van Meeuwen, M.S., "Preliminary Revisions of Some Gene- ra of Malaysian Papilionaceae. V. A. Census of the Genus Desmodium," Reinwardita 6(1962), 239-276. Mason, D.L., "Host Parasite Relations in Spot Anthracnose of Desmodium," Mycologia, 61(1969), 1124-1141. Moore, J.A., "The Vascular Anatomy of the Flower in the Papi- lionaceous Leguminosae," American Journal of Botany, 23(1936), 279-290. Park, Soon Jai, "Genetic Studies in Spanish Clover, D. sand- wicense. I. Inheritance of Flower Color, Stem Color and Leaflet Marking," Crop Sciences, 8(1968), 467-470. Rau, M. Anataswamy, "The Development of the Embryo of Cya- mopsis, Desmodium and Lespedeza with a Discussion on the Position of the Papilionaceae in the System of Embry- ogenic Classification," Phytonorphology, 4(1954), 418- 430. Schubert,, B. G., "Desmodium: Preliminary Studies. III," Rhodora, 52(1950), 135-155. Schubert, B.G., "Desmodium: A New Species of Desmodium from Africa," Kew Bulletin, 25(1971), 61-63. Schubert, B. G., "Desmodium: Preliminary Studies. IV," Jour- nal of the Arnold Arboretum, 44(1963), 284-297. Segelquist, C.A., "Moistening and Heating Improve Germination of Two Legume Species," Journal of Range Management, 24(1971), 393-394. Snow, R., "Alcoholic-Hydrochloric Acid-Carmine as a Stain for Chromosomes in Squash Preparations," Stain Technology, 38(1963), 9-13. Tiwari, R.D., "Physcion 1-glycosyl Rhamnoside from Seeds of D. pulchellum," Phytochemistry, 10(1972), 1922-1929. Vail, A.M., "A Preliminary List of the Species of the Genus Meibomia Heist., Occurring in the United States and British America," Bulletin of the Torrey Botanical Club, 19(1892), 07-118. 70 Walters, H.J.,"Host Range and Some Properties of Desmodium Yellow Mottle Virus," Phytopathology, 62(1972), 125- 128. Young, J.O., "Cytological Investigations in Desmodium and Lespedeza," Botanical Gazette, 101(1940), 839-850. Reports Crowder, L.V., Hybridization of Desnodium canum (Gnel.) Schin, and Desmodium uncinatum (C-.) DC., Plant Breed- ing Paper No. 631, Cornell University, Ithaca, New York, 1972. Rotar, P., Crossinand Flowering Behavior in Spanish Clover, D. sandwicense E. Mey. Technical Progress Report No. 164, Hawaii Agricultural Experiment Station, University of Hawaii, January, 1971. Rotar, P., Morphological Variation and Interspecific Hybri- dization AmonR D. intortum and D. uncinatum, Technical Bulletin No. 82, Hawaii Agricultural Station, University of Hawaii, January, 1971. Rotar, Peter, Some Agronomic Observations in Desmodium Spe- cies: Seed Weights, Technical Progress Report No . 177, Hawaii Agricultural Experiment Station, University of Hawaii, June, 1966. Wilbur, R.L., The Leguminous Plants of North Carolina, Agri- cultural Experiment Station Technical Bulletin No. 151, 1963. Unpublished Materials Isely, D., unpublished manuscript, Papilionoid Le'gumes of the United States, Department of Botany, Iowa State University, Ames, Iowa, 1975. Schubert, B.G., personal communication, Arnold Arboretum, Harvard University, Cambridge, Massachusetts, October 4, 1977.