Variation in Audrotpoom Gerard I Vitiaan and A

VARIATION IN AUDROT POOM GERARD I VITIAAN AND A. SCOPARIU3 TAX. IN TWO OHIO PRAIRIE AREAS

DISSERTATION Presented. In Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By HENRY RAWIE DE SELlil, B.Sc. in AGR., B.Sc. In EDU., M. So. The Ohio State Univei’slty 1953

Adviser TABLE OF CONTENTS

Page

Introduction...... 1 Literature Review Taxonomy ...... 9 Ecology of Big Bluestem...... 11 Ecology of Little Bluestem...... 12 The ecotype and related concepts . 14

Methods and Results Chromosome number...... 19 Germination...... 23 Stem elongation...... 31 Morphological variation...... - ...... 39 Big Bluestem Size classes - Ohio ma t e r i a l ...... 43 Size classes - Ohio vs. prairie-center 46 Scatter diagrams -Ohio material . . . 47 Little Bluestem Size classes - Ohio ma t e r i a l ...... 49 Size classes - Ohio vs. prairie-center 52 Scatter diagrams - Ohio material . . . 53 Discussion and Conclusions...... 55 Summary...... 62 Literature Cited...... 129 Acknowledgements...... 137 Autobiography ...... 139

\ 1.

INTRODUCTION

One of the moot constantly challenging problems of

ecological plant geography Is the interpretation of vegetation In time and space, especially when that vegeta tion is disjunctive. Although mapping the distribution of species, floral elements, and vegetation types Is a pre

cursor to their complete Interpretation, other techniques are helpful. The problem here lies In genetic differences (Infra

specific variation) In Andropogon Gerardl^- and A. scoparlus, hereinafter referred to as Big and Little Bluestem respec tively, which occupy two contrasting prairie habitats in

Ohio. These are the edaphic prairies of southern Ohio and the climatic prairies of the drift plains. It will be shown that there are, in both species, small but definite physiological and morphological characteristics which are expressed to different degrees within the species in the 2 two prairie areas. Whether the differences are greater

^Nomenclature followed is essentially that of Fernald, 1950. 2 The writer uses the term prairie to denote undis turbed grassland communities related floristioally and his torically to the Tall Grass Prairie Formation of the Middle West as mapped by Transeau and many students. Prairie rem nant is used to denote small remnants of these former ex panses left after the destructive actlvltes of man. Prairie area refers to a group of prairies (now remnants) formerly centering In Marion and Adams Counties, Ohio (figures 1, 2, and 3). Figure 1. Map of Ohio showing the two sampling ereas. Prairie remnants from which samples were taken are symbolized by solid circles. The drift areas are Illinoian, 111; Tazewell, T; and Cary, C. (from Gold- thwaite, 1952). eS* ai*

A S,H T A 8 U L A PC A

1__ G E A LI G A

l..: \— T R U W B U L L -i 1 __j P A U L O I IN G,

____ i _ . UIHONING -*1 c ' 1 V A N W E R T OT . CRA VFORO SHLANO. w A v R 1C H LA NO' COLUMBIANA — u

M»A R 1 O N 1— T! TUSCARAWAS;t ILL L_.

G u e r n s e y AC I-- .._ r I— ROE

— Is.

ILL /

L ER MONTl 39 ~T JA C K S O N f a R O w

S c a le o>f M i! e s 2 0 SO 40 lA 'A RE NCt U

ay 4* or less than those (unmeasured) differences between, for example, the Marion-Urbana prairies or the Marion-Toledo oak openings is not known. However, differences do exist between samples of Big and Little Bluestem from the Marion and Adams county prairie areas of Ohio. Big and Little Bluestem were used as test plants be cause of their abundance, indeed their dominance of for mer Ohio prairies, wide geographical distribution, and known genetic variability. One of the earliest statements in the literature con cerning the northern area is that of Atwater, (1819) who wrote: "...Sandusky Plains lying on the high ground be tween the headwaters of the Whetstone fOlentangy} branch of the Scioto River and the waters of streams running into Lake Erie, are still more extensive than those of Pickaway, covered with a coarse, tall grass, intermixed with weeds, with here and there a tree, presenting to the eye a landscape of great extent."

This prairie area occupies part of an undlssected Cary till plain except in the north-west one-third where it is transected by the inconspicuous Wabash moraine. Typically it is gently rolling to flat, the latter especially to the north in the Killdeer Plains, where the Toledo silt loam is widespread and well developed. This soil is known to residents as "Jack-wax" and is almost impervious to water (Morrison, et al, 1918). The area is now thoroughly drained and much of the former wet prairie of Spartlna pectlnata where otherwise undisturbed, has gone over to 5. the Bluesteras. A complete account of these prairies is included in the excellent vegetation survey of the Northern

Virginia Military Lands by Dobbins (1937). Sampling areas, which were mostly along railroad rights of way and little used pastures are indicated in figure 1. Such sampling areas, while showing little dis turbance, may contain genetically unstable populations. Comparisons of other populations with these give unreli able results as noted by Wiegand (1935).

The earliest account, known to the writer, of the Adams county prairie area as it formerly existed is that of John Locke (1838), an early geologist and a qualified botanist. He wrote:

"When it is left in conical mound-like out liers, the marl is often barren of trees, and produces some oeculiar prairie like plants, as the prairie docks, wild sunflowers, scabish and rudbecklas, etc. These places are called 'bald hills' and 'buffalo beats'. Several occur with in a mile of West Union in a northerly direction and would be quite a paradise for the botanist."

The grassland areas of Adams county lie primarily in a discontinous belt east of and roughly parallel to Ohio Brush Creek. They are best developed and most extensive on slopes underlain by Silurian limestone and dolomite, where soil profiles are shallow and poorly developed, and internal drainage is excessive (Jones, 1945; Taylor et al, 1938). These prairies, largely dominated by Little Blue stem and side-oats grama (Bouteloua curtlpendula), have Figure 2. View of a prairie remnant, northwest of Locust Grove, Adams county, Ohio. August, 1953.

Figure 3. A prairie remnant in southwestern Craw ford county, Ohio. In the background is a grove of uercus macrocarpa. In front of it is an extensive §ig Bluestem community. September, 1952. X

F- i jure ^

F ig ***■« 3 8.

been studied by Braun (1928 b) , and Jones (1944, 1945).

The location of remnants from which grass samples were taken are shown in figure 1. All of those indicated occur over dolomite. Two kinds of collections were made. Population sam ples of each species, and sod block transplants were col lected at most prairie remnant stations. The latter were replanted and grew to maturity in the greenhouses of the

Botany Department at the Ohio State University. Morpho logical characters of the latter compared with each other and with herbarium material from the prairie center. The results of well over 20,000 measurements of various types are included here. In addition chromosome counts were made, rates of growth measured, and germination experi ments executed.

It should be apparent that techniques of this kind could contribute knowledge toward a better understanding of species of interest to the eoologiet other than those of the prairie. LITERATURE REVIEW

1. Taxonomy

The Andropogoneae Presl. of the family Gramineae is a large "natural" tribe abundant in tropical and sub

tropical grasslands and savannas. It extends into tem perate regions in North America (Bews, 1929). In a study of recent by Hartley (1950), it was concluded that it Is a groupA

origin with a center of distribution In southeastern Asia.

The tribe contains, according to Pilger (1940), six

subtribes. The genus Andropogon L. is contained in the

subtribe Andropogoninae Stapf. Andropogon is further di vided by Pilger into five sections of which two are native

to North America. These are Schizachyrium, and Lepto- pogon Stapf (Arthrolophls). Amphilophis usually treated as a section of Andropogon is given by Pilger as the genus Botriochloa 0. Kuntze. Species with racemes solitary on each peduncle are referrable to the first section, e.g. Little Bluestem; in the second are species with racemes digitately clustered and Joints of the rachis slender, e.g. Big Bluestem. The other taxon contains no species con sidered in this study.

Variation shown by the two species here considered has not gone unnoticed by the taxonomist. Biotypes with one extreme character, others applying well to the species, have been given names ranging in taxonomic status from 10.

species to form* Big Bluestem, for example, has had at least five spe cific names other than the present one. In addition six varieties have been described, some of which are reductions

from specific to varietal rank. Little Bluestem has been placed In three genera other than Andropogon, has had seven specific epithets and eleven subspecific, varietal, or subvarietal descriptions. Some of these are reductions of species to lower rank. In

addition two forms have been described. While such synonomy appears, at the onset chaotic, it forceably points up the variation In morphology exhibited. There seems to be agreement that the genus Andropogon Is of southern origin. Larson (1947), on the basis of morphological and physiological evidence, suggests a "low latitude" origin of Little Bluestem. Gleason (1908) in discussing some Illinois prairie problems places the origin of the prairie in the southwest; while Harvey (1980) places the post Pleistocene center of dispersal in northeastern Texas, eastern Oklahoma and southern Kansas. Gleason (1923) places the origin of Big and Little Bluestem in the east and considers that they migrated wewt during the Xerother- raic Period. Hubbard, Gleason and P e m a l d agree that the typical (oldest?) form of Little Bluestem Is a Coastal Plain plant. The basis for this conclusion Is not clear. 11. 2. Ecology of Big Bluestem

The distribution of Big Bluestem given by Hitchcock and Chase <1950) Is Quebec and Maine to Saskatchewan and Montana, south to Florida, Wyoming. Utah, Arizona and Mexi co. In the central part of its range It "is the most Im portant dominant of the grasslands which occupy the broad lowland valleys of the larger streams..." (Weaver and

Fitzpatrick, 1934). The Big Bluestem prairie type is described in Kansas by Schaffner, (1926); Oklahoma by Bruner, (1931); and In Texas by Tharp, (1926). Its importance is well recognized in eastern Iowa by Shimek (1925a). Northward it extends into Manitoba (Bird, 1927; Shimek, 1925b). Sampson (1921) describes Big Bluestem in Illinois as "the dominant grass of the upland prairies". Transeau (1935) recognizes that the eastern prairie oenisula is not transitional In nature and Jones (1944) states that in Ohio, at the time of set tlement, the Big Bluestem grassland type was well repre sented.

Weaver and Fitzpatrick In 1932 and 1934 state that the Big and Little Bluestem types constitute 80^ of the cover in the areas sampled. Characteristics of Big Blue- stera resulting In dominance are the dense, rapid developing, sod forming type of growth, great stature resulting in dense shade, and early and pronounced tillering. As many as 200 flower stalks per square meter may be formed. 12.

Roots below the sod or* bunch fill the soil, the long er to a depth of 5-7 feet (see Weaver 1919, 1920). The relationship between soil tyoe and root development of the Bluestems and other grasses has been described by Weaver and Darland (1949). As pointed out by Davidson and Rom berg (1952) the correlation described between root dis tribution and soil type does not take into account possi ble genetic variations within the species. Although oercentage of embryo germination is low and winter killing of seedlings is severe the first year, 14 to 40 inch roots develop within two months after germina tion. In seven to eight weeks a bunch is formed (Weaver and Fitzpatrick, 1934).

Studies on Big Bluestem have been made regards of injury to latitudinal strains by low temperatures (Rog- ler, 1943); dormancy (Shepard, 1938); drought resistance of seedlings (-.Tuehler and Weaver, 1942.); the relationship between litter removal and stem density and height (Cur tis and Fartch, 1950); loss of vigor following Inbreeding (Law and Anderson, 1940); growth following soil deposition upon rhizomes (Mueller, 1941); and growth and ohotoperlodi- city (Rice, 1950).

3- Ecology of Little Bluestem

Hitchcock and Chase indicate the range of Little Bluestem to be from Quebec and Maine to Alberta and Idaho, 13.

and south to Florida and Arizona. In the prairie center It is the "principal dominant of the most Important up land (vegetation) type". Its areal extent is described as many times that of Big Bluestem. To the north it In termingles with Stlpa spartea (Weaver and Fitzpatrick, 1954). To the southeast it competes mainly with Big Blue stem, westward extending Into the mixed prairie (Shimek, 1915; Shantz, 1911) with relicts as far west as Colorado

(Livingston, 1952). In Illinois Little Bluestem original ly occurred on "the more xerophytic uplands and exposed

clay soils" (Sampson, 1921). In Ohio the type occurs on ridges and moraines; within the plateau It is the dominant grass of the small prairie openings frequently associated with Sorghastrum nutans and side-oats grama. There Big Bluestem occurs on low seepage slopes and well watered alluvial deposits. In the prairie center, Little Bluestem ordinarily forms an uninterrupted sod with a 50-90 percent cover on steep slopes or when underlain by loess. Where sod Is formed a dense, compact grass cover results with 50-300 stems per square dm. Hoots fill the soil between and beneath the clumps to a depth of five feet. First year seedlings may be five inches tall with three to four tillers and eight 12—IB Inch well branched roots each by late July (Weaver and Fitzpatrick, 1934).

A. general loss of vigor has been shown In Little 14.

Bluestem following inbreeding (through selfing) although to a degree not as great as that in the proceeding species (Anderson, 1940; Anderson and Aldous, 1938). Seedling drought resistance (Mueller and Weaver, 1942) and photo period by Larson (1947) and Rice (1950) have been studied.

Cornelius (1947) has shown that latitudinal strains, when planted in Kansas, vary in time of flowering, number of grains formed, dry weight and height of the plant, and de

gree of winter injury.

4. The Ecotype and Related Concepts

Since its Inception with Turesson (1922a), the con cept of the ecotype has been applied to many plant species. Originally defined as "the product arising as a result of

the genotypical response of an ecospecies or species to a particular habitat"'3, Turesson applied the concept to such as meadow and shade forms of Lyslmachla nummularla.

Strains of this species when transplanted from the environ ments in which they were growing to a common one retained the phenotype previously thought by botanists to have been environmentally induced. Since that time the ecogenetic

concept of plant variation has been expanded to include, among others, the ecospecies " a group of plants within

3Clausen, Keck and Helsey (1940) amplify further, "It is characterized by its fitness for a particular en vironment within the range occuoled by the ecosoecies as a whole". 15. the cenospecles whose members are able to exchange their genes without detriment to the offspring. Ecospecies are separated from one another by internal barriers that pre vent such free Interchange" (Clausen, Keck and Helsey,

1940). Another la the cenospecles, a group of plants "of common evolutionary origin, so far as morphological, cytological, and experimental facts indicate" (Turesson,

1922b). Among the many ecotypes described by Turesson (1922b and 1925) are dv/arf salt meadow forms of Aster trlpollum. a calcareous rock ecotype of Artemesla campestrls L., and an alpine ecotype of Myosotls sllvatlca. In a paper in

192? he comments upon the wide, high latitude distribution of Poa alplna and suggests that the lowland ecotype is poorly developed due to Pleistocene biotype depletion.

Although the modern foundation of the understanding of infraspecific variation in plants is based upon the works of Turreson, Faegri (193?) suggests that much of the ecotypic distinctness described has resulted from in adequate sampling.

Another Continental worker interpreting variations is J. W. Gregor (1938, 1939; Gregor, Davey, and Lang,

1936). Working with Flantago marltlma he reports that transplants of this species retain the phenotype which characterized them under natural conditions. In 1938, 16. using coastal forme from cliffs, salt marshes and grass lands he reports that "differences between populations are largely dependent on the frequency with which certain quantitative characters, belonging to a continuously grad ed series, are represented". Gregor (1939) has also been instrumental in developing the concept of the cline. This term was first Introduced by Huxley in 1938 to refer to characters showing geogra phic regularities of distribution. Clines have been des cribed in side-oats grama for reaction to photoperiod by Olmstead (1944) and in such characters as frequency of glandular hairs on the calyx of Teucrlum canadense. grading from 100 percent in the Rocky Mountains to none in southern Florida (McClintock and Epling, 1946).

Clines have been described not only on the regional but also on the local level. Transitions between ecotypes are usually clinal in nature, Terrell (1952) has described forest-prairle ecocllnes in Galium boreale.

In reality the clinal and ecotypic approaches are two different ways of attacking the same problems. Generally, however, the latter has more frequently led to causal in terpretations of infraspecific variations.

Hall (1929) was the first to initiate large scale ex perimental garden studies in North America. With this im petus, they were carried on by two groups of investigators, one headed by F. E. Clements and. the other by Jens Clausen. 17.

The former's work has recently been summarized (Clements, Martin, and Long, 1950) but the views expressed "are at variance with practically all reliable scientific evidence" (Ownbey, 1952). The work of Clausen, Keck, and Heisey (1940, 1943, 194Q) on the other hand, has been widely cited. The re sults with Achillea, 1948, are particularly enlightening in view of the eleven climatic races found in a 200 mile transect across California. The ecotypes differ from one another in their physiological reactions to varying tem perature and moisture regimes. Other characteristics, of a morphological nature, which "must be assumed to have survival value", while being fairly well correlated habi tat to habitat are less convincingly shown to have the same causal selective value in that habitat. Lawrence (1945) has stated that, in Deechampsla caespltosa (L.) Beauv,, correlation between morphology and ecotypic phy siology depends upon gene linkage. In this species eco types are poorly defined morphologically, that is, "phy siological differentiation has been Independant of mor phological differentiation".

While use of the "transplant method" has resulted in much new knowledge about "microevolution", one other method, that of local population sampling has also been of value. First introduced by Anderson (1941) as "mass col lection", two other workers have used it to particular 18.

advantage. Fassett (1941) was able to relate Great Lakes

populations of Rubus parvlflorus to those of the Sierras and suggest a continuous population “during a post glacial cool humid period and subsequent bisection of the range by the aridity of the Great Plains" (the Xerothermic Period?). Woodson (1947) from studies of samples of Ascleplas tuber ose suggests a center of origin for each of Its three sub species. Of particular Interest also is the outward "dif fusion" from an Ozark center of a special type of leaf shape associated with characters of supposed selective ad vantage.

Literature surveys pertinent to the methods and re sults of the ecotyoe and related concepts are In the works of Heisey (1940); Gregor (1944, 1946); Turrill (1946); Stebbins (i960); and Eaker (1952). METHODS AND RESULTS 1. Chromosome Number

Among: the techniques which may be used in determining the degree of relationship between species and among strains of the same species is that of chromosome number and behavior. The only example of such work in the North American Andropogoneae is that of Gould (1953) in section Amphilophis. The sections Schizachyrium, containing Lit tle Bluestem and Arthrolophis, containing Big Bluestem, have never received such intensive treatment. The value of long term exhaustive studies has been made apparent by Babcock (194?) in the genus Creols. During August, September and October, 1952, sod blocks of Big and Little Bluestem were collected from several prairie remnants in each of the two prairie areas in Ohio. Soil was washed from the roots and rhizomes and they were planted in the Botanic Gardens of the Ohio State University in a modified Geneesee silt loam. During Novem ber and December the plants were subjected to the fluctu ating temoeratures of a mild winter. January 3 through 8 soil blocks, containing the roots and rhizomes were re moved to the greenhouse (figure 5). There supplementary light from four banks of flourescent 120 cm. white tubes, operating on a 14 hour day was supplied. The light sup- olied an irradiance of 13—120 foot candles depending upon 20.

the distance of the sod blocks from the source. From these sod transplants, growing; stems were re moved three to four weeks before anthesis, one each from each station from which the two species had been original ly collected. Anthers two mm. long were first killed in Carnoy's solution, later transferred to 70 percent alcohol, and then smeared in aceto-carnlne according to Sear's

technique recently reviewed by Smith (1947).

Chromosome counts in mlcrosporocytes of Big Bluestem revealed a 2n number of 60 (figure 4). The chromosome

pairs all appeared to behave as diploids. The same 2n number has been reported by Church (1940) for plants from

Manhattan, Kansas; and from Fayetteville, Arkansas by Nielson in 1939. Church (1940) also reports a tetraploid strain from Anderson County, Kansas.

The chromosomes of Little Bluestem, sampled as above revealed a 2n number of 40. As in the other species, the pairs all appeared to behave as diploids. The same number is reported from Texas and Kansas by Church (1940). In olants from the vicinity of Rosebud and Damascus, Arkansas, Nielson (1939) reports two morphologically separable tetraoloid strains. Church (1929, 1936) found, in variety frequens from the vicinity of Boston a 2n number of bo

with 20 bivalents and 10 univalents, and in variety vlllo- s1ssimus Kearney from Falmouth, Massachusetts a 2n number of 40. Figure 4. Camera lucida drawings showing chromosomes during meiosis of the species used. Upper pair of figures is of Big Bluestem, left from the north, ac cession 27-4; right from the south, accession 31-4. Lower oalr is of Little Bluestem, left from the north accession 47-4; right from the south, accession 17-30 Scale equals lO microns. jui-e 2. Germination

One of the most critical stages in the life history of a seed olant is that of germination of the embryo. The need for such studies, recently emphasised by Pelton (1953), has not been overlooked by students of prairie ecology. Studies of this tyoe on species including the Bluestems have been made by Blake (1935), Greene and Cur tis (1950), Coukos (1944), and Anderson and Aldous (1938). Spikelets were collected in the field in each prairie remnant on three successive weekly two day periods e.g. the Marion prairie area on 18 and 25 October and 1 Novem ber; on the Adame County prairies: 19 and 26 October and 2 November. Equal weights of spikelets from each prairie remnant of a prairie area were thoroughly mixed, selected to ex clude "runts", bagged in lots of 50 each, and stored at about room temperature for 82 to 95 days. More than 20,000

spikelets were thus treated. They were planted in lots of 50 or 100 in 50 grams of course white sand contained in petri plates, or In soil in flats. Planting in plates was at the rate of 50 per

o1ate at about 3 ran. depth in the sand. Sufficient alka line tap water (13 cc.) was added to saturate it to a level

of within 2 to 3 mm. of the spikelets. These were then subjected to five temperature treatments. One set was 24. left at room temperature; the second set was subjected to

-10°C. for 50 days; groups 3, 4, and 5 were subjected to

temperatures of 10°C. for 30, 50, and 70 days respectively. The number of splkelets per treatment Is Indicated in table 1. All splkelets remained in darkness and if cold treated were retained In darkness during their return to room temperature at the end of the indicated period. Only

during the periods of examination which followed were the splkelets exposed to light: this of a few minutes duration at two day intervals.

Beds for planting of splkelets in soil (mentioned above) were prepared in flats 51 x 38 x 7.7 cm., each di

vided into two parts widthwise. The flats were lined with brown wrapping paper and filled to a depth of 3 cm. with medium-grained sand. Soil from around the roots of sod

transplants from each prairie area was sifted to remove organic and mineral particles larger than 6 mm. in diamet er, and placed to a depth of 2 cm. In the flats. Spike-

lets, 50 or 100 to a row were planted and covered with ad ditional soil to a depth of 1 cm., half of each flat con

taining soil from each prairie area. Each half flat con tained 15 rows of splkelets: six of Big Bluestem, three from each prairie area; and nine of Little Bluestem, six from the southern and three from the northern source. Table 1.

Number of splkelets used In each treatment.

Big Bluestem Little Bluestem source source Ungl. 01. Ungl. 01.

f Rra. °C 600 600 1200 600 30 days at -10° C 300 300 600 300 M it "Sand ^ " 10° C 300 300 600 300 50 u " 10° C 300 300 600 300 70 it " 10° C 300 300 600 300

Treatments f Rm. °C 30C 300 600 300

. Outside 300 300 600 300 G-l. 1 30 it " 10° C 300 300 600 300 50 M " 10° c 300 300 600 300 70 II " 10° c 150 150 300 150 Soil <1 isource) / Rm.°C 300 300 600 300 Outside 300 300 600 300 Jjngl. i 30 It " 10° C 300 300 600 300 50 II " 10° C 300 300 600 300 70 II " 10° C 150 150 300 150

i v 01 26.

The flats were disposed of as follows: One remained at room temperature; one received 70 days of mild winter

and spring temperatures (22 January through 2 April); and three, one each, were exposed to 30, 50, and 70 days at

10° c.

Upon removal of cold treated flats to room tempera tures they were examined twice weekly and emergent seedlings marked with tootholcks (see figure 6). The resultant data are plotted in figures 7 through 11. On the abciesa are the number of days following planting or removal to room temperature. On the ordinate is the real germination percentage (Lawrence, et al, 1947), that Is, number of embryos germinating those able to germinate. The latter is calculated from samples of a source identi cal to those in the tests. Splkelets were dissected and the percentage of undeveloped ovularies of the sample was calculated. It was presumed that the embryos of fully de veloped grains were capable of germination under favorable conditions.

The data in figure 7 indicate the very marked effect of treatment upon the rate of germination in Little Blue stem in coarse sand. Regardless of grain source, those receiving no cold treatment, as well as those receiving 30 days freezing, germinated at a slow rate and the final germination percentage is low. This is in contrast to those frrains chilled for 30 to 70 days. A similar series Figure 5. The grasses growing under greenhouse conditions. Photograph taken about seven weeks after removal of dormant rhizomes and roots from the botanic garden.

Figure 6. Flat of seedlings. Each toothpick marks a young plant. This flat received no cold treatment. Z8.

fljjulrt S ’ Figure ?. Germination of embryos of Little Bluestem in sand. Gl. = grain source the northern one, Ungl. the southern. Real germination 90 80 40 50 70 60 30 20 10 ubr f as t om e perature tem room at days of Number 155 0 5 30 25 20 Fijvte. Gl. 70 Ungl. Ungl Ungl nl 50 Ungl 7 5 5 50 45 0 4 35 30 . scoparius A.

70 Gl-IO nl r temp. Ungl. Ungl-10 rm — Gl> temp. rm. 30. 55

31. of "germlnatIon rate curves" could be drawn for Big Blue stem, as well as for grains of both species germinating In flats. The results of germination of grains of Big and Little Bluestem In sand is summarized in figure 8. In the former, grains from the southern source have a somewhat higher germination percentage while In Little Bluestem the oppo site Is true.

Averages of real germination percentages In the two soil types, discussed above, are shown In figure 9. The same general relation between germination and treatment

Is shown here as In the proceeding figure except that grains of Big Bluestem from both areas germinate about e- qually well. The effect of soil type Is indicated In figures 10 and 11. In Little Bluestem the higher germi nation percentage Is again evident with longer cold treat ment, as well as the genera1ly higher percentage in soli from the prairie area from which the grains came. The number of germine.ting grains of Big Bluestem from one or the other prairie area, on the other hand, appears to bear no relation to the soil from that same area.

3. Stem Elongation

Stem elongation is a function of the interaction of many Internal and external complexes. Thus measurement of Figure 3. Germination of both species in sand, the abscissa D.F. = days at -10 C.

Figure 9. Germination of both species in soil. Points are averages of germination percentage in two soil types. In the abscissa, Out = 70 days of mild winter and spring temperature. Reol Germ % Reo) (j6frn ♦» 100 0 4 too 0 6 60 ao 20 ZO 40 mC Out Rm*C 30 50 et ent Treatm 0 7 Ungl x TREATMENT 7 O i 8 Fi C DF D *C m R Ungl mC u 30 50 70 7 0 5 0 3 Out Rm-C o x 3 0 3 s t A. ■A. copariu* scoparius A _ 7 O3 0 5 0 -xU ngl O Gl i _ o G I -o Ungl x 33. Figure 10. Germination of Big Bluestem in two kinds of soil. G 1 , G = northern grains in glacial soil; Ungl., R. = southern grains in residual soil; Ungl., G = southern grains in northern soil; and Gl., R. = northern grains in residual soil.

Figure 11. Germination of Little Bluestem in two kinds of soil. REAL GERM.% Real germ IOO 0 6 oo to 0 2 80 20 0 4 80 60 m° Ot 30 Out Rm.°C m* Out Rm.*C T N E M T A E R T T T tm ent rea Figure (I 10 e r u g i F 30 . Gerardi A. scoparius A 0 70 50 0 5 7 O 7 Ungl,G i G, R Gl,• 0 G Gl, Ungl.R x

36. stems of plants from different sources growing together in a fairly uniform environment may give some measure of the differential genetic constitutions of the plants under ob servation. Stem length of 600 plants was measured weekly, 150 of each species from each prairie area. Measurements were begun 39 days after the buds of Big Bluestem were first seen to enlarge, and the leaves of Little Bluestem first were observed to begin elongation. Height here is con sidered the distance between the soil level and the high est visible ligule. Figure 12 shows that inflorescenees of Big Bluestem from the southern prairies emerge at a somewhat greater height than those from the northern prairie area, although the average dates of flowering are almost identical. The height of Little Bluestem was the same at flowering in both areas, but those transplanted from the south flowered 26 days later than those from the north. Field observations bear out all these facts but the last. The time of flow ering of Little Bluestem on the Adams county prairies is about the same, or at the most a week later, than Little

Bluestem to the north. Presuming that floral initiation took place two months before the average flowering date in each strain, then that time of initiation was one during which the two strains were subjected to almost the same photoperiod. The 3 7

Figure 12. Stem elongation and flowering in Big and Little Bluestem. 3©.

160

140 y * * Ungl i A.A Gerardi 20 __o Gl

J= 100 o*

Ungl

60 A. scoparius 4 0

4 Mar 2 4 3 Apr I 3 2 3 2 Jun 2 2 58 88 128 158 178 No days in greenhouse

Fi ^urc ^ 39. artificial photoperiod had been exceeded by the natural

only by a few minutes. However, examination of the green house temperature record reveals quite different tempera ture reglfbes proceeding flowering. The Adams county strain was subjected to an average 13 hour dally temperature period 21cC. or above, while the northern strain was sub jected to an average 8 hour daily temperature period 21°C. or above. If the southern strain is one in which flower ing is favored by relatively higher temperatures than the strain to the north the temperature-flowering observations above may be correlated.

4. Morphological Variation

Three to four weeks after flowering, twenty plants were collected from greenhouse sods representing each prairie remnant. The number taken was divided as evenly as possible among the blocks per remnant. The specimens were pressed and certain characters noted or measured, many with a 10 X binocular dissecting microscope equipped with an ocular micrometer marked in tenths of a millimeter. The characters measured are listed below. Those preceeded by the letter Ng" were measured on Big Bluestem only, those measured on Little Bluestem only are preceeded by the letter Hs". Those unmarked were measured in both species. Reference to figures 13 and 14 will aid in inter preting the characters. 40. 1. Stem length to uppermost Inflorescence. 2. Number of nodes on stem. 3. Number of racemes. g 4. Length of the longest branch of the terminal raceme. e 5. Number of sessile splkelets per terminal raceme. 6. Longest trichome of the bearding at the base of the rachis. This and measurements below refer to a splkelet pair (sessile and pedlcelled) taken, in Big Bluestem, from the estimated middle of the longest branch of the terminal raceme; in Little Bluestem, from the actual middle pair taken from the terminal raceme. 7. Longest trichome of the bearding at the base of the second glume. 8. Length of the rachis. 9. Length of the glabrous base of the rachis, i.e. v’lth trichomes one-half mm. or less in length. 10. Ratip: length of the glabrous base of the rachis/ length of the rachis. s 11. Density of rachis bearding. Number of trichomes along l£ mm. of the rachis beginning l£ mm. from its base. 12.. Longest trichome of the rachis. 13. Length of the pedicelled splkelet. 14. Length of the pedicel. 15. Length of the glabrous base of the pedicel, i.e. with trichomes one-half mm. or less in length. 16. Ratio: length of the glabrous base of the pedi cel/length of the pedicel. s 17. Density of pedioel bearding. Number of trichomes along ijt mm. of the rachis beginning 1^- mm. from its base. 18. Longest trichome of the pedicel. Many of the above measurements were Also made on pop ulation samples of 20 plants from the same prairie remnant on which the sod transplants were made. Those, in turn, are compared with measurements taken from herbarium speci mens from the following states: Kansas, Missouri, Illinois, and Indiana. It should be pointed out that comparisons be tween population samples collected over a short period, and herbarium material collected over many years, give biased results (Lewis, 1947). Comparisons here therefore 41.

Figure 13. A pair of splke lets of Big Bluestem. The let ter "a" Is at the second glume Behind It Is the first glume. wb B Is at the rachis, "c" at the pedicel and "d" at the pediceled splkelet. Accession 31-4. Approximately 7X.

I mm.

Figure 14. A pair of splkelets of Little Bluestem. The letters refer to the parts as above. Accession 17-2. Approximately 7X.

I mm. 42. are made with caution. Table 2 shows the number of specimens of each type examined. The glacial map of Flint (1945) was used as an aid in determining the origin of out-of-state specimens with reference to glacial boundary.

Table 2. Number of specimens examined according to their source.

Ohio Ohio Prairie center greenhouse field samp. herb. spec.

Big Bluestem glaciated 133 140 73 unglaciated 132 120 40 Little Bluestem glaciated 98 67 109 unglaciated 143 140 61

Two methods of presenting the data are used. Mea surements for a particular character are grouped into size classes, the number per size class being given as the per centage of the sample. In figure 24, for example is plot- tea distribution of data from measurements of the longest trichome of the rachis (of a splkelet) of Big Bluestem. Plotted only are data from Ohio greenhouse sod transplants. The origin is indicated: Gl. * glaciated, i.e. plants from north of the glacial boundary, Ungl. = unglaciated, i.e. plants from the Adams County prairie area. On a work sheet measurements from plants from one prairie area were listed in increasing order of magnitude. The number of 43. plants per eats of sizes (size classes), for example 1.1 to 1.6 mm., was calculated. Since the number of specimens

from each prairie area Is not the same, 133 and 132 (see table 2), raw data from each size class is not comparable. Therefore the percent of the total per size class was cal culated. Points from the same prairie area are connected by a line. It is to be noted In figure 24, that similar curves are produced by handling the data as above. Only in the larger size classes are the populations differenti ated. Here the sample from the southern Ohio prairies are "offset" toward the larger size classes.

The other method employed la that of scatter dia grams. Averages of 20 samples from each prairie remnant are plotted, one character against another.

A. Big Bluestem

1. Size classes - Ohio material. Figures showing the size class distribution of data derived from measure ments on Big Bluestem are divided into three groups. The first group contains those with two curves for Ohio glacia ted and unglaciated prairie area samples. One sample was collected from greenhouse sod transplants, one was col lected in the field. These are figures 15 through 23. On each of the second group are plotted data from Ohio greenhouse samples, figures 24 through 29. 44. The third set shows the same characters as the second hut were based on measurements of Ohio field population samples and from herbarium material of plants from the

prairie center (figures 30 through 35). The distribution of size classes of the longest tri chome of the pedloel (figure 15) Is shown to be very uni form both to the north and to the south of the glacial border in Ohio. This character and the one to follow are distinctive In this group In that they show less variability In the greenhouse than when sampled under field conditions. The fleld-greenhouse relationship, Indicated in figure 16 Is the same but In this Instance the Marlon prairie area sample has slightly longer pedicelled splkelets than the sample to the south.

PlgureB 1? through 20 show measurements of characters in which the difference between samples from glaciated and unglaclated areas are about as great in the greenhouse as in the field. In either, differentiation with reference to the sampling area, is apparent.

Figures 21, 22, and 23 show certain measurements of characters which showed greater differences between green house samples than between field samples. In each there Is a slight, though unmistakable trend of variation. The data presented in figures 24 through 29 are from greenhouse population samples only. A consistent variation In most size classes between the two samples is shown by 45. figures 23, 24 and 25. The larger size classes, in figures 2?, 28 and 29, do not show the strict offset of the pre vious diagrams. In all, however, a distinction may easily

be made between the two population samples. The figures mentioned above (15 through 29) reveal the following differences between the Ohio Big Bluestem populations on unglaciated Adams County, Ohio and those populations of the same species north of the border. The statements below apply to the Adams county Big Bluestem population:

1. The pedicelled splkelet is slightly shorter (figure 16). 2. There are a few more racemes per plant (figure 37). 3. The pedicel usually is somewhat shorter (figure 19).

4. The glabrous base of the pedicel Is deficient In

the middle size classes (figure 20). 5. The stems are somewhat longer (figure 21). 6. Nodes are more numerous per plant (figure 22). 7. The longest trichome of the rachis is usually longer (figure 24).

8. The bearding at the base of the second glume is somewhat longer (figure 25). 9. Hachises are usually shorter (figure 26).

Pedicel and rachis ratios (figures 23 and 29, respec tively) , while having curves differing in some respects 46. according to source of the material, are not consistently differentiated.

Other characters measured (figures 15, 18, 27, and 28), especially the last, are less clearly differentiated.

2. Size classes - Ohio vs. Prairie center material. Prairie center material is compared with Ohio samples in figures 30 through 35. Figure 30 shows that, in the Ohio samples, bearding at the base of the second glume is some what longer than that from the prairie center. Moreover it would appear that in Ohio the samples from north of the glacial border have a higher percentage of larger size classes than In the sample from the south. In the prairie states the same is true.

Figure 31 shows that the length of the rachis is shorter in the prairie center than in the Ohio sample.

The curves also indicate that the Ohio sample from north of the glacial border has longer rachises than to the south.

In the prairie center the opposite Is true. Figure 32, shows that the length of the glabrous base of the rachis is somewhat shorter in the prairie center than in Ohio, and that the peaks exhibited by the curves for glaclated-unglaciated Ohio and the glaoiated-unglaciated prairie center are inverted with reference to one another.

From figure 33 it may be inferred that the length of 47.

the longest trichome of the rachis is somewhat greater In the prairie states than in Ohio. Little difference between Ohio and western samples can be interpreted from the curves exhibited by the length of the bearding at the base of the rachis (figure 34). From size class curves for the rachis ratio, figure 35, it may be seen that the ratio is somewhat smaller in the prairie states than in Ohio. Moreover, in the samples from south of the glacial border to the west, the ratio is somewhat smaller than to the north. In Ohio the opposite is generally true.

3. Ohio material - soatter diagrams. In plotting the average measurements of a particular character from plants of the Ohio prairie remnants (figures 36-43), one set of points is represented for each area, namely the re sults from plants which grew in the greenhouse. In all

instances, populations from north and south of the glacial border, but which grew together in a common environment, overlap. Field samples plotted, but not shown here, over lapped in some characters but did not in others. In no instance Is It possible to predict the two pop ulation positions. Nor is it possible to predict the po sition in either population of any particular prairie rem nant. The two variables Involved in each of these figures are, to a large degree, repetition of previous data. They are presented to emphasize the degree of variability within 48. and the differences between populations. The results of plotting height against number of nodes are shown by figure 31. Here the distinction between the two populations Is well differentiated and Is even more apparent than Indicated previously (figure 12).

A considerable degree of differentiation Is shown by figures 37, 38, and 39: number of Inflorescences vs. length of the longest branch of the terminal Inflorescence, the former vs. number of nodes, and length of the pedicel vs. length of Its glabrous base. Again overlapping Is ap parent.

Figures 40 and 41 characterize populations in which an increase in length of one character is generally accom panied by a similar Increase In length of the character against which it is plotted. In these the above is true of one population but not of the other. In the former it Is the Marlon prairie sample, in the latter the Adams county sample exhibiting this feature.

Figures 42 and 43 are diagrams of "sympatric" popu lations in which overlapping is, or almost is, 100 percent. When plotted, no other character combinations exhibited such little differentiation as is indicated by these. 49.

B. Little Bluestem

1. Size classes - Ohio material. Results of the

analysis of data derived from measurements upon Little Blue st em are presented below in the same fashion as previously.

In general, differentiation Is more concise than In Big Bluestem, a fact not surprising In view of the known vari

ability of the species. The figures may be grouped artificially as follows: figures 44 through 54 show data from four sets of Ohio

plants; two from each side of the glacial boundary, one each collected in the field and in the greenhouse. An

other group of figures (62 through 68) show differences In Ohio and prairie center field samples from glaciated

and unglaclated a±*eas. Accompanying each of these is a figure showing the same character measured on Ohio green house samples (figure 55 through 61).

Flgur« 44 through 49 are plotted variations in Little Eluestem In which the samples north and south of the gla cial border vary more from each other in the field than they do in a common greenhouse environment. In each a

slight but unmistakable difference may be seen In the two

latitudinal populations. In figures 47 and 48 variation in the field is particularly evident but when greenhouse

samples are taken they largely disappear.

Figures 50, 51 and 52 show variations that are as 50. evident in the field as in the common environment. In the latter environment however, the distinction between the Adams and Marion prairie area samples are none-the- less evident.

Figures 53 and 54 are measurements of characters (the latter figure is a ratio) which differ more widely in the common environment than In the field. In each, however, distinction may easily be made between the two populations. The results of plotting data for certain other charac ters shown by figures 55 through 61. In the first virtu ally no differentiation Is apparent. The others however show the offset curves characteristic of the species when data from it are plotted in this fashion.

From these data it is concluded that the differences listed below are evident between populations of Little Bluestem In the unglaclated Adame County area and those sampled from glaciated Ohio. Statements apply to the for mer. 1. The proportion of tall plants is greater (figure 44, cf. also figure 12).

2. The number of fertile spikelets per terminal

raceme I b greater (figure 45).

3. The pedicel is generally shorter (figure 46). 4. The pedicel bearding generally is less dense (fig ure 47). 5. The rachis Is slightly shorter (figure 48). 6. The number of nodes per stem Is greater (figure 51. 50) .

7. Generally there are fewer racemes per plant (fig ure 51) . 8. The sterile splkelet Is usually shorter (figure 53). 9. Glumes are shorter (figure 56). 10. The longest trlchome at the base of the raohls is usually shorter (figure 57). 11. The longest trlchome at the base of the second

glume is usually shorter (figure 58). 12. The longest trlchome of the raohls Is usually shorter (figure 59).

13. The glabrous base of the pedicel usually is longer (figure 60). 14. The longest trlchome of the pedicel is somewhat shorter (figure 61).

It is of interest also that the raohls ratio is some what greater in Adams county populations than in those to the north (figure 52), as is true of the pedicel ratio (figure 54).

It is apparent also that no differentiation is extant between the populations for the characters: degree of rachis bearding (figure 49) and length of the glabrous base of the rachis (figure 55). 52.

2. Size classes - Ohio vs. Prairie center material. By comparison of measurements of certain characters plotted for plants from the prairie center with Ohio samples it may be concluded:

1. O-lume length is somewhat greater to the west. In Ohio, plants from glaciated areas have glumes somewhat longer than those growing in the unglaci ated southern areas. This relationship is not evident in the prairie center (figure 62). 2. The longest trlchome at the base of the rachis Is generally somewhat longer in grasses of the prairie states than in those in Ohio; moreover, the length of the trlchome in the sample from the unglaoiated region to the west is somewhat shorter than that In the sample from glaciated areas. In

Ohio this is also true (figure 63). 3. The length of the longest trlchome at the base of the second glume is similar in all populations

(figure 64) with the exception of populations from the unglaciated prairie center, which is truncate in the middle size classes.

4. The length of the longest trlchome of the rachis (figure 65) of the prairie center sample is slightly longer than In the Ohio sample. More over the curves representing origin are disposed

with reference to one another dissimilarly. 53. 5. In general the Ohio material shows much wider variation in the length of the glabrous base of the pedioel than does the prairie center material. Also, a larger proportion of the samples to the west have shorter glabrous bases than Ohio plants, (figure 66). 6. The length of the longest trlchome of the pedicel (figure 67) is greater in the prairie center than in Ohio. The Ohio sample curve, based on material from glaciated areas, is offset toward the larger

size classes over a great range of size classes. The prairie center sample from glaciated areas is offset only in the largest size classes.

7. Generally, the length of the glabrous base of the rachis (figure 68) is greater in populations in the west than in Ohio. The Ohio-uralrie center populations in unglaclated areas, are greatly dissimilar, however.

3. Ohio material - scatter diagrams. Characters measured and plotted as population averages in scatter diagrams are shown in figures 69 through 77. The populations are quite distinct when certain characters are plotted, figure 69 through 72; others only share one common point, figures 73 and 74; some overlap but little, figures 75 and 76; and in others the degree 54. of overlapping ia considerable, (figure 77).

The greater degree of differentiation in this species is evident when compared to that in Big Bluestem (figures 36 through 43). 55.

DISCUS31ON AND CONCLUSIONS

The late Tertiary evolution and sorting of the grass land flora east of the Rocky Mountdns Is suggested by the findings of Ellas (1935). He reports 18 species of grasses (Agrostldeae and Panlceae) from deposits formed during the establishment of gravsland communities following the rise of the Rocky Mountains. This grassland expansion and the last (post glacial) one are the best substantiated. The post glacial eastward migration of prairie com munities as far as northeastern Pennsylvania has been dis cussed many times. An earlier eastward extentlon is sug- ge ted by Braun (1928a) on the basis of her interpretation of florlstic components of the Adams county prairie area.

More recently (1950) she interprets the presence of south eastern species in the Kentucky barrens as indicating a long period of occupancy, although the presence of barrens on non-pralrie soils and the rapid relnvaslon of forest following man's occupation is suggestive of a shorter period.

The most convincing evidence yet offered (Braun, 1928a) that points to early occupancy of the Adams county prairie area Is florlstic. Certain species, southern and western, are confined to the region, in Ohio, south of the glacial border: Ophloglossum Engelmannl. Agave virginlea. Leaven- worthla unlflora. Hexactrls splcata. Galacta volubllls. 56. and Draba cunelfolla. Other typical species are known In but one station north of the boundary in Ohio, and some times more in Indiana: Salvia lyrata. Lobelia leptostachys A. Dc., and Muhlenbergla ouapidata. The geologic placement of any early Pleistocene west ward movement of the prairie is pure conjecture. The Toronto beds, according to Flint (1947), are of Sangamon age. Shiraek (1940) after a thorough examination of the published evidence states that the beds show a "period of moderate climate similar to that of the present fin Iowa} Lane's study (1941) of Aftonian pollen in an Iowa deposit indicates a climate about as warm and dry as at present. Deevey (1949) has pointed out the need for stratigraphlc study of known deposits, which themselves are too few. The post glacial chronology developed from peat sampling now has an extensive literature mostly concerned with post-Plelstocene vegetation movement, these primarily within the ar--a of glaciation. The use of the Carolina bays (see Frey, 1953) may do much toward completing know ledge of forest sequence on the Coastal Plain. In the central deciduous forest, however, with its local relict communities, a concrete method of attacking problems of climatic shifts and their consequences are lacking. The evidence presented in the preceeding sections is suggestive of a slight although rather consistent differen tiation physiologically and morphologically between popu- 57.

latlons of Big and. Little Bluestem from Adams county and Marlon prairie areas. This variation might have arisen as a result of eootyplc differentiation or drift of an Iso lated population. Ecotyplc differentiation could have arisen, especially to the south, on the thin soil prairies of Adams county;

either in comparatively recent times or during occupation of the land surface throughout parts of the Pleistocene. Genetic drift, the random variation of gene frequen cies especially prominant in small populations (Wright from Dobzhansky, 1951), may have oocured since the Xerothermic Period, in as much as there are but few prairies on the Illinoian till (emphasized by Braun, 1935) between the main prairie areas of Ohio and those of Adams county. If these prairies have existed during much of the Pleistocene their Isolation was indeed extreme during the period that proglaoial Lake Tight covered much of the landscape. That the interval was great is evidenced by the 80 feet of silts deposited on the old lake floor (see Wolfe, 1942).

It Is more likely that these two types of variation have occured simultaneously in parts of the areas studied. That the two samples figured might be part of a lati tudinal pattern of a clinal nature Is not suggested here. Examination of materials from glaciated and unglaciated Ohio and from the prairie center, as indicated in figures 30 through 35, and 62 through 68, reveals that the popu lations bear no consistant relation to one another. For 58.

example, In figure 65, the sample from glaciated Ohio is offset toward the larger size classes compared to the sam ple from south of the glacial boundary. The prairie center samples are nearly superposed. The results of the germination tests on Big Bluestem does not suggest correlation of grain sources and percent germination with 1) cold treatment, or 2) soil type In which germination and Initial growth took place. Little Bluestem however shows a higher percentage germination of northern souroe grains when subjected to cold treatment. Also, percentage of germination is usually higher In the soil types taken from the same prairie area in which the grains are collected. Height of Big Bluestem is greater south than north of the glacial border. A hypothesis based on elimination of tall, late-flowering individuals to the north and their persistence south cannot be used in lieu of the data pre sented for Little Bluestem. In the latter species, the average height of the two populations at flowering are nearly identical. No differentiation has taken place in the time of flowering in Big Bluestem. The differential flowering dates of Little Bluestem are not b o m out by field obser vations, and the cause cannot be explained without use of more precise experimental techniques.

The evidence from the examination of chromosomes in meiosis reveals no difference in chromosome number or 59.

behavior in two populations of either species. Too few counts in Big Bluestem are available to ascertain Its chro mosome number over its entire range. All of the published records are of 30 pairs excepting Church's tetraploid strain reported for Anderson county, Kansas. Its distribution is not known.

The "splitting" perpetrated upon Little Bluestem is, as has been stated, ample evidence of Its variability and was at an early date overdone. It was Hubbard (191?) who reduced the varieties to three but in so doing suggested variety vllloslssimue as the typical. This view has been followed by Fernald (1935), and Gleason (1952). In 1923 Gleason stated that, among others, both Big and Little Bluestem are of eastern origin. Neither, the cytological work of Church (1936, 1940), and Nielson (1939), nor mor phological studies point conclusively to an eastern origin of the species. Larsen (1947) suggests "low latitude" origin for Little Bluestem and two routes of migration north, "through the central grassland region and up the Atlantic Coastal Plain".

The slight morphological differences between the two samples of Ohio material of Big and Little Bluestem suggest simple random variations In gene frequency. This variation Is particularly evident in the former in such characters as: pedicel length (figure 19), stem length (figure 21), length of the raohls (figure 31), length of the glabrous base of the raohls (figure 32), and bearding at the base 60.

of the rachis (figure 34). Pedicel bearding (figure 15), on the other hand, shows virtually no differentiation. Well differentiated characters in Little Bluestem, shown by figures 45, 46, 50, 53, 54, and 56 through 61, contrast sharply with the virtual lack of difference in degree of pedicel bearding (figure 47), length of the rachis (figure 48), and length of the glabrous base of the rachis (figure 55).

A comparison of figures 30 through 35 reveals, in Big Bluestem, that In half of the sets of measurements (figures 31, 33, and 35) the Ohio-Prairle center samples from north of the glacial border are more nearly alike than south of the boundary from both areas. In the rest, figures 30, 32,

and 34, the two populations from south of the glacial bor der are more closely related than those to the north.

This phenomenon suggests a closer relationship between glaciated Ohio-prairie center plants and unglaciated Ohio-

prairie center plants than Ohio glaciated-pralrle center unglaclated and visa versa.

Examination of figures 62, 63, 65, 66, and 68 of Little Bluestem reveals that, in all instances the Ohio- pralrle center populations from north of the glacial border are more alike than the Ohio-prairie center populations to the south. This fact, when coupled with seeming ecotypic development of the Ohio Little Bluestem previously discussed, is suggestive of a greater "opportunity" for evolutionary divergence on the Ohio prairie area south of the glacial border. Whether the opportunity was In the form of "isolation", or due to influx of new genetic ma terial during a climatic shift is not In evidence. 3uch problems, related to the ones mentioned on page 57, await further investigation. 62.

SUMMARY

The interpretation of relict vegetation in time and apace is one of the most difficult problems in ecological plant geography. In Ohio, one of the most prominent and well known communities is the prairie. This study is con cerned with infraspecific variation In two dominant species of grasses, Big and Little Bluestem, from contrasting Ohio prairie areas. One sampling area is the Marion prairies on deep soils of glaciated north-central Ohio; the other is the Adams county prairie area characterized by shallow soils in southern Ohio south of the Illinoian glacial bor der.

Four methods have been used in this study. They are 1) determination of chromosome number and behavior, 2) comparison of results of grain germination in the two species from both areas, 3) measurement of stem elongation after growth was resumed by plants removed from the botanic garden to the greenhouse, and 4) morphological comparisons. In the last, population samples from both areas are com pared with each other and with herbarium specimens collected on both sides of the glacial boundary in the prairie cen ter. Comparisons are also made between sod transplants from both Ohio prairie areas which grew together in a com mon greenhouse environment. Chromosomes numbers (found) are those commonly reported 63. for the speclee. They are: 2n - 60 in Big BlueBtem, and

2n = 40 in Little Bluestem. All chromosome pairs appeared to behave as diploids. Results from the planting of over 20,000 spikelets reveal, generally, that in both species the real germina tion percentage is higher when grains are subjected to 30 to 50 days at 10°C than with either more or less cold treat ment. In Big Bluestem virtually no difference, according to source, in real germination percentage exists. Embryos of Little Bluestem from south of the glacial border ger minate less well with cold treatment than those to the north. Also in contrast to Big Bluestem, the other species exhibits a higher percentage germination when planted in soil from the same prairie area from which the grain was collected.

Data from the measurement of stem elongation, which continued until flowering occured, reveal that Big Blue stem from the south flowers at a slightly greater height than the same species from the north. Flowering occurred at about the same time. The two strains of Little Bluestem flowered at the same height but those from the south flowered 28 days later than their northern counterpart. A differential temperature regime at the time of floral initiation may be responsible for the wide variance In flowering date in this species. Results from the measurement of many morphological 64. characters in Big Bluestem reveal that in the Adams county populations: 1. Stems are somewhat longer. 2. There are more nodes per plant. 3. There are a few more racemes per plant. 4. The rachis is usually shorter. 5. The longest trlchome of the rachis is usually longer. 6. The pedicelled spikelet is slightly shorter. 7. The pedicel is somewhat shorter. 8. The longest trlohome at the base of the second glume is somewhat longer.

In the Adams County Little Bluestem population; 1. The number of nodes per stem is greater. 2. There are generally fewer racemes per plant. 3. The number of fertile spikelets per terminal ra ceme is greater.

4. The longest triohomes at the base of the rachis and second glume are shorter. 5. The rachis and its longest trlchome are usually shorter. 6. The sterile spikelet is usually shorter. 7. The pedicel is shorter although its glabrous base is longer.

8. The pedicel bearding is generally less dense. 9. The longest trlchome of the pedicel is somewhat shorter. 65. 10. The glumes are generally shorter. These slight morphological differences between the two samples of Ohio Big and Little Bluestem suggest random

variation in gene frequency.

Where comparisons are available between Big Bluestem from Ohio and material from the prairie center It is noted

that in almost half the sets of measurements the Ohio- prairie center glaciated populations are more nearly alike than the unglaclated from both areas. In the rest, the two populations from south of the glacial boundary are more closely related than those to the north. This phenomenon suggests a closer relationship between the northern Ohio- prairie center plants and southern Ohio-prairie center plants than between the plants from north of the glacial boundary In Ohio and south on the prairie center and visa versa.

Data derived from measurements on Little Bluestem suggest a closer relationship between the Ohio-prairie cen ter plants from north of the glacial boundary than between those from the same regions south of the boundary. This fact, when coupled with seeming eootyplc development of the Ohio Little Bluestem, previously discussed, is sugges

tive of a greater "opportunity" for evolutionary divergence

on the Ohio prairie area south of the glacial border. Whether the opportunity was in the form of "isolation", or due to influx of new genetic material during a climatic shift is not in evidence. Such problems await further investigation. Figure 15. In this and other figures the first let ter of a pair refers to source of the specimen: from a northern glaciated area, O; or an unglaelated region, y. .Of the second letters, 0- = greenhouse oollected; F * oollected in the field (Ohio); and P • prairie center e.g. in figure 30. Here size class percentages are plotted for the longest trlchome of the pedicel (in mm.). Field and greenhouse samples appear together In figures 15 through 23.

Figure 16. Size class percentages are plotted for the length of the pedicelled spikelet (In mm.).

Figure 17. Size class percentages are plotted for num ber of racemes. GG 50 A. G erardi 40 F i S 30 o

*20 UG

4 -9 10-1.5 I.6-2.I 2.2-27 20-33 34-394.045 P e d ic e l, longest trichome

A. Gerordi 40

- 5 . 3 0 U) 20

48-57 58^.7 68-77 78-87 8097 88-107 108-117 II.8H27 128-137 (38- Length pediceHed spikelet

4C A. Gerardi

2 0 -

GG o'j

0-1 2 - 4 5 -7 8-10 11-13 14-16 17-19 20 22 Number of racemes Figure 18. Size class percentages are plotted for the length of longest branch of the terminal raceme (in cm.).

Figure 19. Size class percentages are plotted for the length of the pedicel (in mm.). % of sample % of sample . 0 4 0 3 20 10 49 0596* 97079 6*. 99 Q19 MI9 2l. Ir3 14-149 I3rl39 12rl2.9 I.9 IM IQ-109 9 *9 0 9 60*0.9 9 7 0 7 .9 60*6 50*5.9 9 -4 0 4 *4 540 .-. 475. 8 5*. 65*70 59*6.4 .8 5 3 5 .2 5 7 4 4.1-4.6 35*4.0 9*34 2 UF eie, length Pedicel, GF Tarmtnal Tarmtnal GG UF aee length raceme, i ^ Fi GG . 0 7 Figure 20. Size class percentages are plotted for the length of the glabrous base of the pedicel (in ram.).

Figure 21. Size class percentages are plotted for stem length to the uppermost inflorescence (in cm.) of sample ^ somPle 50 50 • 40 40 . 20 30 . Gerardi A. 5 0 - 6 .9 7 0 0 9 9 C H 0 9 IICH29 I3 rl4 9 I5.-I6 .9 17-109 19.-209 21:229 19.-209 17-109 .9 I5.-I6 9 rl4 I3 IICH29 9 0 H C 9 9 0 0 7 .9 6 - 0 5 7 011 215 619 23 2.4- 3 -2 0 2 16-1.9 12*1.5 .0-1.1 7 - 4 3 UF UF eie, egh f lbos base glabrous of length Pedicel, l ’ G G egh o pe inflorescence upper to Length GF UG 72.

Figure 22. Size class percentages are plotted for number of nodes on the stem.

Figure 23. Size class percentages are plotted for the pedicel ratio. % of sample % 0f sample 0 5 0 4 20 0 5 IO - 0 4 0 3 20 3C P a i* egh lbosbs/egh pedicel base/length glabrous length Ratio* . Gerardi A. 56 8 91 1-2 31 15-16 13-14 11-12 9-10 -8 7 a5-6 -09 JO-18 -09 GGo ubr f nodes of Number f-T f-T .19-27 . 2 8 - 3 6 .37:45 .37:45 6 3 - 8 2 . .19-27 OfjurC- S 2 . Gerardi A. UF G U 46-54 Figure 24. Shown In figures 24 through 29 are popu lations from greenhouse samples only. Size olass percentages of the length of the longest trlohome of the rachls (in mm.) are shown here.

Figure 25. Size class percentages are plotted for the length of the longest trichome of the bearding at the base of the second glume (in mm.). of sample % of sample 0 5 0 4 20 0 4 0 3 50 0 3 71 1 .-. I72I 22 273I 32-36 2.7-3.I 22*26 I.7-2.I 1.2-1.6 .7-1.1 51 111 172 2-8 93 35-4.0 29-34 23-28 1.7-22 1.1-16 .5-10 erig t ae f eod glume second of base at Bearding UGx a h s lnet trichome Rachis, longest A.Gerardi A. Gerardi Figure 26. Size class percentages are plotted for length of the rachis (in mm.).

Figure 27. Size class percentages are plotted for length of the longest trichome at the base of the rachis (in mm.). of sample % of sam ple 0 3 0 4 20 20 0 3 .-6 34 35- 1-8 74 75- 2 -0 5 7 4 -7 7 6 6 6 * 9 5 -58 51 0 5 - 3 4 2 -4 5 3 4 -3 7 2 1.9-26 6- 911 I 517 20 2J- 26 27* 32 *3 0 3 9 *2 7 2 6 -2 4 2 3 -2 J 2 0 -2 8 L 15-1.7 -I4 2 L .9*1.1 -8 .6 egh f rachis of Length i r. 7 Z Fig ure. erig t ae f rachis of base at Bearding ,UG UG . Gerardi A. . erardi G A. * 7 Figure 28. Size class percentages for the length of the glabrous base of the rachis (in ram.) are plotted.

Figure 29. Size class percentages are plotted for the rachis ratio of sample % of sample 0 4 20 0 3 20 0 3 0 5 0 4 14 .52 2330 33 39-46 47* 62 6370 71-78 0 7 3 6 2 -6 5 5 4 *5 7 4 6 4 - 9 3 3H38 0 3 3 2 .15*22 4 -1 7 0 6 0 - 0 2 . 7 8- 7 L822 2327 28* 33*37 2 *3 8 2 7 2 3 2 2 2 8 L I7 3 L 2 -1 8 -7 .3 :2 0 R a tio : length glabrous bose bose glabrous length : tio a R GG ahs lnt o garu base glabrous of length Rachis, *UG Fijart Fijart St 9 . Gerardi A. . Gerardi A. / egh rachis length . a 9 Figure 30. In this and those through number 35 Ohio field sampleb are plotted with those collected in the prairie oenter. 31ze class percentages are plotted for the length (in mm.) of the longest trlchome at the base of the second glume.

Figure 31. Size class percentages are plotted for the length of the rachis (in mm.). % of sample y % of sample ro oj o O O o o t 5 8 g

,OK‘

o xo o 3" to I

0 ro Figure 32. Size class percentages are plotted for the length of the glabrous base of the rachis (in mm.).

Figure 33. Size class percentages are plotted for the length of the longest triohoige of the rachis (in mm.). of sample % sample 60 0 5 20 40 30 0 5 40 20 30 - 37 81 131 182 2-7 83 3*7 38-4.2 33*37 28-32 1.8*22 23-27 1.3-17 .8-12 3-7 0-2 -0 .-6 72 23- 34 3-. 41- 35-4.0 4 -3 9 2 8 -2 3 2 17-22 1.1-16 5-10 UP* ahs lnet rco e trichom longest Rachis, ahs lnt o garu base glabrous of length Rachis, iGP iue 32. figure . erardi G A. 3 3 . Gerardi A. UP Figure 34. Size class percentages are plotted for the length of the longest trlchome at the base of the rachis (in mm.).

Figure 35. Size class percentages are plotted for the rachis ratio. of sample % of sample 30 0 4 0 3 UF: 68 91 121 1-. 182 2-3 42 2*9 D3 33*35 3D-32 27*29 24-26 2J-23 1.8*20 15-1.7 1.2-14 .9-11 .6-8 06 0-4 52 33 .13 3*6 75 5*2 37 77 79- 7178 63-70 55*62 47-54 39*46 .31-38 23*30 1522 .07-14 6 -0 0 ai lnt garu bs / lnt rachis length / base glabrous length Ratio erig t ae f rachis of base at Bearding 34* . Gerardi A. . Gerardi A. 6£ Figure 36. This figure and those through number 43 are scatter diagrams. X*s are population averages from samples collected from greenhouse plants original ly collected south of the glacial border. 0's are originally from the Marion prairie area. Here number of nodes is plotted against length to uppermost in florescence (in cm.).

Figure 37. Number of inflorescences is plotted a- galnst length of the longest branch of the terminal raceme (in cm. ). Number of inflorescences Number of nodes 12 10 IO 9 11 8 8 1 10 3 10 5 10 170 160 150 140 130 120 110 $ egh o pe inflorescence upper to Length 3$ 8 1 I 12 II 10 9 8 7 egh ftria rcm branch raceme ofterminal Length o X o ° o Gl. O X X o X X ° Gl.o o

o o A. Gerardi x x Ungl. . Gerardi A. Ungl. x t f Figure 38. Number of inflorescences is plotted against number of nodes.

Figure 39. The length of the glabrous base of the pedicel Is plotted against the length of the pedicel (both in mm.). Number of inflorescences CO O

T I* T) ^ Z (D 3 O’ O Q.

5 * * g C ° ° O n o o O -I -I 1 * r - ^ 00 O o

Pedicel, Pedicel, length of glabrous base

J K K J s r s> SI- - -I ro £ O r o x x Ot 39 Pedicel, length Figure 40. The length of the longest trlchome at the base of the second glume Is plotted against that at the base of the rachis (both In ram.).

Figure 41. The length of the rachis Is plotted against the length of the glabrous base of the rachis (both In mm. ). 23 92

% 2 2 oo o c O _r 2 .1 oo 0 61 o» ■d CVJ a> o u> c o tt o -O S 1.9 - 0 o X X x Ungl

A. Gerardi 1 * i . 1.2 1.4 1.6 1.8 2 0 Bd., base ra ch , Ion. tri o 5.2

50 o x

£ 4.8 o* 4\ a> «/>r 4 6 A x Ungl u o “= 4 4 0 X x x 0 0 6 i 4 2 A Gerardi

.7 9 II 13 15 Rachis, length glabrous base Figure 42. The length of the glabrous base of the pedicel Is plotted against the length of the glabrous base of the rachis (both In mm.).

Figure 43. The longest trichome of the bearding at the base of the rachis is plotted against the longest trichome of the rachis (both in mm.). F

Beard, base of rachis, longest trichome c v w ij

2Al Pedicel, length of glabrous base

2 ahs lnt o garu ba e as b glabrous of length Rachis, 42 a> 8 ro Z

Hr* Figure 44. Field and greenhouse samples are dotted together in figure 44 through 54. Here size class percentages for stem length (in cm.) are plotted.

Figure 45. Size class percentages are plotted for number of fertile spikelets of the terminal raceme. % of sample 20 0 4 60 50 30 20 60 50 70 - 56 - -0 11 1-4 15-16 13-14 11-12 9-10 7-8 5-6 3-4 scoparius A ubr f erie spikelets rtile fe of Number GGo Fi| u r< scoparius A UF e ength g n le tem S 70-89 ^ ^ 6 90-10950-69 UF 130-149 . 6 9 Figure 46. Size class percentages are plotted for pedicel length (in mm.).

Figure 47. Size class percentages are plotted for density of pedicel bearding. % of sample % of sample 20 0 6 0 5 0 3 0 5 0 4 0 3 49-5.5 2 70-76 77-8.3 . 8 - 7 7 6 7 - 0 7 9 . 6 3 . 6 2 6 - 6 . 5 5 . 5 - 9 4 8 . 4 - 2 4 1 . 4 - 5 3 7 - 2 1 * 0 e r e f eie bearding pedicel Degree of 8-13 e i e lengthPedicel 14-19 7 3 - 2 3 1 3 - 6 2 5 2 - 0 2 GFb . scoparius A. G G o . scopariusA. 96. Figure 48. Size class percentages are plotted for length of the rachls (in mm.).

Figure 49. Si^e class percentages are plotted for density of rachls bearding. of sample % of sample i gre f ahs di g in rd a e b rachis of ree eg D 9 4 e r u fig 100 60 50 20 30 80 20 40 ■ 40 - 0 6 Fijwvc^s 4 51 1-0 12 2 36 29 21-28 13-20 5-12 -4 0 -26 2.7-37 3.8-48 49*5.9 6 0 7 0 71-81 82-9.2 93H03 82-9.2 71-81 0 7 0 6 49*5.9 3.8-48 2.7-37 -26 egh f rachis of Length . scoparius A. oGF . scoparius A. UF xU oGG toe . 101.

Figure 50. Size class percentages are plotted for number of nodes on the stem.

Figure 51. Size class percentages are plotted for number of racemes per plant. % % of sample sample 40 20 50 30 40 20 60 30 50 iue o ubr f nodes of Number fo Figure *wC ubr f racemes of Number P*SwrC 9 1-9 02 3-94*9 09 06 77 80-89 7079 60-69 5059 40*49 30-39 20-29 10-19 -9 0 * 78 *0 11 1*4 6*16 13*14 11*12 9*10 7-8 5*6 GG oGF . scopariusA. G(G o scoparius A U,G *oz. Figure 52 . Size class percentages are plotted for the rachls ratio.

Figure 53. Size class percentages are plotted for the length of the sterile splkelet (in mm.). of sample % 0f sample 0 5 0 5 20 0 3 20 0 4 0 3 ai1 egh lbos ae egh rachis length / base glabrous length Ratio1 08-17 .8-1.9 20-3.1 3.2-43 4.45.5 3 6 -6 7 6.8*79 8.0-91 92-103 IG 4 I 5 I 4 IG 92-103 8.0-91 6.8*79 7 -6 6 3 4.45.5 3.2-43 20-3.1 .8-1.9 . 18-27 ,ue 3 6 p,£ure . UF» 28-37 trl siee, ength g n le spikelet, Sterile

38*47 UF

. scoparius A. 7 5 8 4 . scoparius A. . 58-67

|04* Figure 54. Size class percentages are plotted for the pedicel ratio.

Figure 55. Figure 55 through 61 are plotted for greenhouse samples only. Here size class oercentages are plotted for length (In ram.) of the glabrous base of the rachls. % of sample % °f sample 50 40 20 50 20 40 30 • a i- egh lbos ae length glabrous lengthpedicel base / Ratio- - .8-14 .1.5-21 0-7 2 2-2.8 2.9-36 37*4.3 4.4-5.0 1 .16-23-15 24-.3I 32-3940 47 48-55 56-63 ahs lnt o garu base glabrous of length Rachis, F*j >r*. S' 4* GG P ij«r< 5 5 f P ij«r< GF A. scoparius scopariusA xUF / 06 . Figure 56. Size class percentages are plotted for first glume length (in mra. ).

Figure 57. Size class percentages are plotted for the length (in mm.) of the longest trichome at the base of the rachls. of sample 0 3 2 0 5 j i F 0 4 20 0 4 - 0 3 5 60- 77 74- 18 88* 5I IQ2-IQ8109- I 95-IQ 4 *9 8 8 81-87 0 -8 4 7 67-73 6 -6 0 6 -59 4 5 $7 35 8 .-. 121 1517 .-. 2.1-2.3 I.8-2.0 1.5-1.7 1.2-14 .9-1.1 -8 6 .3-5 B ase of ra ch is, is, ch ra of ase B UGx egh f ume m lu g of Length tGG ogs trichome longest . i s riu a p o c s A. . scoparius A. /oe. Figure 58. Size class percentages are plotted for the length (in mm.) of the longest trichorae at the base of the second glume.

Figure 59. Size class percentages are plotted for the length of the longest trichome of the rachls (In mm. ). **— 6 a. o CO o % of sample 50 40 20 30 50 40 30 Fijwm a erig bs o scn glume second of Bearding, base ca .5-8 9 S -2.0 ogs tihm o rachis trichome of Longest UGx 21-28 912 316 .-0 .-4 52 29-32 25-28 2.1-24 1.7-20 13-1.6 .9-1.2 2.9-3.6 37-4.4 4.5-5 2 5.3-6.0 2 4.5-5 37-4.4 2.9-3.6 GG GG . scopariusA. A. scoparius //a Figure 60. Size class percentages for the length (in mm.) of the glabrous base of the pedicel.

Figure 61. Size class percentages for the length (in rani. ) of the longest trlchorae of the pedicel. of sample o t 50 40 40 30 20 awed Pdcl erig lnet trichome longest bearding, Pedicel d e w la f Fi«vr< co 37 812 .-7 1.8-2.22B-3.2 33r3.7 1.3-17 2.3-27 .8-1.2 .3*7 11-16 17-2.2 23-2 8 2.9*34 35-40 4.1-4.65.3-58 4.7-52 35-40 2.9*34 8 23-2 17-2.2 11-16 eie, egh f lbos base glabrous of length Pedicel, GGo UG oGG . scoparius A. scoparius A H2. Figure 62. Figure 62 through 60 are of Ohio field population samples and prairie-center samples. Here glume length (in mrn. ) is given as size class oercentages.

Figure 63. Size class percentages are plotted for the length of the longest trichorae at the base of the rachis Tin mm. ). % of sample 30 3 5 0 5 iwc3 ae f ahs lnet trichome longest rachis, of Base pi^wrc63 0 5 0 4 0 4 20 20 0 3 10 i e 62 Lnt o glume of Length 2 6 re u fi^ 45 60- 77 7-0 .-7 8UB-94 95*10.1 8.1-87IQ2-I08IQ9- 74-80 67-73 6 -6 0 6 54-59 3r.5 .6 -8 .9-1.1 1.2-14 15-1.7 1.8-2.0 2.1-2.3 1.8-2.0 15-1.7 1.2-14 .9-1.1 -8 .6 3r.5 ; gp oGF p u A. scoparius A. scoparius Figure 64. Size class percentages are plotted for length of the longest trichome at the base of the second glume (in mm.).

Figure 65. Size class percentages are plotted for the length (in mm.) of the longest trichome of the rachis. o i a> o. of sample 0 5 0 4 0 5 3V Badn, ae f eod glume second of base Bearding, l3uV* 0 3 20 0 4 so .-. 2128 .-6 .-4 .*2 360 6 .3 5 4.5*52 3.7-44 2.9-36 1-2.8 2 I.3-2.0 .5-8 9-1.2 L3-1.6 1.7-20 2.1-24 2 5 '2 8 2 9 32 9 2 8 '2 5 2 2.1-24 1.7-20 L3-1.6 9-1.2 .5-8 Longest trichome of rachis UFx oGF OF UP GP . scoparius A. . scoparius A. UP Figure 66. Size class percentages are plotted for length (in mm.) of the glabrous base of the pedicel.

Figure 67. Size class percentages are plotted for the length (In mm.) of the longest trichome of the pedicel. % % of sample 30 E ° 50 0 4 20 0 4 pPe , ^ u d r i e c e 6? lb e a r d i n g ,l o n g e s tt r i c h o m e 0 3 20 F juk u ij 6 pdcl lnt o garu base glabrous of length pedicel, 66 .- . 17222328 .-43*. 4I46 752 5.3*58 47*5.2 4.I-4.6 35*4.0 1.7-2.2 2.9-34 1.1-1.62.3-2 8 3 -7 B-12 1.3-17 I.8-2.2 2.3-27 2JB-3.2 33-37 2.3-27 I.8-2.2 1.3-17 B-12 -7 3 UF* . scoparius A. oGP . scoparius A. Figure 68. Size class percentages are plotted for length (in mm.) of the glabrous base of the rachis.

Figure 69. The remainder of the figures are scatter diagrams plotted as were numbers 36 through 43. Here length of the rachis is plotted against length of the longest trichome of the rachis (both in mm.). /Zo. UFx

6 0

50 p G P ,GFo

on A. scoparius 30

» — 0 -

0 -7 8-14 15-21 22-28 2 .9-36 37-43 44-50 pijureCa Rachis, length of glabrous base

o o X 0 0

°GI o cu X X Ungl

A scoparius

2.6 2.8 3.0 3.2 3.4 3 6 3 8 4 0 F ig u re 6 3 Longest trichome of rachis 121.

Figure 70. Stem length (in era. ) Is plotted against glurne length, (in mm.).

Figure 71. Length of the longest trichome of the pedicel Is plotted against that of the rachls (both In mm. ). tzx.

f (|WV<

A. scoparius I 90 tn

® 80 Ungl** o» S 70 o Gl.

(A O 30 X X □ or X x Ungl. X 22 2.6 3.0 3.4 3.8 Pedicel, longest trichome

Fijwt-e *11 Figure 72. The length of the longest trichome at the base of the rachls Is plotted against glume length (both In mm.).

Figure 73. The density of rachls bearding Is plotted against the length of the longest trichome of the rachis (In mm.). 0* Longest trichome, base of rachis c Degree of rochis bearding ^ * to Ot to OJ * N ® T ur« -T-

■J 0) length Glume 72 Ul . 7 8 8 8 92 88 84 80 76 7.2 o N

r o c 3 * scoporius o> 3 a to X

w ok X (A o * f> X o o O c o 3 ■o 3 (A o C c CA o CD 01

Figure 75. The number of racemes Is plotted against number of nodes. / * 6 .

a> a« / > J3 (A A. scoparius §22 -O O Ungl. * in 20 X X £ 1.8 o* c aj X x> Gl y tro 2 8 30 3.2 34 36 38 4 0 Figure 74 Longest trichome of rachis x

30 • • • 9 n in h i 2E y 20 <

10 0" Gl x -Ungl A scoparius

— t 9 0 10.0 11.0 12.0 NO. NODES 127.

Figure 76. Stem length (in cm.) Is plotted against number of nodes.

Figure 77. Number of racemes is plotted against number of spikelets per raceme. .o o E a> Ave. no. rocemes ° o CO 10 20 0 3 II i o c 4 egh ofstem Length Fijofc 74 lO A. scoparius 0 9 0 8 0 7 0 6 scoparius A ? 7 v. o s klt / / ikelets sp no. Ave. X o Gl. OX) UngI. x 8 x * * X 100 raceme x«Un gl * Gl o*

Z . /ZS 10 129.

LITERATURE CITED

1. Anderson, E. 1941. The technique and use of mass collections In plant taxonomy. Ann. Missouri 3ot. Card. 28: 287-292. 2. Anderson, K. L. 1940. Effects of inbreeding little bluestem. Jour. Amer. Soc. Agron. 32: 159-160. 3. ______and A. E. Aldous. 1938. Improvement In An'dropogon scoparius Michx. by breeding and selection. Jour. Amer. Soc. Agron. 30: 862-898. 4. Atwater, C. 1819. On the prairies and barrens of the west. Amer. Jour. Sci. 1: 116-125. 5. Babcock, E. B. 1947. The Genus Crepls. I. Univer sity of Calif. Publ. Bot. Vol. 21. 6. Baker, H. G. 1952. The ecospecles - prelude to dis cussion. Evolution 6: 61-68. 7. Bews, J. W. 1929. The world1s grasses. Longmans, Green and Company. London. 8. Bird, R. D. 1927. A preliminary ecological survey of the district surrounding the entomological station at Treesbank, Manitoba. Ecol. 8: 207- 220. 9. Blake, A. K. 1935. Viability and germination of seeds and early life history of prairie plants. Ecol. Mon. 5: 405-460. 1 0 . Braun, E. L. 1928a. Glaoial and post-glacial plant migrations indicated by relict colonies of south ern Ohio. Ecol. 9: 284-302. 11. _____ . 1928b. The vegetation of the Mineral Springs region of Adams County, Ohio. Ohio Biol. Surv. Bull. No. 15.

12. 1935. Affinities of the flora of the Illlnolan till plain of southwestern Ohio. Rhodora 37: 349-361.

13. ______. 1950. Deciduous forests of eastern North America. The Blaklston Corapany7 iP'hiTa- delphla. 130.

14. Bruner, W. E. 1931. The vegetation of Oklahoma. Ecol. Mon. 1: 99-188. 15. Church, G. L. 1929. Meiotlc phenomena In certain graminea. II. Paniceae and Andropogoneae. Bot. Gaz. 88: 66-84. 16. ______. 1936. Cytologlcal studies In the Cramlneae. Amer. Jour. Bot. 23: 12-15. 17. . 1940. Cytotaxonomlc studies In the (TFSfflYJTCSe - Spartlna. AndropoRon. and Panloum. Amer. Jour. 6ot. F 2^3-271. 18. Clausen, J . , D. D. Keck, and W. fcl. Hiesey. 1940. Experimental studies on the nature of species. I. The effect of v ried environments on wewtern North American plants. Cam. Inst. Wash. Publ. 452. 19. ______1945. Experimental studies on the nature of- species. II. Plant evolution through araphl- ploidy and autoploidy, with examples from the Madllnae. Cam. Inst. Wash. Publ. 564. 20. ______1948. Experimental studies on the nature of species. III. Environmental responses of climatic races of Achillea. Cam. Inst. Wash. Publ. 581. 21. Clements, F. E., E. V. Martin and F. L. Long. 1950. Adaptation and origin In the plant world. The role of environment In evolution, fhe Chronica I^otanica Company. WaTtliam, Mass. 22. Cornelius, D. R. 1947. The effect of source of little bluestem seed on growth, adaptation, and use In revegetation seedings. Jour. Agr. Res. 74: 133- 143.

23. Coukos, C. J. 1944. Seed dormancy and germination in some native grasses. Jour. Amer. Soc. Agron. 36: 337-345.

24. Curtis, J, M. and M. L. Partch. 1950. Some factors effecting flower production in Andropogon Gerardi. Ecol. 31: 488-489. ------

25. Davidson, J. F. and Paul F. Romberg. 1952. Genetic variation In Andropogon. Madrono 11: 203-206. 131.

26. Deevey, E. S. 1949. Biogeography of the Pleistocene. Bull. Geol. Soc, Amer. 60: 1315-1416, 27. Dobbins, R. A. 1937. Vegetation of the northern "Virginia Military Lands" of Ohio. Dlse., Ph.D. Ohio State University. (Unpublished) 28. Dobzhansky, T. 1951. Genetics and the origin of species. Third Edition. Columbia University Press, New York. 29. Ellas, M. K. 1935. Tertiary grasses and other prairie vegetation from High Plains of North America. Amer, Jour. Sci., Ser. 5. 29: 24-33. 30. Faegri, K. 1937. Some fundamental problems of tax onomy and phylogenetics. Bot. Rev. 3: 400-423. 31. Fassett, N. C. 1941. Mass collections: Rubus odoratus and R. parviflorus. Ann. Missouri Bot. SarST "£0: 299337T;

32. Fernald, M. L. 1950. Cray1s Manual of Botany. Eighth edition. American Book Company. New York. 33. ______and L. Qriscom. 1935. . Three days of botanizing in southeastern Virginia. Rhodora. S7: 129-157, 167-189.

34. Flint, R. F. 1947. Glacial geology and the Pleisto cene Epoch. John Wiley and Sons, Inc., tfew York.

35. ______and others. 1945. Glacial map of North America. Geol. Soc. Amer. Spec, papers 60. 36. Frey, D. G. 1953. Regional aspects of the late- glacial and post-glacial oollen succession in southeastern North Carolina. Ecol. Mon. 23: 289- 313.

37. Gleason, H. A. 1909. Some unsolved oroblems of the orairles. Bull Torrey Bot. Club. 36: 265-271. 38. ______1923. The Vegetatlonal history of the Middle West. Ann. Assoc. Amer. Geogr. 12: 38-85.

39. ______1952. The new Britton and Brown lllus- trated flora of the northeastern UnTCed Stated and adjacent Tanada. The Mew York Botanical'"Gar den, New York. 132. 40. Goldthwaite, R. P. 1952. What the glaciers did to Ohio. Ohio Div. Geol. Surv. Eauc. Leaflet No. 3. 41. Greene, H. C. and J. M. Curtis. 1950. Germination studies of Wisconsin prairie plants. Amer. Mid. Nat. 43: 186-194. 42. Gregor, J. W. 1938. Experimental taxonomy II. Initial population differentiation in Plantago marltlma L. of Britain. New Phytologiet. 37: T 5 - 4 9 . — 43. 1939. Experimental taxonomy IV. Pop- U T 5 T T 0 T Y differentiation in North American and European sea plantains allied to Plantago marltlma L. New Phytologiet. 38: 293-322. 44. ______1944. The ecotype. Biol. Rev. 19: 20-30. 45. ______1946. Ecotypic differentiation. New Phytologiet 45: 254-270. 46. V. M. Davey, and J. M. S. Lang. 1936. Experimental taxonomy I. Experimental garden technique in relation to the recognition of small taxonomic units. New Phytologiet. 35: 323-350. 47. Hall, H. H. 1926. The taxonomic treatment of units smaller than species. Proc. Internet. Cong. Plant Science (Ithaca) 2: 1461-1468. 48. Hartley, W. 1950. The global distribution of tribes of the Gramineae in relation to historical and environmental factors. Australian Jour. Agric. Research. L: 355-373. 49. Harvey, L. H. 1908. Floral succession in the prairie- grass formation of southeastern South Dakota. Bot. G&z. 46: 81-108, 277-298. 50. Hlesey, W. M. 1940. Environmental influence and trans plant experiments. Bot. Rev. 6: 181-203. 51. Hitchcock, A. S. and A. Chase. 1950. Manual of the gras B e g of the United States. U. 3.' TS. "a7” mTsc. Publ. 200. Second edition.

52. Hubbard, F. T. 1917. Andropogon scoparius in the United States and Canada. Rhodore. 15: 100-105. 53. Huxley, J. s. 1938. Clines: an auxiliary taxonomic principle. Nature 142: 219. 133.

54. Jones, C. H. 1944. Studies In Ohio floristics - III. Vegetation of Ohio prairies. Bull. Torrey Bot. Club. 71: 536-548. 55. . 1945. Vegetation of the Adams-Soloto Quadrangle. Dies. Ph.D. Ohio State University. (Unpublished). 56. Lane. G. H. 1941. Pollen analysis of interglacial peats of Iowa. Iowa Geol.Surv. Vol. 37, Ann Repts. 1934-1939: 233-262. 57. Larsen, E. C. 1947. Photoperlodic responses of geo graphical strains of Andropogon scoparius. Bot. Gaz. 109: 132-149. 58. Locke. John. 1838. Report on Adams County. Second Ann. Rept., Ohio Geol. Surv. 59. Law, A. G. and K. L. Anderson. 1940. The effect of selection and inbreeding on growth of big bluestem (Andropogon furcatus Muhl.). Jour. Amer. Soc, Agron. 32: 931-943.

60. Lawrence, D. B. et al. 1947. Data essential to com pleteness of reports on seed germination of native plants. Ecol, 28: 76-78. 61. Lawrence, W. E. 1945. Some ecotyplc relations of Deschampsia caespitosa. Araer. Jour. Bot. 32: 298-314. 62. Lewis, H. F. 1947. Leaf variation In Delphinium varlegatum. Bull Torrey Bot. Club. ?4: 57-59. 63. Livingston, R. B. 1952. Relict true prairie communi ties in central Colorado. Ecol. 33: 72-86. 64. McCllntock, E. and C. Epling. 1946. A revision of Teucrlum In the New World with observations on its variation, geographical distribution and history. Brittonia. 5: 491-510. 65. Morrison, T. M., Oliver Gossard and G. K. Sivasllan. 1918. Soil survey of Marion County, Ohio. U. S. D. A, Bureau of Soils. Advance Sheets. 66. Mueller, I. M. 1941. An experimental study of rhi zomes of certain prairie plants. Ecol. Mon. 11: 165-188.

67* _____ and J. E. Weaver. 1942. Relative drought resistance of seedlings of dominant prairie grasses. Ecol. 23: 387-398. 134.

66. Nielson, E. L. 1939. Grass studies. III. Additional somatic ohromosome comoliments. Amer. Jour. Bot. 26: 366-372. 69. Olmstead, C. E. 1944. Growth and development in range grasses. IV. Photoperiodic responses in twelve geographic strains of eide-oats grama. Bot. Gaz. 106: 46-74. 70. Ownbey, G. B. 1952. Adaptation and origin in the plant world, review of F. E. Clements, E. V. Mar tin, and F. L. Long. 1950. Adaptation and origin in the plant world. The role of environment"!!? evolu€ion. Chronica I^otanica,-Waltham, Rass. 71. Pelton, J, 1953. Ecological life cycle of seed plantB. Ecol. 34: 619-628. 72. Pilger, R. 1940. Gremineae XII (Panicoideae) in Die Naturllchen Pflanzenfamlllen von A. Engler und K. Prantl. Wilhelm Englemann,Leipzig. 73. Rice, E. L. 1950. Growth and floral development of five species of range grasses in Oklahoma. Bot. Gaz. Ill: 361-377. 74. Rogler, G. A. 1943. Response of geographical strains of grasses to low temperature. Jour. Amer. Soc. Agron. 35: 547-559. 75. Sampson, H. C. 1921. An ecological survey of the prairie vegetation of Illinois. 111. Nat. Hist. Surv. 13: 523-577.

76. Schaffner, J. H. 1926. Observations on the grass lands of the central United States. Ohio State Univ. Studies, Qontr. Bot. 178: 1-56. 77. Shantz, H. L. 1911. Natural vegetation as an indi cator of the capabilities of land for crop pro duction in the great plains area. U. S. D, A, Bur. PI. Ind. Bull. 201. 78. Shepard, H. R. 1938. Studies in breaking the rest period of grass plants by treatment with potassium thiocyanate and in stimulating growth with arti ficial light. Trans. Kansas Acad. Sci. 41: 139- 153.

79. Shlraek, B. 1911. The prairies. Bull. Lab. Nat. Hist., Sta^e Univ. Iowa 6: 169-240. 135.

80. Shimek. B. 1915. The plant geography of the Lake OkoboJI region. Bull. Lab. Nat. Hist., State Univ. Iowa. 6: 1-90. 81. ______. 1925a. The persistence of the prairie. Univ. Iowa Studies Nat. Hist. 11: 3-24. 82. . 1925b. The prairie flora of Manitoba. Univ. Iowa Studies Nat. Hist. 11: 25-36. 83. ______. 1948. The plant geography of Iowa. H. S. Conrad, editor. Univ. Iowa Studies Nat. Hist. 18: 1-178. 84. Smith, L. 1947. The acetocarmlne smear technic. Stain Tech. 22: 17-31. 85. Stebbins, 0. L. 1950. Variation and evolution In plants. Columbia University Press, New iforlcT 86. Taylor, A. E . , et al. 1938. Soil Survey of Adame County, Ohio. U. S. D. A. Series 1932, No. 29. 87. Terrell, E. E. 1952. Morphological variation in Galium boreale and its relation to the habitat. Thesis. Ph.D. University of Wisconsin (Unpub lished) . 88. Tharp, B. C. 1926. Structure of the Texas vegeta tion east of the 98th meridian. Univ. Texas Bull. ^606: 1-100. 89. Transeau, E. N. 1935. The prairie oeninsula. Ecol. 16: 423** 437, 90. Turesson, G-. 1922a. The species and variety as eco logical units. Heredltas. 3: 100*-113. 91. ______1922b. The genotypical response of the plant species to the habitat. Heredltas. 3: 211-350.

92. _____ 1925. The plant species in relation to habitat and climate. Contributions to the know ledge of genecological units. Heredltas. 6: 147- 236. 93. _____ 1927, Contributions to the genecology of glacial relicts. Heredltas. 9: 81-101. 94. Turrill, W. B. 1946. The ecotype concept, a consid eration with appreciation and criticism, especi ally or recent trends. New Phytologist. 45: 34-43. 136. 95. Weaver, J. E. 1919. The ecological relations of roots. Cam. Inst. Wash. Puhl. S86. 96. ______1920. Root development In the grassland formation. Cam. Inst. Wash. Publ. 292. 97. ______and Darland. 1949. Soil-Root relation- ships on certain native grasses in various soil types. Ecol. Mon. 19: 303-338, 98. ______and T. J. Fitzpatrick. 1932. Ecology and relative importance of the dominants of tall- grass prairie. Bot. Oaz. 93: 113-150. 99. and 1934. The prairie. ScoT.' 'Mon. 4rTb'5-S35.--- 100. Wiegand, K. M. 1935. A taxonomist's experience with hybrids in the wild. Science, N. S~. 81: 161-166. 101. Wolfe, J. N. 1942. Species isolation and a proglacial lake in southern Ohio. Ohio Jour. Sci. 42: 1-12. 102. Woodson, R. E. 1947. Some dynamics of leaf variation in Ascleplas tuberose. Ann. Missouri Bot. G-ard. 34: 353-432. 137.

AC KNOW LED GE :,'.ENT S

To all whose words and deeds are found here under the guise of the writer's In this p^-per, he herein makes grateful acknowledgement. Thanks are due particularly to Dr. J. N. Wolfe of the Department of Botany and Plant Pathology, the Ohio State University, whose helpful suggestions and criticisms, es pecially of the manuscript, have been thougit provoking. The writer has had correspondence with Dr. Frank Gould of Texas A. and M. College and Dr. E. Lucy Braun, formerly of the University of Cincinnati. He has consulted with Dr. E. S. Thomas of the Ohio State Museum concerning the location of prairie remnants and has used research maps of Professor Emeritus E. N. Tranaeau for the same purpose. Dr. Alvah Peterson has identified an insect lar-

vafe Inhabiting the growing stems of the Andropogon. To these and others with whom this problem has been discussed the writer is grateful. Acknowledgement Is made to Herbarium curators who have loaned specimens used in this study. They are: A.

L. Delisle, University of Notre Dame; R. A. Evers, Illinois State Natural History Division; Margaret Fulford, Univer

sity of Cincinnati; F. C. Gates, Kansas State College;

F. D. Grover, Oberlin College; C. B. Heiser, Indiana University; W. H. Horr, University of Kansas; G. N. Jones, 138. University of Illinois; W. H. Welch, DePauw University. Specimens were also received from E. L. Core, University of West Virginia. They were examined but the data are not presented here. My wife Is due a special debt of gratitude for help ful field companionship and gracious assistance in assem bling data. In addition she has drawn most of the figures and has typed the first two drafts of the manuscript^ 139.

AUTOBIOGRAPHY

B o m at Columbus, Ohio, November 1, 1924, I was given the name Henry Rawie DeSelm. Formal public school education was begun in 1929 at Crestvlew Elementary and completed in 1942 at North High

School in Columbus. College training began in 1942 at Ohio State Univer sity where I entered the Plant Science Curiculum. My soohmore year was spent at Miami University, Oxford, Ohio, as a member of the U. S. Marine Corps V-12 Unit. Marine Corps training took over another year follow ing which I want aboard a Navy Transport, the U.S.S. San Saba as transportation officer in charge of passenger per sonnel and equipment. Release to inactive duty occurred in August, 1946, and two months later my formal education was resumed at Ohio State. I received in 1948 a B.Sc. in Agriculture, a B.Sc. in Education in 1949, and M.Sc. in Botany in 1950. June 11, 1948, I married Mary Elizabeth Hersee. Between the autumn of 1950 and the summer of 1953 I held the position of Graduate Assistant in Botany. The autumn of 1953 I was appointed temporary Assistant Instruc tor, a position which I now hold. Work toward the Ph.D. degree was begun in 1950 and has continued until the pre sent.