ANNALS OF THE MISSOURI BOTANICAL GARDEN ERRATA May cover, last line of Contents—186 should read 184. Page 65, line 29—var. nov. should read Sp. nov. Page 113, line 34, and on following pages of Gage's paper—chinensis should read chinense. Annals of the Missouri Botanical Garden Volume XL 1933 With 30 plates, and 41 figures Published quarterly at Galesburg, Illinois, by the Board of Trustees of the Missouri Botanical Garden, St. Louis, Mo. Entered as second-class matter at the post-office at Galesburg, Illinois, the Act of March 3, 1879. Annals of the Missouri Botanical Garden A Quarterly Journal containing Scientific Contributions from the Missouri Botanical Garden and the Henry Shaw School of Botany of Washington University in affiliation with the Missouri Botanical Garden. Information The ANNALS OF THE Missourl BOTANICAL GARDEN appears four times during the calendar year: February, May, September, and November. Four numbers constitute a volume. Subscription Price |... $10.00 per volume Single Numbers _____ 2.50 eac Contents of previous issues of the ANNALS OF THE Missouni BOTANICAL GARDEN are listed in the Agricultural Index, published by the H. W. Wilson Company. COUR! BOTAN 5 УЕ ne Rig t аак we MISSOURI BOTANICAL GARDEN STAFF Emeritus Director GEORGE T. Moore Acting Director Jonn S. LEHMANN Assistant Director EDGAR ANDERSON Executive Assistant LEICESTER В. Faust CARROLL W. DODGE, ROBERT J. GILL Mycologist Research К ROBERT E. WOODSON, GEORGE B. VAN SCHAACK, Curator of the Herbariu Honorary Curator of Grasses HENRY N. ANDREWS, JULIAN A. STE K, М н ist тенби ан d уле lite Rorra M. Tryon, Joun D. Dwyer, Assistent Curator of the Research Associate CUTLE ELL C. Hor Curator Museum of ае d | Editor Economic Plants of Publications GERALD ULRICI Business Manager BOARD OF TRUSTEES President JoHN S. LEHMANN Vice-President DANIEL К. CATLIN Second Vice-President EUGENE PETTUS DUDLEY FRENCH GEORGE Т. MOORE HENRY HITCHCOCK A. WESSEL SHAPLEIGH RICHARD J. Lock woop ETHAN A. H. SHEPLEY ROBERT BROOKINGS SMITH EX-OFFICIO MEMBERS ARTHUR Н. Сом RAYMOND К. Tuck Chancellor о of Washingt 1 Mayor of. d e City wy ARTHUR C. LICHTENBERGER, STRATFORD L Bishop ¢ ыы the Diocese of President of the Academy of Mis Science of St. Oscar EHRHARD President. of the Board of dn: of St. Louis GERALD ULRICI, Secretary ` TABLE OF CONTENTS Bluegrass Pasture Almanac. Alfred G. Etter The Popcorns of Turkey аваг Anderson and William L. Brown A Preliminary Survey of.the Milpa System of Maize Culture as Practiced by the Maya Indians of the Northern part of the Yucatan Peninsula. R. A. Emerson. А Contribution to the Lichen Flora of Arizona and New exico Emanuel D. Rudolph The Analysis of Suspected Hybrids, as Illustrated by Berberis X gladwynensis gar Anderson Variation in Cob Morphology among о. Archaeo- logical and Ethnological Races of M kontod H. Nickerson The Cytology, Morphology, and Systematic Relation- ships of Delphinium X Belladonna Hort. ex Bergm. Marilyn Amy Gage Some American Medullosas Henry N. Andrews and Sergius H. Mamay The Distribution of Diospyros R i с. arry R. e Wildwood—A Study in Historical Ecology |... Alfred G. Etter Some X-Ray Effects in Petunia. Henry A. McQuade Some Lichens of Tropical Africa... Carroll W. Dodge General Index to Volume XL PAGE 17-729 33— 48 51— 62 ег 73.78 795711 113—182 183—209 411-223 227-258 259—270 271—401 403—412 Volume XL Number l Annals e d. up rt Missouri Botanical | m n Annals | 4 о cof the о: врея Garden ar and the Henry Shaw School of dont of ity i in affiliation with. the Missouri Botanical Annals of the Missouri Botanical Garden Vol. 40 FEBRUARY, 1953 No. 1 BLUEGRASS PASTURE ALMANAC* ALFRED GORDON ETTER Malvern B. Clopton Experimental Farm, Clarksville, Mo. Washington University School of Medicine JANUARY The long summer leaves of bluegrass have turned brown, making the sere meadows and pastures of winter. In a few places, however, where these long leaves have been grazed off, pastures are still rather green. Bluegrass is also surprisingly green under certain trees and on the north slopes of hills. Isolated plants and small clones beneath apple trees in the orchard are brilliant green and are used by rabbits in preference to the tangle of brown plants in adjacent sod in full sun. Shading produces plants with leaves proportionately high in protein and low in fiber, which seem to satisfy the cottontails. Shading also restricts rhizome development suffi- ciently so that plants do not become crowded and thus sod-bound and nitrogen- deficient. After tree leaves are shed in fall, plants that were shaded all summer stay green well into the winter. When blades of these plants do die, they wither into very thin and delicate tufts, low in fiber, that contrast with the tough per- sistent dead leaves of sun-grown plants. Hogs on the hill pasture are working up moist areas into black patches of open soil. Where a bed of limestone outcrops above a bed of shale there are seeps which keep the ground soft and wet. Hogs have made the contact of these two forma- tions clearly visible across the hill (fig. 1). Their trails across the orchard-grass * The observations recorded in this study were made during bug! — s three-year residence on Brookhill не.) near Clarksville, Missouri. This farm and des pecial research fellowship were given ngton Univers ity н thé late Dr. Malve : S CH га? tes interest o Washin was in the касса of living things to the land. Grateful est is due Dr. Edgar Anderson of i Henry a ac y of dr of Washington University, and Dr. Arthur H. Compton, S Chan тА arranging this unusu niei Se = -— author. Much credit is also due the nis e "Aldo Leopold of the уник of Wisc hose inspiration err a background for these studies. nerous grant from the О. M. ro са Sons Company of Marysville, Ohio, =~ — it e As the author to organize his field notes and prepare the almanac for publica- e present pes is a companion pd to the author's "How Kentucky Bluegrass Grows, whic cb е eee in the September issue of the ANNALS for 1951 (Vol. 38, pp. 293-375). References to the — and орай 2 bits of toc tae will be made much more clear by referring to some of the diagrams in that p (1) [Уог. 40 2 ANNALS OF THE MISSOURI BOTANICAL GARDEN pasture are also very apparent now. Їп a part of this pasture which was left un- mowed, the orchard-grass plants became rank and have been avoided by the hogs, not only in their feeding but also as they walk about the pasture. Consequently the pasture there has deteriorated into a rough tussocked run (fig. 2). During winter when the ground remains frozen for long periods, bluegrass leaves become thin, wiry and twisted, with tips curving inward, probably as a result of the difficulty of obtaining moisture from the frozen soil. January leaves are very short, only an inch or so long. In areas closely grazed since fall, stock are hard-put to it to find enough to eat. Horses are able to eat close enough, but hay is being fed cattle. The grazing habits of stock are distinctive and play an important role in determining the appearance of a pasture throughout the year. An important fact is that cattle and horses prefer grass which has grown in soil fed by the nutrients contained in their urine. They instinctively conserve the nitrogen, potassium, phosphorous, and trace minerals that urine contains and benefit from the stimula- tion which these substances, as well as dissolved hormones and vitamins, give the soil microorganisms and the herbage itself. This they do by actually seeking out the deep green grass which grows where urine has fallen (figs. 7 and 11). Analyses made by the author of grass in these spots has shown that it contains twice as much vitamin А (carotene) as ordinary grass, and Neven (1941) reports that the protein content of such fertilized grass may be nearly doubled. Horses graze small areas intensively, but cattle range widely. Urine is dis- tributed as the animals graze so that characteristic patterns of urine-fertilized grass develop. The pattern produced by cattle seems to be a rather random distribution over the pasture as a whole. With horses, however, urinations occur in small areas where the grass is kept closely grazed (fig. 7). The urine characteristically falls either on this short grass or, when grazing pressure increases, along the edges of these grazings (fig. 12) in which case it has the effect of continually enlarging the intensively grazed area (fig. 13). Of equal significance is the distribution of manure. Stock avoid grass around their own manure clumps as instinctively as they seek out that fertilized by their urine. Cattle manure is scattered rather indiscriminately though there is some considerable concentration in loafing and bedding areas. Horse droppings, how- ever, are somewhat controlled and deliberately concentrated in certain spots around the pasture. Often these spots are low places or fence corners; almost invariably they are weedy, full of rank vegetation. By keeping the manure thus concentrated and out of the way, their system of intensive grazing on small urine-fertilized areas is not interfered with. Since in summer, stock avoid grass around clumps of their own manure, "rough" areas of rank grass are protected from consumption (figs. 9 and 13). Considerable forage is thus saved for winter use. When the manure is frozen it no longer seems to be objectionable. In January, when new grass growth is at a 1953] ETTER—BLUEGRASS PASTURE ALMANAC 3 minimum, these areas are brought into use. Clumps of manure previously pro- tected by grass are now exposed, and to some extent are being scattered about the pasture. These "roughs" provide a convenient emergency supply of fodder when freezing rains and sleets fall. Urine grazings become a solid sheet of ice, but roughs, drifted with a little sleet, are still rather soft underfoot, and horses can paw them with their hooves, scrape away the crust of snow, and obtain all the grass they need (fig. 3). Cattle are not quite so adaptable and must be fed. No matter how heavy summer grazing may be, reserves of long grass around manure will usually be saved for overwintering. If fall months are very dry, however, grass growth will be slow throughout the pasture and “rough” reserves will be invaded during October and November so that before winter is over roughs may almost disappear. The past summer certain small blocks of pasture were enclosed with fence so that stock could not reach the grass within them (figs. 1 and 10). This protected grass has grown rank, and I can be almost certain that in each “‘exclosure” a rabbit will be spending the day. In fall and winter rabbits like the rich proteinaceous grass nourished by stock urine. By using these protected islands in the pasture they have access to much of that fertile grass, and their pellets are often found in closely grazed urine grass. The rabbit’s midwinter desire for proteinaceous food, which apparently ac- companies the advent of the breeding season, is satisfied by grazing intensively small plots of grass which have been fed urine by dogs, stock, man, or by the rabbits themselves. These grazing places are eaten down into the dirt with nothing but bare soil remaining (fig. 4). Such rabbit pastures are common sights along the runways of rabbits in winter orchards. They can also be seen occasionally in the summertime (fig. 20). In some places mice grazings are just as conspicuous. In October some urine was poured in a circle around a stake placed in an old Canada bluegrass meadow. Now the grass which received urine has been completely consumed and the ground is bare in a circle around the post (fig. 5). Even the rhizomes have been eaten. More striking still, where some of the urine had fallen on the post, the wood was chewed up by the mice. Gophers are no longer busy bringing up their piles of dirt as they were during fall (fig. 6). The extensive bulldozing they did then is slumping down with the frost and being scattered by treading hooves and being undermined by rhizomes of bluegrass. Most fall gopher mounds are built after temperatures are low enough to keep weed seeds from germinating. Also in fall and winter rhizoming and tillering of bluegrass are very active, and the new loosened aerated soil in the mounds encourages the vigorous vegetative development of bluegrass plants around its edges. The invasion and conquest of the mound go on to some extent all winter, for the loose soil does not freeze deeply. Revegetation of these mounds is [Vor. 40 4 ANNALS OF THE MISSOURI BOTANICAL GARDEN also accomplished by plants which have been buried beneath the soil. When leaves and crowns are covered, internodes elongate and bring the crown up to the surface of the mound. Branching of rhizomes, while infrequent in most places, is the order of the day in gopher mounds. FEBRUARY Wherever the pasture thaws during the winter, stock cut the sod with their hooves and create large scars by sliding and slipping. These scars usually heal rapidly if made in early winter, but may remain for a considerable period if made in late spring. When the snow is soft, horses moving about build up pads of snow on their hooves. When these pads later slip off, many weed seeds and even vege- tative parts are carried about and planted. From a single pad picked up in Feb- ruary, 28 plants were germinated. Sixteen species were represented, half of which were grasses (Etter, 1948). During these cold months the rhizomes of bluegrass which began to grow in late October and November behave much as do the bulbs of such early spring plants as jonquils. Being situated near the surface of the soil, however, they are more subject to variations in the weather. They grow slowly in late fall, fitfully with thaws in the winter, and rapidly in the spring, shooting up and developing into new plants. At any time in winter one may find, in places where growth is rather vigorous and fall grazing was not too intensive, small rhizomes which are all white. Those which did considerable growing in fall have slightly green tips, and others may show a few short leaves. The dominant process, of course, during winter in every bluegrass plant is tillering. In old meadows this proceeds at a snail’s pace, but along streams, in loose, rich soil, where plants have room to stretch and root, it proceeds rapidly. More subtle is the secret development of the bluegrass seed head within enclosing sheaths. This process may have begun as early as November or may wait as late as March, depending on variety and environment. Winter thaws are important to bluegrass. With only a little rise in the soil temperature above freezing, fat white roots develop from the new tillers which have originated during the previous months. These roots, in turn, appear to stimulate new buds to develop into additional tillers. Where the soil is protected by an overburden of long leaves, thawing is slower and less frequent and roots do not develop very rapidly. Well-grazed or isolated plants, however, begin to send out roots a few hours after the soil thaws. This may play an important part in the earliness and the more vigorous growth of open-grown or spaced plants often observed along paths, in alluvium, in gardens, or near buildings. It is certainly true that bluegrass is often early to flower on the south sides of buildings and hills, above subterranean steam pipes, under eaves that drip warm water frequently in winter, or in places that are protected from snowdrifts as on the south sides of evergreens. 1953] ETTER— BLUEGRASS PASTURE ALMANAC 5 In meadows, now, there are well-camouflaged tunnels through which mice circulate in contempt of the winter weather. These are not mere pathways to and from, but feed alleys along which mice find their food. The stems of grass are cut near the ground to make the tunnel. When new leaves grow from the crowns of these severed plants, they stick up along the pathway. This growth is tender, bleached, and proteinaceous and is nipped off by the mouse as he navigates his pasture. These tunnels occur in heavy meadow grass where the overburden of leaves keeps the soil from freezing deeply, so that a small amount of blade growth takes place throughout the winter. Small feeding pockets extend short distances from the tunnels and are frequently filled with the grass-green pellets of mice. Undoubtedly the urine is also there. The floor of the tunnels and pockets is littered with clipped stems and pellets and is penetrated by a maze of bluegrass roots. When the ground thaws, the pellets disintegrate and a rich bed of mold develops, fed by these excretory products. The reserves of the rhizomes and crowns which feed the grazed plants are depleted by winter-long grazing and lack of sun, and tend, in spring, to produce weak growth which is often unable to reach through the overburden to the light. When plants such as chicory and plantain, which have bulky storage crowns, grow isolated in rank meadows they are apt to be consumed over winter by mice or gophers. Two plants, one of plantain, one of chicory, which I had marked last summer have been cut down and their roots consumed and transformed into a pile of pellets. MARCH A little warm weather in late February and March is soon followed by a slight greening-up of pastures. Earthworms return to the topsoil and robins arrive in time to take advantage of them. Flocks of starlings and purple grackles are often found feeding on these early spring pastures. During December and January bluegrass leaves tend to become short and prostrate, and leaves grow only about a half centimeter a week. When the days become longer and thaws are frequent, leaves and plants straighten up. New leaves shoot up from tillers and rhizomes which have developed during the winter. Since the snow, ice, wind, and treading cattle have broken the old tufts of grass in much of the pasture, these new leaves are not encased in the long sheaths of the previous year, and so become visible almost as soon as growth begins. Where leaves reach the light with a minimum expenditure of food reserves, they are in- clined to be more vigorous, broader, deeper green, and more attractive to livestock than those which have had to struggle up to find light. During this greening-up period, little further tillering occurs, and this is one time of the year when virtually no rhizomes are initiated. Any rhizomes still underground turn up and proceed to develop leaves. Often after this green-up first begins, the weather becomes cold again and leaf blades become reddish at the tips and growth almost stops. This redness, apparently anthocyanin accumulation, is especially common in the [Vor. 40 6 ANNALS OF THE MISSOURI BOTANICAL GARDEN tips of grass recently grazed or cut. It seems to be a characteristic response of plants which are all ready to begin active growth and are prevented from doing so, either by injury, drought, low temperature, or excessively bright sun. During this spring renaissance, reserves in the crown, roots, and rhizomes are used up to produce new leaf growth. Sods become remarkably weak and may be pulled apart more easily than at any other time. | Wherever urine has fallen on the soil since last September, bluegrass shows ап early difference. Grass fed nitrogen does much more growing at low temperatures, below 40°, than does unfertilized grass (Blackman, 1936). At higher tempera- tures the nitrogen seems to have relatively less effect. Urine-fertilized grass is deeper green, thicker, and broader-leaved. Growth is also early in other sites where the ground is rich or moist, such as bare places once occupied by haystacks, near manure piles, in seeps, and around the edges of spots of sod burned out by urine last summer. Now that the sod is becoming soft, hogs are taken off the permanent pastures. Where they have been fed is a litter of corn cobs and mud which will become a rank bed of rich weeds in a few months. Horse grazings are now strikingly differentiated from the rest of the pasture which still is rather bleached (fig. 7). Tufts of grass fed by steer urine during fall and winter are now becoming visible, and cattle run in these early pastures seek out these patches. It is interesting that horse urine does not burn bluegrass when it falls on frozen ground. By mid-March, however, severe burning may occur. Cattle urine seldom burns grass even in mid-summer. APRIL In early April dormant buds on the bluegrass crown begin to swell, anticipating the spring flush of rhizomes that accompanies flowering. At the same time, blue- grass florets and panicles are growing, getting ready for the grand elongation of the stem which will occur in a few weeks. Long green blades of bluegrass are waving in the wind. Where pastures were not grazed last fall the blades are long, three or four inches, while in grazed places they are only an inch or so long and form small rosettes. There are many more blades in the grazed area, however. In rich soil, closely grazed, there may be twenty or more blades to a square inch. Trails and paths made through old meadows last summer and fall show up conspicuously now as stripes of green grass in a tangle of old whitish leaves. Break- ing down the sheaths and shading growth gave short fall and winter leaves access to winter sun. This stimulated tillering. A rough count of leaves along such a pathway showed twice as many new spring leaves as in undisturbed parts of the field. This path was used only one day, when it was walked over only five or six times. It demonstrates very simply how fall discing or mowing and raking of neglected pastures will do much to improve them. Down along the creek the succession of plants on newly built alluvial banks _ leads eventually, under grazing, to bluegrass. After gravel bars are stabilized by ee a eer мыд) 7999 1953] ETTER—BLUEGRASS PASTURE ALMANAC 7 sweet clover (Melilotus officinalis L. (Lam.), bouncing-Bet (Saponaria officinalis L.) and willow sprouts, coatings of alluvium begin to support a variety of weeds. Composites such as species of bur marigold (Bidens sp.), cocklebur (Xantbium sp.), sunflower (Helianthus sp.), as well as great ragweed (Ambrosia trifida L.), will also grow here, but will not withstand grazing and trampling. Rhizomes of Muhlenbergia racemosa (Michx.) B.S.P., broken loose from clones upstream, lodge on banks and begin to establish a rank stand of coarse grass which ties down the soil and speeds up deposition. As time goes on, bluegrass becomes established and invades and eliminates the other species. Without grazing, the succession would proceed from the coarse grass and weeds to shrub and tree growth. Grazed alluvial banks support excellent stands of bluegrass in the older established areas. Other grasses are developing rapidly now. Orchard grass (Dactylis glomerata L.) is flowering, and quack grass (Agropyron repens (L.) Beauv.) and smooth brome (Bromus inermis Leyss.) are beginning to develop rhizomes, while Canada bluegrass (Poa compressa L.) sods, devoid of new leaves, show many long white rhizomes which have not yet turned up. Redtop (Agrostis alba L.) is still almost dormant, a few green leaves beginning to appear. Purpletop (Triodia flava L.) has made virtually no growth, although its white dead stems are prominent in parts of certain pastures (fig. 8). On south slopes dandelions are just appearing, and in thin places a variety of small seedlings now appear. Old chicory (Cichorium Intybus L.) plants show only a few small leaves. Earthworms are abundant and are throwing up extensive castings within the secrecy of the grass. This subtle cultivation, accompanied by an increase in the numbers and activity of nitrogen-fixing bacteria in the soil, comes at the time when bluegrass needs it for rapid nitrogen-consuming leaf growth and inflorescence development. Moles are busy now too, tunneling in well-drained places, but gophers seem strangely absent. Cottontail rabbits are using wads of last summer’s long grass to hide their nests in. Young cottontails will be venturing from these nests very soon. By mid-April the blades on the flowering shoots of bluegrass have finished their development, but they are still telescoped within each other, waiting for the elongation of sheaths and internodes. The panicle or flowering head can be found just within the sheath of the newest leaf. It is striking to note that shoots of bluegrass which are going to bloom this spring can be easily selected now, because they have shorter and wider leaves than the blind or non-flowering shoots, and they are conspicuously a deeper bluer green. The short winter leaves are beginning to turn yellow, and this dulls the green- ness of the pasture very slightly. When stock graze the pastures now they are getting not only leaves but the developing flowers within. Mowing or close grazing at this time will greatly reduce the flowering of pastures. It will also produce a silage or feed as rich in protein as bluegrass can produce. It may, how- ever, be a little short on sugar. By waiting a few weeks for the panicles to develop [Vor. 40 8 ANNALS OF THE MISSOURI BOTANICAL GARDEN a little further a heavier yield will be forthcoming and the carbohydrate content will increase, though the per cent protein will drop accordingly. Old gopher mounds are now drying out and are very evident in pastures grazed continuously, for the grass there has not been given much chance to cover them (fig. 6). Bluegrass plants which grow on these mounds are very distinctive; the leaves are short and wide, deep blue-green, are prostrate, and often have much anthocyanin in their blades. Flowering shoots are consistently early to appear and grow almost horizontally. Winter pigs are on pasture now and they thrive on the high protein of the new grass. They are also busy rooting for worms and other foods in the loose soil under walnut trees. In a few weeks they will be seeking clover. Where hogs were pastured during the cold months, what grass remains is now a brilliant green. Hog urine has an especially strong, though brief, effect on bluegrass, and that combined with the well-distributed manure and treading stimulates the grass as it is stimu- lated in few other places. The excretory behaviour of swine is quite as interesting and significant as is that of cattle and horses. Hogs, when penned, are very careful to drop their manure only in certain corners. When on pasture they are careful to keep both urine and manure from contaminating the bedding area. Consequently, both waste products accumulate just outside the sleeping area, where sleepy hogs politely retire. Swine are just as careful not to defile their rooting areas, although here the prejudice is limited to the solid droppings. They do not hesitate to urinate in their excavations or in their wallows. Since hogs have need for various trace elements and antibiotics, it is quite possible that wallowing may be more than merely a way of keeping cool. The continual dabbling and rooting and munching that goes on while pigs search their wallows may provide them in a natural way with nitrogen, animal protein factors, and trace elements. Almost invariably, after rooting and wallowing, hogs leave along some much- used path, travel a few hundred feet, drop dung just off the path, and proceed with the day’s affairs. Where pastures are intensively grazed and where corn supplements, minerals, and water are provided all close together, these patterns are broken down and feeding areas are littered with manure. Frequent worming then becomes a necessity. In a pasture grazed last fall for a few months by steers at a rate of about two head per acre, there now appear as many as 20 urine spots in a plot 45 feet square, or about 1 square foot of urine grass to 40 unfertilized. In a pasture grazed longer and more intensively the spots are so close that I can easily step from one to the other. An important result of the stimulation of these pastures by urine is that bulky and less nutritious plants which might ordinarily be avoided are con- sumed to provide the necessary bulk. It is a common observation that the more rich proteinaceous food an animal has the more bulk it can handle. 1953] ETTER— BLUEGRASS PASTURE ALMANAC 9 It is rather easy to tell approximately when the urine fell which stimulated a certain piece of sod, for the vegetative response to fertilization depends on the season.when it is applied. Urine applied in fall promotes vigorous tillering which leads to short leaves and many short inflorescences with heavy loads of flowers and a decided blue-green color. Spring applications produce leaves as long as the in- florescences. ‘These shoots are weak, short, bear few well-developed flowers, and often lodge. With such information it is possible by examining urine spots to tell approximately when and how long an area was grazed. It is interesting that pig and calf urines have no such effect as do hog, beef, human, or mare urines. The difference in the reaction of the grass may be due partly to the characteristically lower nitrogen content of urine from a growing animal, or to a different hormonal content. The effect of urine on the grass of pastures is not limited to a single season, although the striking flush of growth in spring is the response most commonly seen. ere is a more lasting influence as a consequence of the fact that urine stimulates the development of grasses more than it does of weeds, especially when applied in winter. This rapid growth smothers out weed seedlings, clover plants, and results in a pure stand of bluegrass. Urine holds down weed development in a second way. From controlled tests made in flats kept in a greenhouse and planted to a variety of pasture weed seeds and grasses, it was found that germination of weed seeds was strongly inhibited in flats treated with urine, while grass seeds suf- fered little if any such inhibition. This effect was just as marked when urine was applied to plots of bluegrass in the pasture or on a lawn (fig. 14). In pastures grazed last summer by the horses there is much clover. Horse pastures characteristically contain more clover than those grazed by cattle. This is due in part to four facts: (1) Close grazing encourages clover, and horses graze quite close. (2) Compaction of the soil in certain areas maintains a favorable moisture condition for germination. (3) Breaking of the sod during winter in- duces germination of clover seeds, and horses cut the sod more than cattle. In a simple chopping test where a rank sod of pure bluegrass was cut with numerous strokes of a hand axe while the ground was frozen, clover germinated the following spring to such an extent that boundaries of the plot could easily be discerned. Winter discing also favors clover. (4) Most interesting relationship of all, how- ever, is that involving urine and grass. Horse urine that falls on growing grass usually kills it; cow urine does not. Burned patches soon become bare ground. While weather and soil conditions influence the flora that develops on these spots it is especially characteristic that spots which develop in dry mid-summer or fall remain bare until the following spring. At that time clover seedlings take over the bare ground and by April the patches are solid in clover. The attraction which urine-saturated soil has for earthworms and many other burrowing organisms may lead to extensive aeration of these spots and so encourage clover germination also. By the last week in April dandelions and peppergrass (Lepidium virginicum L.) are beginning to elongate, having already produced nutritious rosettes. Not only [Vor. 40 10 ANNALS OF THE MISSOURI BOTANICAL GARDEN has bluegrass begun active production of rhizomes, but the panicles of early- flowering bluegrass plants have already been exserted. Silica deposition in the soft culms is going on and stock are avoiding them now. Where grazing is close and intensive the tough stem interferes with the efficiency of the grazing procedure. The stem does not tear when the cow pulls on the leaves, but instead the entire plant comes up by the roots. Dead plants pulled up in this manner litter the turf in places. At this season when growth is more rapid than at any other, a surplus of herbage soon develops and stock begin to pick and choose their preferences. А striking example of the attraction which urine-fed grass has for cattle now is shown in fig. 11. May During the first part of May flowering stems of bluegrass elongate rapidly, at least a centimeter a day. Ungrazed pastures become a sea of grass. Inflorescence elongation seems to have a depressing effect on the non-flowering tillers of the same plant, for many of them die during this period. They give the appearance of having been deprived of water. It is obvious that death of some plants and shoots must occur in a stand of bluegrass each year in order to maintain an even density over the years. In ungrazed meadow grass the greatest number of these deaths occurs at this time of year. There is also a gradual dying of the leaves on the inflorescence itself, all but the upper one or two having turned yellow at the tip. Beneath the soil new rhizomes are more abundant than ever. Some will turn up in the next few weeks; others may turn up sporadically throughout summer. All will turn up by early fall. It is a striking fact that new roots virtually cease to develop after April, though elongation of existing roots continues. Rank growth in the unknown meadows of last year, especially the prominent white stems of last summer’s purpletop (fig. 8), begins to blacken with mold, weaken and collapse. Lespedeza seedlings are coming up in patches of last year’s crabgrass, and both orchard grass and wild barley are showing their flowering heads. Cows drop their manure indiscriminately on the grazing places of horses, ignoring the careful system which the horses would maintain. Horses will avoid the grass around these cow-manure spots for a short time but in general the taboo against using grass around manure applies only within the species. In Germany, it was often the practice to apportion horses and cattle to pasture in such a ratio that rough areas were kept eaten clean, the horses cleaning up cow-dung patches, the cows the horse-dung grass. Cows are even now using some of the grass around piles of horse manure. Where horses graze with sheep, roughs are more evident (fig. 13); both species prefer the short grass, the sheep avoiding the grass around horse droppings almost as religiously as the horses themselves. Sheep drop most of their manure in loafing areas. What manure is dropped on the pasture is so easily scattered about that it seems to have very little residual effect. Grass fertilized 1953] ETTER—BLUEGRASS PASTURE ALMANAC 11 with sheep urine, however, has the same attraction as that fertilized by other stock. During the last part of April when bluegrass leaves are soft and green, aphids are active, sucking juices from them. Very soon, however, lady-bug larvae begin reducing the aphid population, searching every blade its entire length for their green prey. In two weeks more these larvae will fatten and develop into full- fledged lady-bugs. Now, in the last part of May, white clover (Trifolium repens L.) is flowering, and the inflorescences of Canada bluegrass are leaving the protection of surround- ing leaves. Cheat (Bromus sp.) is ripening, yarrow (Achillia Millefolium L.) has begun to bloom, ironweed (Veronia sp.) is 18 inches high, and the first blades of purpletop are beginning to show. Kentucky bluegrass has mostly flowered and seeds are ripening. Steers have shifted from bluegrass to rosettes of redtop and are grazing them intensively. Since redtop blooms later than bluegrass it is still in an early stage of development, growing rapidly, and consequently more palatable and proportionately higher in protein and lower in fiber than the more mature plants of bluegrass. This is a demonstration of the general tendency of stock to seek during each part of the summer those plants which are immature. After redtop begins to flower, the leaves of pigeon-grass, purpletop and paspalum will be sought, and new lespedeza will be consumed eagerly. By the time these late-developing species have matured, cool-weather species such as bluegrass will have begun to develop palatable new fall growt It is interesting to see young rabbits cutting the tender flowering shoots of bluegrass that grow on rich creek-bank soil, then eating them, flowers and all. Occasional they will prefer a dandelion stalk, which is consumed, stem, puff, and all. Dung beetles are actively digging their tunnels into the soil, then excavating fresh manure which they carry down into their burrows. The soil brought up is spread out over the top of the cow dung, serving to keep the dung from drying out, inoculating it with a variety of soil organisms, and hastening the course of disintegration and the incorporation of this organic matter into the soil. Dung beetles play a significant part in maintaining the fertility of a heavily used pasture for they keep the cycle of fertility operating efficiently with little waste of time or material. Where pastures are lightly grazed, or merely mowed occasionally, there is not enough manure to maintain a dung beetle population, and manure that falls often remains undisturbed and undecayed for years. Under those conditions little if any fertility is restored to the soil by way of animal waste products. It is also significant that stock on rank pastures, where there is little clover or rich green grass, produce manure which is bulky, hard, and low in nitrogen. This has little attraction for dung beetles hunting a rich medium for their eggs to develop in. Manure of stock on rich pasture is loose and moist, and decays readily. Machines which have been developed to cultivate intact sod by spiking or plugging the soil are following the precedent of the dung beetles, especially those devices which bring up small cores of soil and distribute them over the surface of [Vor. 40 12 ANNALS OF THE MISSOURI BOTANICAL GARDEN the grass. Much of the benefit of this machine comes from the loose soil which it adds to the surface of the ground and mixes with the litter of grass stems, en- couraging their decay. In a small way, it simulates the natural process of flooding, whereby thin additions of alluvium from flooding streams are made to pastures on low ground. Plots of grass experimentally burned in February are now decidedly more yellow-green than adjacent grass not burned. Inflorescences are less green and less full. Since the ground was moist and frozen when the plot was burned and the fire not at all hot, there was little if any consumption of decayed nitrogen-rich litter from the soil surface. The only loss was the overburden of summer leaves. The evidence of nitrogen deficiency in the plants and flowering shoots of the burned plot suggests that the tangle of leaves makes an important contribution to the nutrient economy of an ungrazed meadow. A bluegrass pasture that is maintained by light grazing or mowing reaches a stability that is the result of a complex interrelationship of many factors. If we ook down at it closely, it is a veritable jungle with many kinds of insects walking in and out among the fallen stems. There are worm castings, mouse droppings, white molds, and minute fungi. A few attenuated weed seedlings try to reach the light. Bluegrass roots come up during the summer into this community and feed, and they are as dependent on a balance of these organisms as they are on rain. As a matter of fact, in this litter moisture is retained, and the molds and fungi, thus encouraged to develop, release nutrients and build up important organic com- pounds. Prolonged mowing experiments which remove leaves to be dried and weighed for assays discourage this natural process and inevitably interfere with production. Many such tests, repeated year after year, report unexplainable de- clining yields as years go by. There is little doubt that destruction of the reigning balance of nutrient release inhibits forage production. Summer roots have no place to feed. Disruption of the balance in a meadow sod by burning or continuous removal of clippings with consequent weakening of the cover encourages perennial summer grasses such as purpletop or broom-sedge which demand less nitrogen. With re- peated burning it has been shown (Curtis & Partch, 1948) that fire can weaken bluegrass enough to allow prairie grasses and forbs to become established. This weakening involves, in all probability, the discouragement of tillering. JUNE In early June bluegrass inflorescences are still mostly rather greenish-yellow, though some patches tend to be more purple, or red. There is also much difference in the heights and leafiness of the flowering shoots. At first glance one might believe that the different patches are clones, displaying inherent differences but such is not always the case. Bluegrass is a very flexible plant, and it is subject to an unbelievable variety of influences and environments. Urine spots alone cause the development of many false clones, either by stimulating certain patches, or by 1953] ETTER—BLUEGRASS PASTURE ALMANAC 13 encouraging grazing in certain places. The differences between fertilized grass or grazed grass, and adjacent grass which has not been treated is often so striking as to seem uninterpretable except on a genetic basis. Time and degree of fertilizing influence the length and number of inflorescences and leaves. Close fall and winter grazing will produce plants with numerous small inflorescences and short leaves. Lack of fall grazing results in fewer and much longer flowering stems and longer leaves. Spots entirely avoided by stock for one reason or another will con- trast with those which are grazed. Treading also has its effect. Where pigs pastured intensively in winter, blue- grass is hardly flowering; the few inflorescences present are much shortened. The combined influence of actual killing of crowns by hooves, and the weakening effect of continual injury and nitrogenous fertilization played a part in preventing blooming. It was found, however, that hoeing of all top growth of grass from the frozen surface of the soil during February did not prevent crowns beneath the surface from sending up flowering shoots in spring, but it did greatly shorten the inflorescence, and weakened recovery growth. This weakness persisted throughout the summer, and the area was invaded by many plants, such as horseweed (Erigeron canadensis L.) , milk purslane (Euphorbia maculata Г.) , Paspalum pubescens Muhl., boneset (Eupatorium sp.), purpletop, and white clover. Panicles on this plot ripened several weeks early, were abnormally white, and are already dropping seeds. During this flowering season inflorescences are often found which have barely appeared out of retaining leaves and then died. The heads are а gray-white color, and the condition is often called silver-top. It is caused, in many cases, by the larvae of a thrip which consumes the soft tissue of the developing stem at one of the upper joints. This severs the panicle from its source of water and food and prevents its further development. While this thrip sometimes causes trouble for seed-raisers, the severance of the seed-head at this early stage undoubtedly con- serves much vigor for the vegetative parts. Where silver top is common, the grass is commonly deeper green, but whether this is cause or effect is difficult to say. Chicory is beginning to flower now. In Missouri this plant is an abundant weed along highways and in pastures. It does not become especially noticeable until June, spending both spring and fall as a prostrate rosette. Stock like to graze it in its immature stages, and to some extent even after its milky fruiting stems have begun to elongate. The seed of the plant is used as an ingredient of some pasture seed mixtures in England and has even occasionally been tried in this country. It has a long large tap-root which penetrates the soil deeply and thus serves some of the functions of sweet clover and alfalfa. Where pastures are not grazed in spring, chicory plants develop into tall bushy weeds full of pale blue flowers. Although adding a touch of beauty to a pasture, they become tough and inedible and leave their resistant stems to clutter up fall and winter grass. In many cases where chicory has become objectionable, it is because the affected pasture was consistently not grazed in spring. If stock are allowed access to it at that time they will graze it closely and reduce its vigor. Fields along streams are LI [Vor. 40 14 ANNALS OF THE MISSOURI BOTANICAL GARDEN especially subject to chicory infestation for the simple reason that floods frequently break through fences during the spring season. Procrastination in repairing these fences often gives the chicory time to develop and seed, and build up reserves. Then when pastures elsewhere are grazed out, the fences are repaired and stock turned in on the low fields full of mature chicory. This gets rid of the overburden of summer grass, so that in fall when the chicory plant again assumes a rosette type of growth and begins to store up food supplies, it has full access to sun. А single year's experiment to determine the effectiveness of spring grazing on restricting chicory growth was tried in some paddocks which had been ungrazed for some time and contained a heavy stand of chicory. Stock were allowed access to one paddock during April but were kept off the others till June. By July a strong reduction in height and vigor of the chicory stand was noted on the grazed paddock (fig. 18). The prominence of many weeds in certain fields can thus be traced back to the interrelationships of farm management and seasonal events. By mid-June the tedious mowing of permanent pastures has begun (fig. 15). Although it is customary to wait until the seed head reaches its full growth before any mowing begins, it is quite probable that earlier clipping might be advisable so that the resources expended on production of the inedible mature panicle might be conserved for vegetative development. The observations already made of tillers dying during the period of elongation indicate that the plant is under stress at this time. Earlier clipping would also prevent swaths from being so heavy. The blue- grass plant in the early flowering stage, when the panicle is just barely visible, is still fairly nitrogen-rich and soft and decays rapidly. As weeks pass, however, stems grow tough and siliceous, lie heavy on the new growth and disintegrate slowly. Even though the swaths are apparently obliterated in a couple of weeks by the rapidly growing June leaves, these leaves are forced to grow unnecessarily long through the weak light of the overburden and so deplete their food reserves. While it is essential to return this organic matter to the soil, it would be better to send it through the more efficient system, the animal, instead of trusting to the slow and expensive plan of decay. This is where the importance of grass silage comes in. Cut when grass is rich and generally in excess, and at the period when it is most beneficial to the plant, it is converted into manure and urine, rather than allowed to sink down slowly in an expensive disintegration. In well-grazed pastures there are now three main types of vegetation: (1) There are those areas in present use for grazing which are short, rich, green and non-flowering. These places indicate where urine fell this winter and spring.’ They are full of clover, buckhorn plaintain (Plantago lanceolata L.), and chicory, all of which are being readily eaten. The rich nitrogenous grass is eaten closely. (2) There are also those areas that were grazed last fall and winter, but are neglected for the time being. These show up now with short but well-developed inflorescences of bluegrass, mingled with chicory which has developed short flower- 1953] ETTER— BLUEGRASS PASTURE ALMANAC 15 ing shoots. Chickweed (Stellaria media (L.) Cyrillo) is common here, testifying to the fact that close fall grazing weakened the stand and exposed enough soil so that seed could germinate and the young seedlings receive sunlight all winter. Being winter annuals, they thrive best when winter competition is eliminated. Paspalum pubescens is also common in these places, since by the time fall and winter grazing occurs, this warm-weather species has already built up its reserves. Since fall-grazed bluegrass is slowed down in spring and its leaves shortened, the shoots of paspalum get off to a good start in May. Foxtail grass (Setaria lutescens (Weigel) Hubb.) is also beginning to become prominent in this short bluegrass. (3) A third kind of area distinguishable on the pasture is the rough about manure, where tall chicory and rank bluegrass grow, and little if any grazing is taking place. While these three conditions can be separately described as general tendencies where the pasture has been continuously grazed, they necessarily grade into each other to some extent. Heavy rains after a dry spell freshen up the bluegrass rapidly. Much of this rapid greening-up results from rhizomes which, caught underground during the dry period, were unable to move. With the rains and resumption of rapid growth many of them come to the surface. Most striking response is that of the grass along shallow drainages in the pastures. Much of the reaction in this case is due not only to the availability of moisture, but to the concentration of the nutrients washed down or leached from the pasture hills above. In response to the accumula- tion of soil and nutrients in these shallow drains, they commonly support a vege- tation distinct from that on adjacent slopes where removal of nutrient substances prevails. In one pasture, while the yellow soil of the slopes was thinly covered with a pure stand of Canada bluegrass the dark-soiled draws were distinctly marked off by tongues of Kentucky bluegrass. In another pasture, very poor and overgrazed, the hills were sparsely covered with triple-awn (Aristida oligantha Michx.), and lespedeza, with clumps of goldenrod (Solidago altissima L.) protecting a few re- maining patches of old top-soil. In the gullies were chains of undernourished cocklebur. Canada bluegrass is flowering now, and timothy heads are abundant in a few pastures and in the orchard. Redtop panicles are just now appearing. Ironweed flower buds are forming, lespedeza is coming up in pastures rooted by hogs this spring, and squirrel-tail grass (Hordeum jubatum L.) whitens the marshy swales in the hog pasture (fig. 16). This plant has the reputation of causing consider- able difficulty with stock, especially when its heads find their way into hay. The bristles on the grain cause mouth injuries and open the way for more serious in- fection through diseases. Occasionally now small grasshoppers appear in the grass, and the first horseflies are giving the saddle-horses trouble. In a few parts of a Canada bluegrass pasture which I fenced off, rabbits have grazed the grass very closely. In these spots, opened up to the full strength of the sun, seedlings of buttonweed (Diodia teres Walt.) and also of sheep sorrel (Rumex Acetosella L.) come up, giving the appearance of small clones. [Vor. 40 16 ANNALS OF THE MISSOURI BOTANICAL GARDEN In a 15-acre pasture where a single horse had been grazing for 12 days I counted 18 newly burned urine spots on a plot 50 feet square. Since a horse averages around 7 urinations a day, in 12 days there would have been 84. If these urinations had been distributed evenly over the pasture there would have been only a third of a spot on 2500 square feet of the counted area. The concentration was actually 54 times that, indicating the extent to which grazing and urination are localized. The site of these counts was near the exclosure on the terrace in fig. 1. The soil was a rich fill and benefitted from subirrigation from the slope above. Cattle, after having been turned into a pasture where pigs had run all winter, found many of the rank weeds to their liking. They ate avidly of the lamb’s- quarters (Chenopodium album L.), tall smartweed (Polygonum lapathifolium L.), giant ragweed, and chicory (fig. 17). In a few days they had almost cleaned out even the Mexican tea (Chenopodium ambrosioides L.) and cockleburs (Xanthium sp.). They seemed to thrive on these weeds, even though the cockleburs are some- times considered quite poisonous in their early stages of growth. About the only plants that the cattle could not stomach were common smartweed (Polygonum Hydropiper L.) and lady’s-thumb (Polygonum persicaria L.). Consequently these two plants are taking over the old hog-feeding area. This indulgence in weeds took place at a time when plenty of good pasture was available. JuLy The grass of pastures which have gone ungrazed is rank and lax, and is dis- regarded or trampled underfoot by stock. Where bluegrass has been kept short by continuous grazing, weedy annual grasses such as foxtail and crabgrass (Digitaria sanguinalis (L.) Scop.) come in. The former is especially common in closely grazed urine places, and actually contributes considerable forage. It is beginning to bloom, however, and that reduces its palatability. In the pastures not so closely grazed, stock are feeding on summer perennials such as purpletop and Paspalum pubescens, new shoots of sweet clover (Melilotus sp.), prickly lettuce (Lactuca sp.) and new lespedeza. Pastures on low ground, partly as a result of extensive rains in June and July, are becoming much more rank and weedy than during previous drier summers. Bur marigold (Bidens sp.), Croton capitatus Michx., wild carrot (Daucus Carota L.), chicory, and Paspalum pubescens are coming up everywhere. When these weeds are cut during July mowing they leave a heavy swath which smothers out the grass and provides open areas where new weeds such as corn gromwell (Litho- spermum arvense L.) can come in. At the same time, in an exclosure on this moist and rich low ground where no mowing at all went on last summer, much grass has drowned itself out, and weeds such as bull nettle (Solanum carolinense L.), black nightshade (Solanum nigrum L.), sedges (Carex and Cyperus sp.), and pokeweed (Phytolacca americana L.) are invading. The new poke plants are lim- ited strictly to the vicinity of the fence, emphasizing that the chief means of 1953] ETTER— BLUEGRASS PASTURE ALMANAC 17 dispersal of poke seeds is by way of the alimentary tract of birds that eat the berries. Wherever convenient resting places such as fence rails, brush piles or fallen trees occur, there poke soon appears. As these rank weeds develop in the grass, they shade it out. They are soon followed by giant ragweed and other rank bottomland vegetation, and eventually by elms and walnuts. On such rich wet ground removal of some of the rankness of the bluegrass is absolutely necessary for its own perpetuation. This can be most profitably accomplished by grazing, though careful late summer burning occasion- ally can be used. Mowing and raking is an expensive alternative. Upland blue- grass is much more stable and less subject to this self-destruction. Although mowing in July is often done, it has little effect on pastures then except to shorten the grass for a time and postpone a little the flowering of certain prominent weeds. This postponement of flowering might help to reduce seed output were it followed later on with a second mowing. It often happens, how- ever, that such mowings are never accomplished. Actually, many of these weeds are rhizomatous species, and occasional mowing seems to help more than hinder their spread. St. John’s-wort (Hypericum perforatum L.) and chicory cut a few weeks ago already show flowers again. Ironweed (Vernonia sp.) comes back strong, and broad-leaved dock (Китех obtusifolius L.) is scarcely affected. When Canada bluegrass growing on exposed hills with little top soil is mowed, the hot summer sun has free access to the soil surface, which becomes dry and powdery. The slow process of restoring organic matter to these sterile places is consequently made even slower by indiscriminate mowing. Many farm practices such as manure spreading, fertilizing and liming, mowing and seeding, are done without reference to the natural conditions of the land. Мо effort is made to use ecological information to reduce the expense of wasteful blanket measures. Seed is not suited to the site, and treatment is not suited to the plant. The manure spreader is put into gear and manure flies over raw points and rich fills alike. As much seed is used on rich as on poor land. Alfalfa may be sown not only on suit- able slopes, but on the poorly drained swales between. Mowing is performed when and where the opportunity offers without reference to the developmental stages of the plants or the nature of the soil. Heavy swaths smother bottomlands, and on thin soils the shrivelling litter offers no protection from sun and rain. To achieve maximum production and balance on an uneven farm, the ecological factors must enter into the picture. Now that the spring drouth has been broken, rhizomes of bluegrass are elongat- ing again. The drouth period is recorded on the rhizomes in the form of several extremely short joints, or internodes which contrast with the new ones nearly an inch long. While the axillary buds on these rhizomes seldom develop under ordinary sod conditions, this new flush of growth has resulted in considerable branching from the first one or two nodes of new growth. This branching seems to be a characteristic response of many plants to a break in drouth. [Vor. 40 18 ANNALS OF THE MISSOURI BOTANICAL GARDEN Gophers are throwing up mounds again. The fresh excavations are often dis- turbed and scattered by the mower bar during clipping of the pasture. They are also scattered by quail that use them for dusting places (fig. 21). These summer mounds usually develop a flora quite different from that of mounds pushed up in late fall. They are most commonly invaded by fleabane (Erigerom annuus (L.) Pers.), though also common are ironweed, boneset (Eupatorium sp.), thistle (Cirsium lanceolatum (L.) Hill), and other plants which germinate in summer, live for a fall and winter, and then flower the following summer. Old gopher diggings are often identified from a distance in summer pasture by patches of fleabane such as that shown in fig. 19. Bare soil areas are also prominent where horses stand head to tail, brushing constantly at the horse-flies that are now abundant. The continual stamping of their feet digs up the soil over a considerable area. On windy days horses usually retreat to some exposed point high on a hill where their efforts are assisted by the wind. In low shady places where the steer manure of last winter has remained moist, seed of honey-locust which passed through cattle undamaged have germinated, demonstrating one of the ways by which this plant spreads over pastures and waste land. Seeds of other plants also are able to navigate the alimentary canal of stock. From a small amount of hog manure planted in a flat with sterilized soil, bluegrass, nimble- Will (Muhlenbergia Schreberi Gmel.) , goosegrass (Eleusine indica (L.) Gaertn.) and Mexican tea germinated (Etter, 1948). AUGUST Although August, like July, is usually a hot month, its days are noticably shorter, and it is perhaps this difference that makes bluegrass change for the better toward the end of the month. Often following a good soaking rain leaves seem to become a deeper color and broaden out slightly. Stock will leave their summer diet of purpletop, foxtail, lespedeza and weeds to graze the bluegrass where it has | been kept short by summer traffic or close mowing, or where it has been grazed over by chickens for some time. A little anthocyanin may appear on the upper surface of leaves and on rhizome tips beginning to turn up. Some of the dormant summer buds on the crown of the plant have swollen. Most significant is the resumption of root initiation. As the month nears its end, short sprout-like rhizomes often develop, especially on meadow plants which have flowered. A rare plant or two, growing in a rich aerated nitrogenous soil, may even show the first development of a new tiller. This new activity depends on the availability of moisture, and if August is very dry resumption of growth may be delayed, to occur in a flush of activity after the first rain. While no amount of water in July can make bluegrass really palatable or healthy, irrigation in August and September is a different matter. By allowing fall growth to get away to an early start while days are still warm enough to encourage growth, total forage can be much encouraged and its nutritional 1953] ETTER— BLUEGRASS PASTURE ALMANAC 19 quality will be reasonably satisfactory. Where bluegrass pastures are depended on to provide overwintering forage in the fields, dry falls can be very serious. Other early signs that day length is getting short enough to produce results are the first fall flowering of dandelions and of annual bluegrass (Poa annua L.) and of the development from dry Canada bluegrass stems of new fall leaves which attract rabbits. The fall crop of grasshoppers also finds that this grass is begin- ning to have what they need to bring them to maturity, but many of them still prefer other more succulent sources of food. In lespedeza fields which were being saved for seed quite a few weeds were developing, especially prickly lettuce (Lactuca scariola L. and L. saligna L.). Grasshoppers invaded the field and while completely ignoring the lespedeza, they consumed leaves, seeds and stems of the weedy species, and almost entirely eliminated the weed problem before combining began. Grasshoppers are so uniformly castigated that it is interesting to observe also that they sometimes attack the giant ragweed with particular fury, consuming leaves and flowering heads before much pollen has had a chance to spill. Grass- hoppers are often just as choosy as cattle or horses about which plants they eat. In a field of ironweed, a very troublesome plant, grasshoppers may strip certain plants completely, leaving only the bare stems. At the same time adjacent plants may be untouched. lronweed is a highly variable species, and there are considerable differences in the leaf as to its hairiness, thickness, toughness, and greenness. There are probably equivalent differences in the chemistry of the tissues. It is small wonder that when they can afford it, the grasshoppers are particular. ugust is a time of spider webs, and in the morning dew-wet webs may be seen scattered widely over the pastures. They are also abundant in the woodland, and can sometimes be seen flying overhead on the wind. Mowing in mid or late August, while not a common practice, is very bene- ficial to pastures where summer grass has grown abundantly and deep. Mowing should be done just late enough to catch the last flowering shoots of the summer grasses such as purpletop, broom-sedge (Andropogon virginicus L.) crabgrass and foxtail so that no regrowth will occur. The removal of this overburden will per- mit the short fall blades of bluegrass to reap the benefit of all the fall sunlight which plays a big part in the amount of tillering that occurs as well as the width of leaves and the nutritiousness of the grass. This mowing should be done early enough, on the other hand, to allow rears fall top growth to occur. In souri, the third week in August seems a good time. Later mowing will reduce the crop of grass which may be used for fall grazing or saved for winter subsistence. SEPTEMBER The change which bluegrass began to feel in August has now brightened up the pastures except in ungrazed places where long summer leaves have begun to die and turn brown. Growth is rapid, summer rhizomes are turning up and leafing out, and in good soil a few plants may begin to form tillers. Just as in March, this September rebirth of activity results in a temporary absence of rhizomes. [Vor. 40 20 ANNALS OF THE MISSOURI BOTANICAL GARDEN Except for a few short sprouts in the meadow or on wounded plants, they are almost impossible to find. In good soil which has had plenty of air, water, and sun during August vigorous plants develop, and these often begin producing rhizomes in September and continue all fall. This fall growth is at the expense of stored reserves, and the growth it produces is rather proteinaceous, though low in sugars. Summer rhizomes in the soil are depleted and weakened. Sods tend to fall apart rather easily as new roots develop. New leaves are definitely shorter and broader and stock show renewed interest in bluegrass. They may lose weight, or gain very little during this period as a result of the low sugar-high protein composition of the forage, and at the same time dairy cattle often show signs of acetonemia, or sugar shortage. Аз the days get shorter and the nights cooler, however, sugars will begin to accumulate and by November bluegrass will have a fairy high carbohydrate content provided the pasture has not been excessively grazed in fall. While work horses take readily to this new fall grass after being turned loose, they seem generally to prefer to consume considerable quantities of the more mature higher carbohydrate growth of nimble- Will and crabgrass before settling down to bluegrass and clover. Winter annuals such as wild brome or cheat are now developing, just in time to replace the summer grasses such as purpletop, now flowering, and crabgrass and setaria, which are becoming rank or dying out. Cheat is now four to five inches high and furnishes good pasture. It is much sought after by stock and will con- tinue to provide pasture until well into winter. It will also return early in spring to give stock an early bite. This is the wild counterpart of such crops as winter rye or wheat, often planted for late fall and spring grazing. Аза matter of fact, while man has laboriously devised through trial and error pasture systems involving winter annuals and hot-weather annuals such as lespedeza or Sudan grass to sup- plement permanent grass, nature does it automatically on pastures, and stock take advantage of it instinctively. On a hill which was seeded several years ago to lespedeza and timothy there is now a striking demonstration of the advantage of mowing at the correct time (fig. 22). Part of this field was mowed the latter part of June, while the timothy heads were still green and the stalks tender. The lespedeza seedlings were just getting a good start. The other part of the field was not mowed until more than a month later. The heavy strawy swaths can still be seen on this pasture. The lespedeza seedlings, shaded during their early development and then covered with grass, have scarcely produced any fall forage. Lespedeza in bluegrass also benefits by early mowing in late May or early June. OcTOBER If the fall has been moist enough, October pastures may be as green as those of spring, but drouth in September and October, a frequent occurrence, cuts seriously into fall grass production. The greenest grass, as usual, is to be found in disturbed areas, where vehicles have passed, animals or men have walked, or chickens 1953] ETTER— BLUEGRASS PASTURE ALMANAC 21 scratched, in gardens or where new soil has been deposited either by flooding creeks or eroding rains. In these places fall bluegrass plants produce many rhizomes and tiller early and are sought by stock. If the plants have been extensively injured during late summer, however, as by burning, or by the cleats of farm machinery or by hooves of swine, the recovery growth may be very deep green and actively tillering, but very few if any rhizomes will develop. In general, such rhizomatous pasture or meadow species as quack grass, smooth brome, or redtop follow patterns of development similar to bluegrass, though slight differences in time of maturity are reflected in vegetative responses. None of these species is actively developing rhizomes now. Hogs are feeding on bluegrass again, as they did during spring, testifying to the return of high protein forage. Cool weather and short days cause bluegrass to produce short leaves and prostrate plants. Stock graze closely seeking the nutritious foliage. They are invading grass protected during summer by rank foxtail or purpletop (fig. 23). Areas where urine fell last spring are still preferred, though little holdover effect is apparent. In pastures grazed during September spots of green grass are beginning to show where cattle urinated, but it is apparent that urine spilled in the hot weather of July and August produces virtually no results. The combined heat and dryness of summer permit the nitrogen to escape. Plots on which urine was applied artificially in August and September gave an amazing demonstration of the difference a month made (fig. 24). This fact has important meaning for grazers, for while it is commonly felt that animals return to the soil most of what they take from it, the truth is that much of the value of urine is lost when it goes on summer pastures. This might well influence the type of man- agement which the farmer chooses to practice during that period, especially the dairy farmers who return the waste products of their cows to the fields during summer. It is quite possible that summer applications of manure lose so much nitrogen as to be unprofitable. Storage under protected conditions until cold weather might be more efficient. Where pastures were not mowed in late summer the grass is very slow to develop and few if any tillers and rhizomes have formed. Fresh horse urine still kills the grass, but invasion of these burned spots is rapid now, at least in moist soil where adjacent plants have begun active rhizoming. Stock are avoiding the tall green grass around manure clumps. The cessation of root initiation during summer causes an accumulation of buds without roots. During fall this backlog is reduced, and by October roots develop as soon as buds mature. These fall roots are fat and white and contrast with the slender wiry roots of summer. There is now much mole activity, and gophers begin to build mounds out into closely grazed places seeking the succulent plants of dandelion and chicory. Clover is blooming a little, and aphids are back on the soft grass leaves. Under the oak trees where horses stand, fallen acorns are cracked beneath their hooves, and the soil becomes dry and powdery. Quail come to these trees, enjoy dust baths, and [Vor. 40 22 ANNALS OF THE MISSOURI BOTANICAL GARDEN pick up fragments from the cracked acorn shells. In the woods pasture they follow the hogs that crack hickory and walnuts and acorns with their teeth. Frag- ments dropped to the ground provide the quail with a convenient meal. Bluegrass is often noticably superior under walnut trees. While this may be a complicated relationship involving microorganisms and obscure biochemistry, it also involves in part the fact that walnut leaves are shed early in fall, sometimes by mid-September, when bluegrass begins its important fall growing. Other trees often retain their leaves until November when all warm weather has passed. The shade cast in summer by walnuts is not dense enough to weaken growth, but is sufficient to limit rhizome development so that the grass does not become sod- u The sun of fall then allows tillering to progress and winter reserves to build up. To test the influence of shade on bluegrass between September and November, a shade of muslin sack was placed a foot above the sod at the same time that walnut leaves were shed. This shade was left on until late-shedding trees such as maples had become bare. It had a noticeable effect on grass leaves, making them more lax, waxy yellowish green, and delayed the death or browning of summer leaves. In addition it was later noticed that the body of the fall leaves was much different. The shaded ones were fragile and shriveled and gray, while those in full sun retained their shape well and died to a tough yellow brown. The repercussions of this treatment were not followed the next year, but undoubtedly influenced the quality of recovery growth in spring, as well as the size and shape of the florescence. NovEMBER In November bluegrass begins to suffer from the short days and the cold. Following the first frosts and the first freezes the blades shorten and cling to the ground. Rhizomes in vigorous grasses continue to be initiated, and to turn up, as they will, more slowly through the early winter. Nearly every new bud that develops now becomes a tiller. Gopher mounds are being invaded by rhizomes from adjacent plants. When the fall has been dry, stock begin to graze the rough areas where long grass remains. Clover leaves have been killed by the frost and swine are feeding on bluegrass, young chicory, and acorns. Canada bluegrass is just beginning to send out its rhizomes, which will continue growing off and on all winter and will turn up next spring to produce flowering shoots. astures where sows and pigs were running during August were so softened by heavy rains that the unringed sows were able to attack the sod and uproot much of it in their search for worms, roots and other food. These rough bare areas still remain, testimony to the results of delaying too long ringing of the sows (fig. 25). In spring these rootings will be a maze of weeds, DECEMBER Occasional frost has little effect on bluegrass, but in December when the soil freezes, bluegrass growth slows down greatly. With several inches of the top-soil frozen the leaves often turn dark, become flushed with anthocyanin. Wherever 1953] ETTER— BLUEGRASS PASTURE ALMANAC 23 bluegrass would like to grow but can’t, as in frozen ground, dry ground, intense sunlight, or because of injury to growing leaves, it has a tendency to develop this purple coloring. Uprooted rhizomes become pink at the tips or occasionally for their full length. This is especially noticable in Johnson grass (Sorghum hale pense (L.) Pers.) where the large fat rhizomes often turn a brilliant cerise when dug up and allowed to remain in the sun. Many other species seem to obey a similar compulsion. Around manure clumps, the grass remains surprisingly green, and bluegrass rhizomes work beneath the clumps with indifference to the cold. The manure insulates the ground and prevents its freezing, thus allowing bluegrass to continue its growth and to build up its reserves. As a result of this advantage, the grass in these clumps gets away to an early start in spring. Bluegrass leaves produced in December are very short, only an inch or two long. Except where the grass has been kept short, they are lost within the longer fall and summer leaves. When these old leaves die, pastures turn brown. In heavily pastured places, however, the grass is still green and stock still seek the sugar-rich grass. In many places, they are eating cheat seedlings in preference. Gophers are very active, throwing up mounds overnight, especially in the longer grass areas where the soil is not so solidly frozen. Mice are once again building extensive tunnels in the sod, and when snow melts away, elaborate feed- ing runs bordered with cut leaves can be found in the shorter grass, ending here and there in large “haystacks” of reserve food. Most fall rhizomes have turned up and now rest at the surface of the ground, though a few have been caught underground in the frozen soil. Tillering is still occurring, although the new shoots take several weeks to become large enough to be seen outside the retaining sheaths of their axillant leaf. Even more important than tillering, however, is the beginning of flower initiation. This process began on a few early plants last month, but is now general. A careful examination of the growing point shows a slight accumulation of extra plant segments at its tip, and it is from these that the panicle will develop. So it is that preparation for June pastures, full of waving bluegrass, is made even during this cold month half a year in advance. BIBLIOGRAPHY mec G E. (19 36). The ice cam of pro ger - available nitrogen supply on the owth of pasture in the spring. Jour. Agr. Sci. 26:6 Curis, Ja Au and Max L. App (1948). чаи of fire be Mere competition between bluegrass and ain prairie plants. Amer. Midl. Nat. 39:437-4 Etter, А. С. (1948). Seeds Hat db livestock. r M Gard. Sera 36:170-172. — — —, (1951). How Kentucky bluegrass grow . Gard. 38:293-37 ne У. В . (1941). Cow's urine as a fertilizer fe quies e Jour. De Sci. 24:761- [Ver. 40, 1953] 24 ANNALS OF THE MISSOURI BOTANICAL GARDEN EXPLANATION OF PLATE PLATE 1 Fig. 1. January. On a pasture hill near the top, rooting hogs have defined the con- tact of a limestone and shale bed. Scepage along the contact favors their activity. Result- ing changes in the vegetation will mark this contact for some time. Fig. 2. Janu ry. Gilts in an orchard-grass pasture that was not mowed last summer beat paths among the resulting tussocks. Fig. 3. February. During periods when snow and ice cover the ground horses feed on tall grass preserved in manured or weedy areas. There they can easily scrape the snow away with their hooves. Fig. 4. February. Rank grass in orchards provides cover for cottontails. For sus- tenance, however, the rabbits prefer special spots where the grass is short and deep green. hese spots usually show evidence о aving received applications of urine. often grazed completely bare, roots, rhizomes and leaves all being consumed. Fig. 5. March. Urine applied in a circle around this post in fall resulted in com- plete utilization of all grass, including subterranean parts, by mice. Even the post was aten where urine had fallen on it. Fig. 6. April. Old gopher mounds of last fall are partly covered with grass, but new ones remain barren and spring winds whiten them. These mounds, especially the spring diggings, provide opportunity for weed seeds to germinate, but also produce grass which is darker green, productive, and preferred by stock. : ANN. Mo. Вот. Garb., Vor. 40, 1953 PLATE 1 ETTER—BLUEGRASS PASTURES [Vor. 40, 1953] ` 26 ANNALS OF THE MISSOURI BOTANICAL GARDEN EXPLANATION OF PLATE PLATE 2 18. 7. March. At this time of year the pattern of horse grazing is very apparent. - Rank grass remains uneaten while small areas are consistently grazed. "These spots received continual applications of urine and remained green all winter. Fig. 8. April. Stems of last summer's purpletop stand in the new rank grass of spring. When temperatures rise sufficiently to encourage molds and mildews, these stems will drop into the surrounding grass. Fig. 9. April. Stock do not graze the rank grass round clumps of their own manure : and rough areas thus appear in closely grazed pastures. Where cows and horses are grazed | together these roughs are not so apparent since no prejudice seems to exist for grass con- | taminated by droppings of other animals. 1 ig. 10. May. The spots of darker and taller from steers during fall and winter. The resultin is preferred by stock. This pasture has grass in this picture received urinations — $ grass growth is more nutritious and — not yet been grazed. Fig. 11. June. When stock are turned in on a pasture such as that in fig. 10, they seek out the urine spots and consume the grass that grows there. Fig. 12. June. Horses graze limited areas intensivel these areas gradually expand. Urinations which burn the periphery of grazings. The dead grass will be replaced b will be grazed closely. Nutrit y. Under heavy grazing pressure - y rich weeds and clover which - ious bluegrass will soon invade the spots and restore the turf. ANN. Mo. Bor. GARD., Vor. 40, 1953 PLATE 2 ETTER—BLUEGRASS PA [Vor. 40, 1953] 28 ANNALS OF THE MISSOURI BOTANICAL GARDEN EXPLANATION ОЕ PLATE PLATE 3 Fig. June. When horse and sheep graze together stable rough areas are built up where horse manure accumulates, Sheep, like horses, prefer short grass. As grazing pressure increases rough areas are gradually whittled away. Fig. 14. June. This illustration shows the effects of urine on the vegetation of a lawn. Not only does the grass thrive in the right-hand plot, but weeds are eliminated. Application here was about two-thirds gallon per square yard and was applied in June, the previous year, on moist soil. 8. e. Mowing of bluegrass pastures commences. The heavy swaths will soon disappear beneath long summer leaves, but earlier mowing would have been desirable and would have maintained the food value of the grass at a higher level. Fig. 16. June. In the swales of an old hog pasture squirrel-tail grass heads out. This grass often causes mouth injuries to stock that consume it in hay. Fig. 17. June. Cattle turned into a new pasture in June sought first the rank growth of weeds that had grown up in areas where hogs had fed and urinated. The lamb's-quarters _ in the picture was cleaned out as were other weed species. Fig. 18. June. Chicory is a bad Pasture weed in some parts of Missouri. It can be controlled to a considerable extent by grazing it in s i er t grazed in spring at all while paddock in foreground was Grazing chicory while it is in the rosette stage weakens it and reduces the height of the coarse flowering stems. on zs га > < E un» o < 24 O 2 = р са а н H m i rol + EAE ۴ и [Vor. 40, 1953] 30 ANNALS OF THE MISSOURI BOTANICAL GARDEN EXPLANATION OF PLATE PLATE 4 Fig. 19 у. Gopher mounds in rank pastures are often ries at a distance by a white growth of annual fleabane which germinates in the disturbed so Fig. 20. “August. Rabbits are choosey about where they graze. This small spot of crabgrass was grazed into the dirt. The grass was visibly darker pas and had apparently benefitted from urine of some species Fig. 21. July. Quail use gopher mounds for dusting places. Fig. 22. September. The time to mow is critical in pasture management and it varies We the еске In foreground а field of timothy and lespedeza was mowed while timothy heads were young and green. In distance where fall herbage is much inferior, mowing had bern pois. till July. Fig. 23. October. In late samme and fall ricus oat comes into rich heavily grazed pastures. For a while it provides some good mids azing, but, when seeding, is unpalatable. In fall, hungry stock invade the old foil beat it to the ground ^ November. Urine has à striking effect on bluegrass та in both fall and spring. In this picture, it is shown that this effect dep ends time of application. Grass in the left plot was given urine in late August; on da is a Loisir application in late September. There was no visible effect from the earlier is on. Fig. 25. December. Late s mmer rains softened the soil and unr ringed sows ie. the turf. If this oc. were disced and harrowed slightly i in early fall it odi diced recover. If allowed to go into spring in this condition it will be covered with weeds. ANN. Mo. Вот. Garp., Vor. 40, 1953 PLATE 4 ETTER—BLUEGRASS PASTURES THE POPCORNS OF TURKEY EDGAR ANDERSON, Missouri Botanical Garden N WILLIAM L. BROWN, Pioneer Hi-Bred Corn Company The special significance of popcorns in the history of maize has been recognized by many students of that perennially fascinating problem (see, for instance, Sturte- vant, 1894, Mangelsdorf, 1948). To scientists at large, however, the classification of popcorns and the description of the various ways in which they are used have been matters of such little moment that we were eventually driven to making our own world survey of popcorns and their uses (see Anderson and Cutler, 1950, for a general discussion). It was not until a decade of research had taught us to inquire in some of the most unlikely places that we learned (through Volney Jones) that popcorn is commonly used throughout Turkey and is a characteristic feature of many Turkish villages. At the very moment when we were endeavoring to have a comprehensive collection made for us there, Dr. Jack Harlan returned from that country with a collection of economic plants made for the Division of Plant Exploration and Introduction of the United States Department of Agricul- ture. Fifty-four collections of popcorn made by Dr. Harlan and his collaborators form the basis of the following report. We are indebted to Dr. Harlan for various observations supplementing the unusually complete data turned over to us by the Department of Agriculture, and to Dr. M. M. Hoover and his staff of the Regional Introduction Station at Ames, Iowa. It is a pleasure to acknowledge the efficiency and courtesy of this entire organization. We were welcomed to the increase plots in good weather and in bad, our attention was called to other collections of possible significance, and we were supplied promptly with viable seed of all the cultures we wanted to grow. For a few varieties of particular interest pertinent informa- tion was quickly produced from the files, and remnant seed of the original collec- tions was made available to us. The 54 Turkish popcorns collected by Dr. Harlan were grown in the experi- mental plots of the Pioneer Hi-Bred Corn Company at Johnston, Iowa, 20-30 plants being grown of each collection. А few varieties of particular interest were grown in replicate, and flint varieties and flint-dent mixtures from Dr. Harlan's collections were available for comparison in another plot. Each variety was scored for morphological uniformity, for season, for tassel type. Detailed measurements were taken on plant height, ear height, ear number, number of leaves above the ear, leaf length, leaf width, tassel exsertion, internode length and internode pattern, the width of the central spike of the tassel, glume length, the number of tertiary branches on the lowermost secondary branch of the tassel, the number of branch- lets of the fourth order (if any), the number of secondary tassel branches, and the pubescence of the leaf sheath. Ас harvest time the plants were scored for husk number, for number of apparent nodes in the shank (the difference between these two numbers represents the number of condensed [telescoped] nodes of the (33) \ Уот. 40 34 ANNALS OF THE MISSOURI BOTANICAL GARDEN shank), shank length, shank width, kernel row number, kernel thickness, pith diameter, basal enlargement of the ear, pointing of the kernel, and color of peri- carp and endosperm. At silking time regulation photographs of representative plants were made against a measured background. “Inclusive” herbarium specimens (Anderson, 1951) were made as a permanent record of each collection, each speci- men including a photograph of the entire plant against a measured background and pressed central spikes and lower tassel branches. Sporocytes of each culture were pickled in aceto-alcohol and smeared in the cytological laboratory. The number of chromosome knobs (a variable and diag- nostic feature in maize—see, for instance, Mangelsdorf and Cameron, 1942) were determined and wherever possible knob positions were worked out. The results of these various measurements and scores are presented in Tables I and II. Photo- graphs of representative plants are shown in pls. 5 and 6. As soon as the collection began to tassel it was clear that maize is as hetero- geneous a mixture in Turkey as are many other crops in that area (see Harlan, 1951). The variation within collections and between collections exceeded that in any of the collections of exotic maize from various parts of the world which we have had under observation. Three extreme types and three mixtures and inter- mediates between these extremes were apparent in the collection. One of these extremes was a popcorn identical with the variety known as Japanese Hull-less Popcorn. Since it came from Ankara, the capital city, and since most of the intermediates between it and the other two types came from the vicinity of Ankara or Istambul, the older capital, it probably represents a comparatively recent introduction into Turkey. It was strikingly different from the other two extremes. It had pubescent leaves, a heavy tassel with an extremely thick central spike, and the ear was elliptical in cross-section with a high number of rows of kernels, all of these features being characteristic of the variety Japanese Hull-less. The other two extreme types were unlike рорсогпѕ commonly grown in the United States, and the two extremes, though connected by various intermediates, were quite unlike one another. One was very early to mature; the other was very late. The early type was short, with an exserted tassel, large glumes, and stiff upright tassel branches. Characteristically, the type had a low number of chromo- some knobs, usually from none to three with either no knob on chromosome No. 6, or merely one small one. The late extreme was tall with numerous short inter- nodes above the ear, which was borne high on the plant. The tassel was charac- terized by small glumes and lax branches, at maturity even the central spike remaining drooped over. The tassel was so included in the sheaths of the upper leaves that the lowest tassel branches were never completely free from the sheath of the uppermost leaf. Characteristically, it had seven to eight chromosome knobs, including two on chromosome No. 6 and one on No. 8. All these facts are set forth in the tables and the plates. Figures 1 and 2 show diagrammatically, but to scale, representative plants from the three most extreme collections of the two types. 1953] ANDERSON AND BROWN-—TURKISH POPCORNS 35 The early extreme bears no close resemblance to any known race of maize though it is vaguely similar to several. It is rather like a somewhat degenerate form of the Northern Flint corns which were once characteristic of eastern Nort America. It is shorter and earlier than most of them, with somewhat smaller ears, higher row numbers on the average, and without the more or less enlarged base to the ear which is one of their most distinctive features. The late extreme, on the other hand, belongs to a group of little-known pop- corns with an intriguingly peculiar distribution. They are commonly grown by various aboriginal tribes along the borders of China and India; they were in early prehistoric times the prevailing type of maize on the coast of Peru and Chile; and they are still to be found here and there in various out-of-the-way places in South America. We received some years ago from Lorenzo Parodi, of Argentina, a pop- corn of this general type and have a few scattered collections from Peru, Chile, and Colombia. Mangelsdorf and Oliver (1951) described and illustrated one from Caldas, Colombia. The collections from Turkey agreed with these Asiatic and South American popcorns in various technical characteristics of the inflorescence (see Alava, 1952)—in their knob numbers, the numerous ears high on the plant, the short upper internodes, and the drooping, included tassels. In its height, leaf shape, etc., the collection made by Harlan at Samsun was similar to varieties we have grown from the mountains of Siam, from the Naga tribes of Assam and from the Lushai Hills. Not only did it agree in general with these Oriental and ancient uth American popcorns; with a popcorn collected in the mountains of Nepal (P.I. 166,162 called to our attention by Dr. Hoover), it was virtually identical. The collections from Nepal and from Turkey differed not as much as would two strains of any open-pollinated (ie. non-hybrid) popcorn variety (such as Japanese Hull-less or South American) grown in the United States. The three extreme types described above were apprehended by repeated study of the plants in the breeding plot. The exact association of characters was worked out more objectively by means of pictorialized scatter diagrams. Опе of these is presented in fig. 3. It has been set up to demonstrate the relationship for the seven characters which best differentiate the early flint-like popcorns from the late “Asiatic” extremes. Аз fig. 3 demonstrates, there is in the Turkish collections an over-all association for these seven characters, tassel exsertion, percentage of total height in the internodes above the ear, glume length, chromosome knob number, shank width, plant height, and ear height. At the upper right-hand corner of the diagram are the Asiatic extremes. Аз the diagram demonstrates, they are all plants with ears high on the plant, the internodes above the ears being propor- tionately short. They have narrow shanks, small glumes, and high knob numbers. At the lower left-hand corner of the diagram are the opposite extremes, a more variable lot. They tend to be short, low-eared plants, with exserted tassels and proportionately long internodes below the tassel. They have wide shanks, long glumes, and few or no knobs on their chromosomes. It will be seen that though the Asiatic extremes seem to be somewhat set off from the rest there is a gradual Fig. 1 Three examples of the “Asiatic” race of popcorn, highly diagrammatic but exactly to scale CX 25). The lengths of the internodes, of the leaves, of the shanks, and of the branches of the tassel are all to scale, and the shape of the tassel is copied from photographs of fully mature tassels. Each diagram represents a different collection. Note the short internodes, the drooping included tassels, and the multiple ears high on the plant. NHGWVO 'IVOINV.LOH MNOSSIN AHL AO SIVNNV 10A]‏ ض0 1953] ANDERSON AND BROWN-—TURKISH POPCORNS а. EF /,. Ex d TM TUN da ы Lr а c. и hree examples of the “Aegean” race of popcorn, at the upper part of the plant, and the low ears F ү т p Fig. 2 У e| Lm тү, —. a to the same scale and prepared in the same way as fig. 1. Note the exserted tassels, the longer internodes [Vor. 40 58 ANNALS OF THE MISSOURI BOTANICAL GARDEN EAR HEIGHT IN DECIMETERS : v 13] * v У + 12} y éd 11 є Ф я xi i QN Ф, 6 Ri và 4 ‚ ёе at Ф é * $ у $ Ф, é | ө. 199 % «o é & eood 1 6% % rates VY we 5 € ò 6 © 6 4 e 6 e n „а. نہد ن‎ 11 12 13 14 15 16 17 18 19 20 21 22 23 24 PLANT HEIGHT IN DECIMETERS TASSEL EXSERTION GLUME LENGTH (MM) 0.3 -1 7-9 é Fos Ф v e 6-2 ы и - PERCENT TOTAL HEIGH IN UPPER INTERNODE Е "i 0-1 X li Ф 2-4 w. 5-6 e 35- 9 3-4 e 0-2 Fi Pictorialized scatter dem ее. ue variation in irs n of Turkish popcorns for seven hese age Each d a dia of a single plant for each collection, the plant having € osen in the field as pa e n e ds EK. aee ery Зет: Further explanation in the text. 1953] ANDERSON AND BROWN-—TURKISH POPCORNS 39 transition from one extreme to the other. The diagram presents an exact picture of the way in which two diverse races of maize have been brought together and have intermingled, forming a variable set of intermediates. Some of these are extreme recombinations in which characters of one race have been combined with characters from the other. For the whole group of varieties, however, there can be demonstrated the continuing association, om the average, of those characters which went in together. For those not familiar with the recent work on races of maize it should be pointed out that just such a mingling of correlated variables is characteristic of maize. It has been shown for Mexico y Wellhausen, Roberts, and Hernandez in collaboration with Mangelsdorf (1951, 1952), who have greatly extended the preliminary studies of Anderson and Cutler (1942) and Anderson (1946). For the maize of the American corn belt where purposeful mixing of two races has been carried on intensively, Brown and Anderson (1947) were able through these correlations to work out the history of the mingling by these indirect methods before carrying on the historical and archaeological research which proved that it had indeed occurred in just such a way (Anderson and Brown, 1952). There can be little doubt then that the heterogeneous popcorns of Turkey were largely derived from the mingling of popcorns from the two widely different races of maize, the resulting mixture being made somewhat more diverse by the com- paratively recent addition of still other types of popcorn. After the racial affiliations of the 54 collections made by Dr. Harlan had been worked out, their distribution in Turkey was determined from his collecting notes. The resulting picture is a simple one. The Asiatic race in its purest form is char- acteristic of the mountainous areas in northeastern Turkey. The other race is characteristic of the Aegean shores of Anatolia and we are accordingly referring to it as the Aegean race. It is possible to turn the pictorialized scatter diagram of fig. 3 into a crude sort of index running from 0 to 14. By plotting these index numbers in four grades it is possible to demonstrate the mingling of the Aegean and Asiatic races of popcorn as reflected by Dr. Harlan's collections (fig. 4). Conclusions: From the evidence presented above we conclude that popcorn is widespread in Turkey. Though heterogeneous there even for maize, it can be assigned to two provisional races, the Aegean and the Asiatic. The former is commonest along the coast, the latter in the northeastern mountains. Most of the popcorns of Turkey are various intermediates between these two extremes, occasionally modified by more recent introductions of modern commercial varieties. In the above report there is little or no evidence to indicate when or where or by whom these popcorns were introduced into Turkey. То answer such questions we have begun a series of ethnological, linguistic, and historical inquiries. These investigations have proceeded to the point where we are confident that though the problem is a complex one, it is not chaotic. It seems probable to us, in the . 4, Distribution of ав. pape — made by Dr. чи апа ove тай in fig. The numerous = in and around gre the capitol i т; been ous values o E n turned in crude - runn or D for typical "Aeg varieties to 14 for typical “Asiatic” varieties. Solid bla k dots eR ent c ollectio ons si like the Asiatic мча. e (those in de upper right-ha ad corner of dn ; open circles, Pa most extreme Aegean types; lightly stippled dots, intermediates ыы Аереап; heavily verae dots, "змс ав resembling Asiatic race. N3GWVO 'IVOINV.ILOH ГЯПО$$ПА AHL AO SIVNNV OF TOA] 1953] ANDERSON AND BROWN-—TURKISH POPCORNS 41 light of all the evidence, that the Asiatic and the Aegean popcorns were brought to Turkey at different times, and by different routes. Further discussion must be postponed until other evidence can be presented. LITERATURE CITED Alava, Reino, O. (1952). Spikelet variation = Zea Mays І. Ann. Mo. Bot. Gard.39:65—96. Anderson, Edgar (1946). Maize in Mexico: preliminary study. —€— 33:147-247. ; (1947). Field studies of Guat smalan maize. Ibid. - Ош). Inclusive herbaria. Jour. Ind. Genet. & РІ. 2 11: :1-3. , and William L. Brown Eu The history of the common maize varieties of the United States corn belt. Agr. Hist — — ——, and ‘Hugh С. Cutler ( Ts Races of Zea Mays: Their recognition and classification. Ann. Mo. jo Gard. 29:69-88. ————, — — ——, (1950). CE of corn popping and their historical significance. South- t. Jour. ae 6:303—3 FEES William L., and Edgar A read (1947). A Northern Flint corns. Ann. Mo. Bot. Gard. 34:1-28 Harlan, J. R. (1951). Anatomy of gene centers. ys Nat. 85:97-1 е ret Раш С. (19 si iion role of pod corn in she origin d evolution of maize. Ann. t. Gar “ ч :377 —_ ae s W. E (1942). Western си. 2 е аа center of origin of dire maize varieties. Bot. Mus. Leafl., Har жг 10:217-252. — —— —, and D. L r (1951). Whence came m Aper) ‘bid. 14:263—291. Sturtevant, E L. (1894). Notes on maize. Bull. ne Bor. Club 21:319-343, 503-523. ен п, Е. Ј., et al. (1951). Razas de maiz en Mexico. Foll. Tec. 5, Ofic. Estud. Езрес:, "M y Ganad. Mexico, D. F. 952). Races of maize in Mexico. Bussey Inst. Harv. Univ. TABLE I Turkish popcorns listed in the order of rod dcum pla seed w nt introduction number ach rity, Ja у мы For collection there is given the place where the ted, the season of matu of the plants grown from that dir genda the "bendi tae lant in decimeters, the height of the uppermost ear in decimeters, the number of ears, the y^ gth of the u pou internode in centi- meters, е M of фет! male чур: in millimeters and the Index number on a seven-fold index maa values of 0 nai treme “Aegean” types and of 14 for extreme “Asiatic” types (further ie ine in з s 8 K > ч > ч ta ч ЕР : Be toe c asdaM ia. d 5 sii g Е 2а |а 2 | 25 | iii) BEBO) D$ 4 #23 Я © i © i af | Нес ES ЕЕ - REA Е di - Mok) Mss | 25 |9258 | Wwe | ^^ Б 164,994 | Cayirhan, Ankara Mid Uniform 14 8 2 2.3 10 3 g 165,036 | Nallihan, Ankara Mid Uniform | 15 8 2 2.4 0 1 165,057 Beypazare, Ankara Mid Variable | 16 9 2 1.8 11 1 "i 167,030 | Kirik-Khan, Hatay Late Variable | 18 11 1 2.2 12 5 le 167,9 dd Balikesir Mid Uniform 15 5 3 1.9 10 5 m 167,948 | Bursa Early Uniform | 15 6 2 7 12 0 < 167,949 | Balikesir Early — 16 6 2 1.9 13 3 | 167,963 | Bucak, Burdur Late Uniform 18 9 2 1.8 8 8 Ф 167,967 | Osman Вики, Aydin Mid Uniform | 15 7 2 1.7 9 8 o 167,983 | Koycegiz, Mugla : Very Late pem 23 15 2 1.6 9 12 c 167,989 | Akhisar, Manisa 16 9 2 2.4 9 2 E 168,000 | Istanbul Mid A 16 8 2 24 9 4 Lc 168,006 | Kirklareli Mid Uniform | 15 10 2 2.0 10 4 O 168,007 | Kirklareli Mid Uniform | 16 10 1 ZI 9 4 H 168,008 | Kirklareli Mid Uniform | 18 11 2 Zi 9 5 > 168,016 | Tekirda Mid Uniform | 15 6 2 2.4 10 1 2 168,017 | Tekirdag Mid Uniform | 15 4 2 22 10 2 Ф 168,019 | From a seed s Early 16 9 1 2.3 10 2 р 168,027 | Ayvacik, Canakkale Early — 14 6 2 2% 12 0 168,032 | Edremit id 16 6 2 2.4 9 2 o 168,033 | Balikesir Early Usiforin 12 5 2 24 10 2 > 168,039 | Kulak, Balikesir Early Uniform 15 6 1 1.3 11 2 6 170,877 | Manisa Mid Uniform | 14 6 2 2.5 11 1 tr 170,879 | Soma, Manisa Early Uniform 13 4 1 Pi 12 2 2 170,880 | Istanbul Mid — 13 6 2 2.7 12 2 170,881 | Cayirova, rtr Mid Uniform | 18 10 3 1.6 11 6 171,896 | Inceoglu, Bartin — 13 10 3 1.4 10 8 e 171,912 | ' Unifor 22 13 3 2.5 9 14 = 171,915 Uniform | 23 13 3 1.2 8 14 3 171,91 Tokat Uniform | 22 15 3 1.0 10 13 è 172,600 indir, ата = 13 8 2 2.0 10 3 173,830 | Malatya Uniform 14 6 2 22 11 1 174,414 | Mardin 3 2 1.7 11 6 ани os " 75,976 167,030 170,881 170,881 171,904 Biga, Canakkale Biga, Canakkale rzincan Ayfon, Karahisa Araplar, ДАДЕ Balikesir > Ss o'm m > 25 B. = > ч p ecik, Ankara \lpagut, Sakarya Eskisehir Altinova, Ayvalik, Balikesir Araplar Sehir, Sebin Karahisar, Giresun Bolce Agac, Manyas, Balikesir Elbistan- us Igdir, Karahis Gemlik, Barsa Karga ili, Kastanou Ka are t-S Zo nguldak Kozam, i Kilic, Kesan, Edirne irne Lalapasa, ым eer a (near) Mani = чора Bafra, Sam Akashic J Kh isifite te, from farmers at Kirik-Khan, Hatay eli 0 Cayirova, Kocaeli Seed dealer in Samsun Early (Mid) Late Very Late Uniform for Uniform Variable Uniform Variable Variable Variable Uniform — ` чаи м سم‎ mm T nw m м олоо ое мм Ч eee NNN © NNN et ЫБ Ne Рә Рә Мә МӘ NEN NUN ж ie ies ка NIN NINN Ri Д wenn о œ N ее о — о їч Umea оошо ооо ооо МЧ Фо (о М фо їл оло? н m n س‎ ROW m | eS each] ide eR T ЧЫТЕ 6 ©‏ سم سر Rt‏ ي ی س طب ما \л‏ ++ m boo م‎ SNOOdOd HSIMYNL—NAOUT аму NOSYAGNY tv [£561 [Vor. 40 44 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE П Numbers and positions of chromosome knobs for the collection of = from Turkey. The ern at ti left are > eges ris introduction numbers of the Bureau of Plant Pu d Introduction. Numbe to the heading t the right refer pe the 10 chrom es of Zea r as rec бшш! a maize cytogenetics The final column, “В”, records the ct EE of “В” chromosomes, if an Wher mber is given, there were no "B" chromosomes. —1 indicates a s kicb on аб one of $ed E Seep, “ү” — small T" = terminal, эш минах Knob position 2 3 4 5 6 7 8 9 10 Р.І. 164994 4 2 —1|— IT >. I. 165036 1 1 >. I. 165057 4 1 1 1 fT P.1. 167030 4 1 1 1 DI P.1. 167947 Ж 4 2 1 LE >. |. 167948 2 0 1 1 0 0 » 1. 167949 3 0 1 1T >, I. 167963 6 1 1? 2 1 1 >, I. 167967 5 0 1 1 1 1 0 —1 |— >, I. 167983 6-7 2 1 Dr >, 1. 167989 2 0 1 0 0 › 1. 168000 4 1 0 —1 | — HT > I. 168006 1 1 0 ›. I. 168007 2 1? 1 >, I. 168008 3 0 1 ҮТ >, I. 168016 3 1 1 1T › I. 168017 4 1 1 1 rr >, 1. 168019 3 1 1 iT P.L 168027 1 1 >». I. 168032 3 1 1 0 1T 0 >, I. 168033 2 1 0 1 >, 1. 168039 0 >, I. 170877 3 1 1 fr >, I. 170879 1 0 1Т >. I. 170880 3 1 0 1 4 > 1. 170881 = 2 1 1 . L 171896 5 1s 1? 2 1T 5$ 171912 8 0 1 0 1? 2 1 1 2 0 » 1: 171915 7-8 15 2 1 1 >, 1. 171917 7? 2 1 1 HI 0 >. I. 17260 6 1s 1 1 2 PT 0 . 173830 3 0 1s 0 1 то '. I. 174414 3 1 1 0 1 0 0 0 „ 1. 175976 4 1 1 1 ae >. I. 175978 3 1 1 Ge 0 . 1. 176801 6 1 2 1 1 1T . I. 176804 1 1 0 . 177114 2 1? 1T . L 177583 5 1? 2 1 Жеш лч. 177598 3 1 0 1 IT] 0 І. 177622 2 1 1T . 1. 177624 4 1 0 1 TES . I. 177625 2 0 Is 0 1 0 0 Р. I. 177626 2 1? 0 1 Р. L17762 6 0 1 0 1 1 2 1? 0 0 Р. 1. 177631 5 1 1 2 1 . L 177637 3 1 1 0 . 1. 177641 5 1 i 2 с : . L 177643 2 1 0 " M а „179127 1 1T . L 179131 5 1 1 0 { iT 1953] ANDERSON AND BROWN—TURKISH POPCORNS 45 Number Knob position — of knobs |" 2 3 7 8 9 10 | B P. I. 179566 2 0 iT 0 P. I. 179568 2 2 P. I. 182324 2 0 1 1T 0 Р. I. 182326 4 1 1 2 P. I. 182327 2 0 1 1T Р. [. 182785 1 sd Р. 1.183750 4 1? iT P. I. 183751 5? IT P. I. 183756 4 2 1 IT P. I. 183764 1 1 P. I. 183771 3 0 2 IT 0 P. I. 183783 5 1 0 0 2 IT 0 P.1. 183787 5 2 P. 1. 183790 5 0 1 1 2 0 Р.І. 183798 4 1 1 1 1 1 P. I. 185045 3 1 1T P. I. 185063 3L 1 1 ІТ Р.І. 185073 2? 1 т P.I. 167030 2 1T P. I. 167030 5L 2 1 1 1T 0 P. I. 170881 3L 1 1 1 P. I. 170881 3 1 1 1 P. I. 171904 8 1 2 0 1 2 0 [Vor. 40, 1953] 46 ANNALS OF THE MISSOURI BOTANICAL GARDEN EXPLANATION OF PLATE PLATE 5 Four plants of Dr. Harlan' s collection from Samsun, Turkey, photographed against a measured background, at the time when the X: were actively shedding pollen. Note the drooping tassels, barely exserted from the eaves, the bs ih on the plant. Specimen No. 4 has had its leaf blades and on $i its leaf күк сы аи The lines іп the background аге 25 cm. apart. PLATE 5 Mo. Вот. Garp., Vor. 40, 1953 ANN. ANDERSON AND BROWN—TURKISH POPCORNS : | Vor. 40, 1953] 48 ANNALS OF THE MISSOURI BOTANICAL GARDEN ExPLANATION OF PLATE PLATE 6 The same “Asiatic” popcorn of pl. 5, photographed in the field at harvest time. e ants photographed in the experimental plot. Note a: ears high on the plane, -de ber g tassels, and the plant- to-plant uniformity. These plants were grown а actually collected in Turkey. The uniformity ot this variety is сысы ot collections made both in Nepal and in Turkey 2. Three plants from the same row as those in fig. 1, laid down on the earth of the дый plot. е the sare ears, high on the Slat: the leaves continuing to the verv base of the tassel, the drooping tassel кан the slender stem, and the short silks. These are all VETTEN А of Asiatic popcorn Mo. Bor. Garp., Vor. 40, 1953 ANN. u5 Z 22 О о a е a m ^ z ра р 1 Z 2 о ec © A Z < Z О Ze] га ыы A Z < А PRELIMINARY SURVEY OF THE MILPA SYSTEM OF MAIZE CULTURE AS PRACTICED BY THE MAYA INDIANS OF THE NORTHERN PART OF THE YUCATAN PENINSULA R. A. EMERSON FoREWORD In January and February of 1935, өн А. Emerson, of von Celery) an МЕ T. HE REDE, then of the U. S. Dep nt "| Agricult ture, studied m e farming a Maya Indians at the request of the Car beg а о My reins тие їп mimeographed form under Dr. Emerson’s name, is of the ta ir e results of this preliminary Though it has been cited in at least two bibliographies its very existence has been unsus- inl by many students of maize. indebted to Mrs. Ida K. he ee р the А Natural e of Philadelphia, for supplying us with a y, and to Dr. Н. E. D. Pollock, of the E of Archaeology of the Carnegie ій "tee Hort ag а to reprint the Up uii the time of his merson was the heart np center of resear rch in maize e on his part, has been identified with much of the research on maize which was ca mid o in the U Department of Agriculture and had taken the lead in зра maize and its dnd in rica. oe the “pe decade, Mexico, in age and Mexico’s — of growing maize, in particular, have changed a n ever- EEE v» e. Dr. Emerson's report, though preliminary and informal, ives us an itu of maize in an out-of-the-way part ie Mexico ат years ago as seen through the eyes of Б. rein maize scholar of his time,.—EpGAR AN PURPOSE OF THE SURVEY In the course of archeological studies of the ancient Maya civilization, con- ducted by the Carnegie Institution of Washington and other agencies, it has be- come important to estimate the density of population that could have been main- tained by the agricultural system then in use. Is it possible that the Mayas of one or two thousand years ago could have used a more intensive type of agriculture than that employed by Mayas of the present day? Could the fact that large cities and their contributing territories were abandoned, some of them to be rehabilitated only after some hundreds of years, have been due even in part to the inadequacy of the agriculture of these regions? Could a possible decline in the productivity of the agricultural land have necessitated abandonment after a century or more? Was such postulated decline in soil fertility related casually to the system of agriculture in vogue? It was not so much to find conclusive answers to such questions as these as it Was to form an opinion of the possible value of a more thorough and sustained agronomic study as an aid in answering them that a preliminary survey of parts of the former Maya region was undertaken. REGION COVERED BY THE SURVEY The survey was conducted under the auspices of the Division of Historical Research of the Carnegie Institution of Washington and under the tutelage of Dr. S. G. Morley of the Chichen Itza project of that Institution. It was begun the latter part of January and continued to the end of February, 1935. The (51) [Vor. 40 52 ANNALS OF THE MISSOURI BOTANICAL GARDEN personnel consisted of Mr. J. H. Kempton of the United States Department of Agriculture, and of the writer. "Throughout the study they had the help of an interpreter and usually of a local milpero. The study of milpas was begun at Piste in the central part of the state of Yucatan near Chichen Itza and about 120 kilometers from Merida. Trips between Merida and Chichen Itza by automobile and by train over different routes, a trip by train from Merida southeast to Peto about 150 kilometers, and another by train from Merida southwest to Campeche about 180 kilometers, afforded glimpses of the western half of the state of Yucatan and the northwestern part of the state of Campeche. From a number of places along these lines of railroad, milpas at dis- tances ranging from a few up to 40 kilometers were visited by automobile, tram- car, mule cart, on horse-back, or afoot. Such points of departure were: Piste, Oxkutzkab, and Peto in the state of Yucatan, and Campeche in the state of ampec TYPE OF COUNTRY VISITED Topography.—The parts of Yucatan and of Campeche that were visited, as is said to be true of all of the northern part of the Yucatan peninsula, are a low, slightly undulating plain, broken in the southwestern part by ranges of low hills themselves only a few hundred feet above sea level. There is not even a suggestion of a river or brook in all that part of: the peninsula visited. Limestone rock.—The soil is underlaid throughout by a rather soft and porous limestone. Everywhere there appear outcrops of this rock. The sides and tops of small low knolls, often seen to be all rock, and the intervening areas, only a few meters less elevated than the knolls, exhibit many outcrops of rocks. In fact, it is never far from the surface to stratified rock. Although this soft limestone seems to harden on exposure after being taken from quarries, there are abundant evi- dences of its weathering to produce the usually scant soil. Everywhere exposed rocks are pitted with holes from the size of pin-heads to a meter across and half that deep. Both vertical and horizontal crevices of variable sizes are to be noted greater or less distances apart. Such crevices extend downward for considerable distances, perhaps even to the permanent water table which lies from a few to 50 or more meters below the surface, the distance apparently depending on the rela- tive elevation of the surface. Numerous caves and even large caverns exist. Not infrequently these extend down to or considerably below the water table. When such caverns are open to the surface, they constitute the natural wells, cenotes, of the region, upon which the Mayas depend for their water supply. Soils.—Evidences of the weathering of limestone rocks to produce soil are so universal that one is inclined to speculate on what would have been the nature of the land surface, had these rocks been exposed for ages to the alternate freezing and thawing in latitudes far to the north. Presumably the low plain would have been covered with some meters of fine loam instead of the scant soil covering now seen. And this might have resulted in a very different type of agriculture from that now prevailing. 1953] EMERSON-——MILPA AGRICULTURE IN YUCATAN 53 On the numerous low knolls the soil is seldom more than a few centimeters deep and on many of them soil is seen only in pockets, or crevices in the rock. Crevices seem to be filled with surface soil and humus to considerable depths, as can be readily observed in recently worked stone quarries. Between the knolls the soil may be a half meter in depth, in some places more, and in many places much less than that. The only relatively deep soils observed are those bordering the ranges of low hills in the southwestern part of the area visited. At Oxkutzkab such border soils were said to be as much as five meters deep. The near absence of surface stone and the luxuriant growth and productivity of citrus trees in this region make such an estimate seem not unreasonable. Evidently the heaviest rains have carried down to the plain enough sediment to build up a relatively deep soil at the foot of the hills. Yet, there is little evidence of even temporary waterways on the hillsides and no evidence whatever of brooks paralleling the base of the hills or leading away from them. The water presumably spreads out in a thin sheet over the plain, depositing much of its sediment before disappearing into rock crevices. These bands of deeper soil at the base of the hills are of very minor importance for the region as a whole, for they are seldom as much as a kilometer wide and in many places are very narrow. The surface soil, to a depth of a few centimeters on land on which the bush has been undisturbed for from ten to twenty years, is made up largely of decaying vegetable matter. During the dry winter season there is a sparse covering of dry leaves, nowhere the heavy covering seen in northern forests. Beneath this layer of leaves is a shallow layer of partly decayed vegetable matter and below that the red or brown loam. Evidently the relatively high temperature and considerable moisture of the region induce such rapid decay of the fallen leaves that no deep layer of decaying vegetable matter ever accumulates. Climate.—The entire peninsula of Yucatan lies within the tropics, the north- ernmost part being somewhat below the 22nd parallel of latitude. The rainfall is light in the extreme northern and northwestern parts of the peninsula. Near Merida and along the gulf coast as far south as Campeche, the annual rainfall is said to be not over 18 to 20 inches. At Chichen Itza the annual precipitation approaches 50 inches. The year is divided into two seasons, a summer rainy season of about five months, usually from May to September inclusive but some- times beginning nearly a month earlier or а month later, and a dry winter season of about seven months from October to April inclusive. Some of the weather records kept at Chichen Itza for the past nine or ten years were examined. The total precipitation for 1933 was 50.1 inches and for 1934, 46.6 inches. During the five months, May to September, of 1933, 86 per cent of the year’s rain fell; during the same months of 1934, only 73 per cent of the annual rainfall was recorded. Of the 730 days of the two-year period, rain fell on 275 days, 185 of them in the two five-month summer seasons. The maximum rainfall for one day, 4.75 inches, occurred in September, 1933. One-half inch or more of rain fell on 56 of the 275 rainy days, and 46 of the 56 were in the Мау- [Vor. 40 54 ANNALS OF THE MISSOURI BOTANICAL GARDEN September periods. During the two-year period there were 168 days on which .2 inches or less of rain fell, 102 of them having been in the five-month summer seasons. The mean of the maximum daily temperatures for the years 1933 and 1934 at Chichen Itza was 91.6? F., and the mean of the daily minimum temperatures was 65.6° Е. The highest temperatures occurred in March, April, and early May, maximum temperatures of 101°, 103? and 105° Е. respectively having been re- corded for these months in 1933. When the rainy season begins the maximum temperatures are not as high. Minimum temperatures are lowest during the period from late November to early March. During 1933 the lowest temperature re- corded was 47^ F., which occurred twice in December; in 1934, 44? was recorded twice in March, and 43° once in December. The lowest temperature on record was 40° Е. on January 27, 1935. Vegetation.—Except for land in crops, the whole of the northwestern part of the Yucatan peninsula is covered by "bush". "This consists for the most part of trees and of vines which cling to the trees. In land not in crops for ten to twenty years, the dominant trees are 10 to 15 meters high and.10 to 20 centimeters in diameter. Under these are many smaller trees that have died or been suppressed by the shade of the larger trees. Relatively few annual weeds and almost no grasses are seen in any heavy bush. Both of these occur, however, along trails and in the more open places in the bush. In bush of only a few years’ growth, many annual weeds and small vines are to be seen. Among these are amaranths, numerous composites, and several kinds of bindweed. As the trees grow and shade the ground more and more, these weeds are less and less common. Even in very recently abandoned milpas little grass is seen except in the extreme northern and western parts of the area visited. Неге the many large henequen haciendas, in which the bush has been kept cut for ten to fifteen years, apparently return to bush less rapidly when abandoned and more grass is seen. The surface outcrop of rock does not seem to interfere much with tree growth. It is not uncommon to see large trees with roots exposed on the rock surface for several meters. Evidently the roots grow out in the thin cover of soil until a crevice is reached, then turn abruptly downward. In rock quarries roots 5 centi- meters in diameter have been found 2—3 meters beneath the land surface. THE MILPA SYSTEM OF AGRICULTURE Location of milpas.—Difficulty was experienced in learning why a Maya Indian selects a particular site for his milpa. The Mexican Government, in some regions at least, has assigned areas of bush around the several villages for the use of the inhabitants of those villages. If a milpa is made beyond the designated area on Federal land, a tax is exacted by the Government. Milpas have been visited which were 8 to 12 kilometers from the village in which the milperos making them lived. The factors that seem to govern the selection of a site are: type of soil, ease of 1953] EMERSON—MILPA AGRICULTURE IN YUCATAN 55 cutting the bush, nearness to the village, the presence of water, etc. There are undoubtedly differences in bush soils, though it is not easy to see the difference tween what the milpero calls good and poor soils. There appears to be a size range in bush trees on either side of which cutting the trees is more difficult. Trees of more than say 15 centimeters diameter are not easily cut with the small axes of milperos. Likewise, small trees sometimes bend under blows of the axe and must be cut with machetes. Moreover, and perhaps of more importance, in bush made up largely of small trees, there are many more trees per unit area. When milpas are made near a village, the milperos return to the village after their day's work. When a milpa is more than an hour's walk away—5 or 6 kilometers—the milpero usually moves to his milpa during the periods when he must work more or less steadily. In such instances, nearness to water, a cenote or a well, is of prime importance. The necessity of building a fence around a milpa, if near a village, where cattle and horses have free range, may often be a factor influencing the selection of a site far enough away to obviate the necessity for this extra labor. Making the Milpa.—A site having been selected, the milpero cuts the bush in the early part of the dry season, November to January, and allows the trees to lie as cut until they are dry enough to burn. The largest trees are usually left stand- ing. Among these are the ramon tree, the leaves and small branches of which are much used as forage for domestic animals, and the zapote, from which the chicle is obtained. The few palm trees, the leaves of which furnish the thatch for Maya huts, are almost never cut. The trees are usually cut at about waist height. Where fences are necessary, many of the smaller trees near the periphery of the milpa are left to be cut as needed for construction of the brush fence. In some milpas, some of the smaller trees are cut at heights of 2 or more meters, the stumps later serving as supports for the bean vines grown with the maize. In March or April, before the rainy season, the cut and then dry bush is burned. The milpero waits for a day of relatively strong wind and starts his fire on the windward side of the milpa. By this time many of the stumps left standing have ‚ Sprouts a meter or more long. This new growth is killed by the fire, and if the bush were a heavy one, the stumps and the large uncut trees may be killed, but usually, though considerably burned, many of them are left alive. Lundell, in his ‘Preliminary sketch of the phytogeography of the Yucatan peninsula,” refers re- peatedly to devastating fires which not only burn the felled bush but also sweep through the whole countryside. In this visit to northwestern Yucatan, absolutely no evidence of such forest fires was seen. Particular attention was given to this Possibility throughout the trip; and in no instance was evidence found of milpa fires having burned for more than a few meters into the surrounding green bush. This study was made before the season of milpa burning, but one large milpa fire near Chichen Itza was watched. Even with a strong wind, the fire burned only the dry leaves and small dead trees for only a few meters beyond the felled bush; and the fire had ceased to burn in the uncut bush while it was still burning hotly ; [Vor. 40 56 ANNALS OF THE MISSOURI BOTANICAL GARDEN in the trunks of the larger felled trees. Sometimes а milpero takes the precaution of removing the cut bush around the edges of his milpa for a few meters before setting the fire, but most often this precaution is to protect from fire his newly erected brush fence. And perhaps this measure is only apparently one of pre- caution; the fence is most easily built with the cut bush that is nearest it. When the rainy season begins, which may be in April or as late as the end of May, the maize is planted. А sharpened stick, usually with a metal point, is used to make a hole in the ground. In this are dropped four or five kernels of maize. Seeds of a native squash and of the common bean (always a black-seeded one) or of a lima bean are mixed with the maize and all are planted together at random. No attempt is made to arrange the hills in rows, and they are spaced irregularly averaging probably a little less than a meter and a half apart. Not infrequently hills are planted in narrow rock crevices or in the deeper pockets in the surface limestone. If this were not done, considerable areas on the low knolls could not be planted at all. Маше seems to grow in rock crevices as well as trees do. In the deeper soils at the base of the range of hills in the western part of the area visited, maize is planted in rows about one and a half meters apart with hills not over half a meter apart in the rows. Sprouts of tree stumps and weeds that appear in a first-year milpa are cut with a machete from once to three times during the growing season of the maize. Very rarely is this done three times and probably more often once than twice. By harvest time the weeds and sprouts have grown to a length of from 1 to 3 meters. When the maize ears begin to mature, the stalks are bent over below the ears, so that the upper part of the stalks, including the ears, hangs down; this is done, it is said, to protect the ears from rain and from birds The maize harvest extends over a considerable period of time. It is not un- common for a milpero to harvest a basket of ears whenever the previously harvested basketful has been used up. By the early part of February, however, most of the maize has been harvested and stored. The ears are gathered into palm-leaf baskets, with the husks on, and piled near a thatched hut in the milpa. The part of the crop that is to be used or sold at once is husked and shelled. Ears that are stored for any considerable period are left in the husks. A crib of stocks is built a little above the dirt floor of the storage hut, and the ears are packed in it tightly in a vertical position with the tips downward. This method of storage is said to lessen injury from rodents and insects. In preparing a milpa for a second crop of maize, the weeds, maize stalks are not cut early in the dry season as was the b The cutting is delayed until toward the end of the dry season, it was said because: (1) The small cut stems dry out quickly and therefore need no long drying period; (2) if cut early, more weeds and stump sprouts would grow before burning; and (3) the relatively small amount of trash does not produce a sufficiently hot fire to kill late-grown weeds and sprouts. As a matter of fact, this trash has to be piled as Cut in order to insure its burning at all. A serious difficulty arising from this stump sprouts, and ush the year before. 1953] ЕМЕК$ОМ—МПРА AGRICULTURE IN YUCATAN 57 delayed cutting, seemingly overlooked by the Maya milpero, is that it affords time for the ripening and dispersal of an abundant crop of weed seeds which are not destroyed in any large measure by the fires which burn the piled trash. The first-year milpa, having been cleared and burned, a second maize crop is planted and tended just as the first one was. For the reason noted above, the weeds on a second-year milpa are much more abundant than on a first-year опе. On visiting a milpa it was rarely difficult to decide at once whether it was a first- or a second-year one, using relative abundance of weeds as the sole criterion. Occasionally, maize is grown on the same milpa for three consecutive years. On the narrow belts of deeper soil at the base of ranges of hills, maize is some- times grown on the same milpa for four years. But, in an overwhelming per- centage of cases, two years is the limit. The milpa is then abandoned to return to bush and a new one prepared by cutting and burning another area of bush. The length of time that a piece of land is left in bush before being cut again for a milpa varies greatly. Rarely, this period is as short as four or five years, and in some instances it is as long as fifteen or twenty years. There is insufficient evi- dence upon which to base a positive statement of the average length of this bush period, but it is probably not far from ten years. It would seem that the essential factor is a length of time sufficient for the dominant trees to reach a size satis- factory for cutting, and to choke out the smaller bush, and incidentally the annual weeds also. YIELDS OF MAIZE GROWN BY THE MILPA SYSTEM As in other places, maize yields in Yucatan vary from season to season. Throughout much of the area visited 1-1 cargas per mecate is a good yield for а first-year milpa, and one or а little less than one for a second-year milpa. It should be explained that a carga is 42 kilos, or about 93 pounds, approximately 125 bushels, and that a mecate is a plot of land 20 20 meters, or 400 sq. meters, almost one-tenth of an acre. While yields of 15 to 20 bushels per acre might we be regarded as unsatisfactory to a farmer in the corn belt of the United States, the surprising thing is that so good yields are obtained under the conditions pre- vailing in Yucatan. DECLINE IN MAIZE YIELDS That in general the yield of maize is somewhat less in a second-year than in a first-year milpa is everywhere evident. The bearing of this fact on the customary abandonment of a milpa after the second crop and its possible bearing on the withdrawal from a whole region by the ancient Mayas makes its cause or causes worth more than passing notice. [Vor. 40 58 ANNALS OF THE MISSOURI BOTANICAL GARDEN Soil depletion—An opinion, apparently somewhat generally held, is that the system of agriculture now practiced by the Mayas results in rapid depletion of soil fertility. Lundell, in his "Phytogeography of Yucatan,” says: The ashes are the only fertilizer the soil receives, so it is quickly exhausted. Little humus ept the land. The cr the s remains after fires have swep op is best the first year; econd year it falls off, and in either the third or fourth year the clearing is abandoned and another site he ab i i in a la is chosen. years, during which | the fertility of the soil is partly restored by the rank vegeta- tio: i EO. uos interva. Apparently continued rotation leads to complete soil exhaustio: Since the trees of northern and western Yucatan are hardwoods, and ashes from their burning undoubtedly provide readily available mineral nutrients for the first crop of maize, it seems unlikely that there can be much leaching of these nutrients, for, even after the maize is ripe, the weeds and stump sprouts continue to grow and are not cut until late in spring. It is true also that the fire which burns the dried bush in preparation for the first-year crop burns the few leaves then on the ground and destroys some of the scant layer of humus. Rarely, how- ever, does this destruction of humus extend for more than a couple of centimeters below the soil surface, and that far only where relatively large tree trunks have burned. The burning of stalks, weeds, and stump sprouts preparatory to the second crop has little effect on the remaining humus. This vegetation is so scanty that it has to be piled to insure its burning at all. One with an agronomic background finds it difficult to believe that milpas, after two crops of maize, have been abandoned because of soil depletion, and equally difficult to conceive of soil fertility, once depleted, being restored by a few years of tree growth. Even without an agricultural background, one might reason- ably question how such weeds as amaranths in a second-year milpa could grow to a height of 2-3 meters with a spread of branches nearly equal to their height if the soil were nearing exhaustion. Weed competition.—In bush of several years standing, such as is felled for a first-year milpa, there are relatively few annual weeds, and these are found mainly along the narrow trails and in the more open parts of the bush. Many of the seeds of these annuals are undoubtedly killed by the intense heat of the burning slash. But some certainly escape, for there is always a considerable number of weeds in a first-year milpa. Since the earlier appearing weeds are cut usually only once, and that during the maize-growing season, weeds continue to grow long after the crop is ripe and are still flowering and only beginning to ripen seed in ebruary; and since these late-growing weeds, along with the maize stalks and stump sprouts, are not cut until near the end of the dry season, there is oppor- tunity for the ripening and dispersal of an abundant crop of seeds. Only a few large weeds may produce enough seed to stock the whole milpa. The burning of scattered piles of trash late in spring can have little destructive effect on the weed seeds which by that time are widely spread over the soil. The system is an ideal one for stocking the second-year milpa with noxious weeds. It seems almost too 1953] EMERSON—MILPA AGRICULTURE IN YUCATAN 59 obvious to require statement that weed competition rather than soil depletion is the factor primarily responsible for the lessened yield of the second-year milpa. It seems equally clear that tree growth after abandonment of a milpa functions primarily in choking annual weeds rather than in restoring depleted soil fertility. The labor differential between first-year and second-year milpas.—Statements of milperos agree that more time is required to cut the weeds during summer in a second-year than in a first-year milpa. Moreover, it takes materially more labor to cut, pile, and burn the maize stalks, weeds, and stump sprouts in preparation for a second-year milpa than to cut and burn the bush for a first year milpa. Only from careful records of labor requirements can this differential be accurately determined, but it would not be surprising to find that the preparation and care of a second-year milpa necessitate from one and a half to two times the amount of labor that a first-year milpa requires. THE MILPA SYSTEM OF AGRICULTURE IN RELATION TO DENSITY OF POPULATION Dr. Steggerda’s study of the dietary of the Maya Indians of the village of Piste indicates that maize furnishes about 85 per cent of the Maya’s diet. е problem of food supply adequate for a given population is, therefore, largely one of how much maize can be grown in the region. Dr. Steggerda's records indicate that an average Maya family of five requires annually about 30 cargas of maize. In so far as can be determined from a hurried survey such as this, the usual yield of maize is 1 to somewhat more than 1 carga per mecate, and an average size of milpa for one family is probably not far from 40 mecates. It would seem, there- fore, that the Maya now grows enough maize for his family plus what the chickens, dogs, and other animals eat, with some to sell to provide the family with the cheap cotton clothes they wear and the few other purchased articles they require. Since a milpa is used ordinarily for only two years and then abandoned, and if about a ten-year interval in bush is necessary before the land is again used for maize, a crop can be produced on the average about one year in six. In other words, about 17 per cent of the entire area might be in maize each year. This statement is based on the assumption that none of the land is unsuited to maize production—an assumption that is almost literally true for the parts of the Yuca- tan peninsula covered in this survey. The water problem could be solved by artificial wells, such as are now in use. What percentage of this area is now in milpas in any one year is not known. But going over the country by rail, auto- mobile, mule cart, and on horse-back, one is impressed by the very small fraction of the country in milpas. This impression was strengthened by observations made in going by air over the entire northern part of Yucatan from Merida to the east coast of the peninsula. While only by accurate records of sample areas can one Obtain the necessary information, it would surprise the writer if such records, when obtained, show so much as 1 per cent of the land in milpa at a given time. [Vor. 40 60 ANNALS OF THE MISSOURI BOTANICAL GARDEN If this guess is not too wild to be credited, 15 to 20 times as much maize might be produced each year as is now grown, thus supporting that many times as large a population as is now found in the Yucatan peninsula, and this without changing the present system of milpa agriculture. Мо evidence was found in this short- time survey to indicate that the present system of milpa agriculture could not maintained indefinitely. It would seem, therefore, that 15 to 20 times as great a population as the rather sparse one now in existence could find an adequate food supply. Even this figure would probably fall far short of the tremendous density of ancient population postulated by some archeologists in moments of ultra en- thusiasm. The factual bases for such estimates are unknown to the writer and are perhaps beyond the powers of comprehension of a mere maize specialist. They do, however, call for a consideration of the possibilities that, in ancient times, a more intensive type of agriculture may have been practiced. In the territory covered by this survey, there is no evidence of the type of in- tensive agriculture inferred from mountain-side terraces such as those seen by the writer some years ago near Inca remains in Peru. Of course, no such things are to be expected in a low and flat country such as northern and western Yucatan. Whether the ancient Mayas employed fertilizers other than the ashes from burned trees or used partly rotted vegetable matter as a manure—as the present- day Maya Indians certainly do not—it is idle to guess. The advanced civilization of the ancient Mayas would certainly have been capable of inventing tillage implements. They had no beasts of burden, but the postulated population should have furnished man power enough to obviate the need of horses or oxen. Even though the ancient Mayas had no metal tools, they could have made plows and cultivators, as they did axes and knives, from stone and hard wood. But why waste time on such details? Even our modern steel implements of tillage could not possibly be used anywhere in Yucatan except in the deeper soils at the base of the ranges of hills. Even if the outcropping rock did not incapacitate such implements in the first half hour of use, the rocks would keep them from reaching the intervening soil pockets. The main basis for the opinion that ancient Maya agriculture was much like present-day milpa agriculture is the fact that no other type can be regarded as having a chance of successful use. The Mayas of today use the only method available; and the ancient Mayas presum- ably used the same system for the same reason. There remains one possibility not yet discussed. Could the ancient Mayas have kept the land in maize more than two years at a time, perhaps many more than two? Or, could they have shortened the interval during which the land was abandoned to go back to bush? Ву either of these ways the percentage of the whole area in maize at any one time could have been correspondingly increased. The agronomic problem involved in such longer or more frequent use of the land for maize is largely one of greater weed competition resulting in lessened yields or of more labor per unit area to hold the weeds in check. 1953] EMERSON—MILPA AGRICULTURE IN YUCATAN 61 Information obtained from Maya milperos indicates that it requires from one and one-half to two times as much labor to care for a second-year as for a first- year milpa. Cutting the weeds, stalks, and sprouts in a milpa as soon as the maize is mature enough to harvest, and thus before the weed seeds have ripened and been scattered, instead of waiting as at present until near the end of the dry season, should materially lessen the weeds to be fought the next season. But there is no way of knowing whether the ancient Maya's routine differed from that in vogue at present. How much time does a Maya spend in caring for a milpa large enough to supply his dependents with maize for the year? If the ancient Mayas used materially more than one-sixth of the land for maize each year, could the milperos have devoted enough time to their milpas to meet the weed problem involved? А milpero cuts about two mecates of high bush a day. In cutting the trash in prep- aration for a second-year milpa, he covers about the same amount of land, but this takes the greater part of a day, while cutting two mecates of bush for a first- year milpa seldom requires all day. Two mecates seem to be the day’s stint, irre- spective of how many hours of work are required. Rarely is more than one mecate cleaned of weeds in one day’s work during the growing season of maize in a second-year milpa. Under the present system, therefore, to prepare land enough for the maize needed by one family—say 40 mecates, or about 4 acres—and to weed it once in summer requires about sixty days of labor for one man. Burning the felled trees or the weeds and stalks, planting, and harvesting could hardly require more than twenty days. It follows that the present-day Maya milpero has about three-fourths of his time free from the labor necessarily connected with his milpa. The ancient Maya milpera grew cotton as well as maize, but the Maya of today grows enough more corn so that he can sell some in order to buy cotton cloth. Moreover, the ancient milperos must have grown enough maize for their priests and rulers in addition to the requirements of their own families. But even so, the ancients could have spent more time on their milpas than do their descend- ants of the present time, and still have had much time in which to build pyramids, temples, and the like. All this, if true, suggests that the present milpa system of agriculture, with only minor modifications, could have supported a very much greater population than is found today in the Yucatan peninsula—how much greater only careful long continued studies can determine. at effects longer and more frequent use of the land for maize growing would have on soil fertility is beyond the writer’s ability at guessing. Certainly soil erosion could have played no part in bringing about the abandonment of a low, flat region like that around Chichen Itza. There is no evidence whatever, so far as one who is not a geologist can determine, to indicate the presence of such “washes” as one sees commonly in the much drier regions, deserts, of California. But these regions are bordered by real mountains, not merely the low hills of Yucatan; and the steepness of their slopes is a surprise to any automobile driver who is likely to have his engine stall on what seems almost level ground. [Vi с. 40, 1953] 62 ANNALS OF THE MISSOURI BOTANICAL GARDEN The absence of erosion in northern and western Yucatan is no argument against the possibility that erosion was an important factor in causing the abandonment of the higher and more broken regions farther south where the old-empire Mayas flourished for a time only to abandon the area later and migrate to the low lands of Yucatan. Whether or not the erosion theory is the correct answer to the aban- donment of the territory of the old-empire Mayas, it is now serving a useful purpose as the goblin of the present erosion campaign in the United States. RECOMMENDATIONS The survey here reported was only a preliminary study. Although statements of milperos are perhaps correct, many of the figures given here are at best on y estimates and some of them are guesses, perhaps very poor guesses. At best a hurried survey could not be expected to do more than to indicate what the elements of the problem are and to suggest the kinds of accurate information needed. he writer is convinced that weed competition and the labor involved in weed control is a more important factor than soil exhaustion, and yet, the latter might enter to a greater or less degree, particularly if the land were used for maize cul- ture for relatively long periods and at relatively frequent intervals. The only way to get reliable information concerning the possible depletion of soil fertility by maize culture is to conduct, for several years, a carefully planned and carefully executed field-plot experiment. But this would require the services of a man trained in modern agronomic technique for at least part of each year, and of a trustworthy helper to keep detailed records during the growing season. But there are problems of perhaps more immediate importance which should be undertaken at once, if their solution is of enough importance archeologically to warrant the expense involved. And, for the solution of these problems also, it is of prime importance to have a well-trained man with a well-founded agronomic background. Such a study should provide accurate information to take the place of the hear-say evidence, the estimates, and the guesses presented in this report. It would seem sufficient to select six or eight sample areas, such as the territory contributory to Piste already studied in part by Dr. Steggerda. It is important to know the number of people gaining their support in a given area, the amount of maize produced and used for food by them and their domestic animals, or sold in order that other necessities may be purchased, the number of mecates cultivated in order to provide this maize, and the relation that the area of all the milpas bears to the entire area. Accurate information should be obtained also respecting the hours of labor per unit area devoted to each separate operation involved and in the preparation and care of milpas. If a field-plot experiment is undertaken, one series of plots should be kept free from weeds at whatever cost, while another series is cared for as the Mayas now tend their milpas. Such an experiment should afford information not only in respect to possible soil depletion, as measured by maize yields, but the labor records should be of value in appraising the possibility that the ancient Mayas may have kept the land in milpas a greater percentage of the time than is now done. ee eer STAFF .- OF THE MISSOURI BOTANICAL GARDEN Director GEORGE T. Moore Assistant Director EDGAR ANDERSON CARROLL W. DODGE, HUGH C. CUTLER, Mycologist Curator Museum of Economie Plants ROBERT E. WooDsoN, JR., Gong P- B. Van AN $снАлск, Curator of the Herbarium of neo N. ANDREWS, JULIAN A. нме botanist Honorary Research Associate Rorra M. Try Nett C. Horner, gat istant Gate ‘of Len Librarian and Editor erbarium of Publications —~ GrRALD ULRICH .. Business Manager | BOARD oF TRUSTEES ОЕ ТНЕ MISSOURI BOTANTAL GARDEN 5 bd | RICHARD J. Lock Vice-President Dant К, К. CATLIN Second Vice-President EucENE Perros . ^. оны S. dcs DE qe ` ETHAN H. SHEPLEY Er um : Romer Brookes: Sur ; о a EX-OFFICIO ‘MEMBERS 3 Автыча Н. Compton, _ : dH _5тадтово L. маек. Chancellor of Washington pe d kt “у = ЕЕ | THUR © LICHTENBERGER, | os Volume XL Number 2 Annals of the Missouri Botanical Garden MAY, 1953 А Contribution to the Tiden Flora of Arizona and New Mexico $i Emanuel D. ). Rudolph The Analysis of Suspected Hybrids, as s Illustrated by | Berberis X Sladwynensis. - . iens Anderson Variation in Cob Morphology among Certain Archaeo : Nort Et The Cytology, a and Systematic Relationships of e map cm p сарат Hort. ex Berg: om = = Marilyn Ашу age n dem and Ethnological Races of Маше . . - A ee 79-1 ш» issouri Botanical Garden. A e A С м L B r ic 1 Annals of the Missouri Botanical Garden Vol. 40 MAY, 1953 No. 2 A CONTRIBUTION TO THE LICHEN FLORA OF ARIZONA AND NEW MEXICO EMANUEL D. RUDOLPH* The lichen flora of the states of Arizona and New Mexico is much less known than that of many other sections of the United States. In addition to the references on distribution given in the standard works of Tuckerman (1882, 1888) and Fink (1935), there are only scattered papers in the literature which are concerned with the lichens of these states. Tuckerman, in 1860 and 1862, reported on the collec- tions made by Fendler near Santa Fe. When these records were brought together in 1866, several collections made by Wright in New Mexico were added. Tuckerman (1878) also wrote up some Arizona collections made by the U. 5. Geological Survey of the 100th Meridian. It was not until some time later that Fink (1909, 19092) studied the lichens of the vicinity of Tucson, Arizona, from where some new species were described by Zahlbruckner (1908,'09). In 1932 Bouly de Lesdain published an account of lichen collections from the vicinity of Las Vegas, New Mexico, made by Brother Arséne Brouard, and, in 1942, on further collections from the vicinity of Santa Fe. Magnusson (1929, 1937) treated the Acarospora of Brother Агзёпе Brouard. Herre (1944, 1950) has provided additional informa- tion on the lichen flora of New Mexico from San Miguel and Sierra counties. Finally, Darrow (1950) has reported on the arboreal lichen flora of southeastern Arizona. The present paper includes forms from some new localities in Arizona and New Mexico of which three are new species and a number are additions to the flora. This compilation is the result of a study of two collections: (1) that of Dr. Robert A. Darrow made in various localities in southeastern Arizona in 1933 and 1934, and in the 1940's which includes only saxicolous and terricolous forms; (2) that made by Mr. and Mrs. Francis Elmore at Chaco Canyon National Monu- ment in northwestern New Mexico in 1952. The latter is for the most part saxicolous, but a few corticolous forms are included. Thus most of these lichens are crustaceous rock-inhabiting ones which form a conspicuous element in the vegetation of these arid mountainous regions. * Missouri Botanical Garden. (63) [Vor. 40 64 ANNALS OF THE MISSOURI BOTANICAL GARDEN The arrangement and nomenclature, wherever possible, follows that of Fink (1935). The treatment of Acarospora is in accord with that of Magnusson (1929a). А complete set of specimens is on deposit in the herbarium of the Missouri Botanical Garden. I should like to thank Dr. Carroll W. Dodge for his valuable guidance and assistance throughout this study. LIST OF SPECIES VERRUCARIA FUSCELLA (Turn.) Ach.—ARIZONA: Tucson, on caliche, 2500 ft., Dec. 4, 1933, Darrow 572; Tucson Mts., on volcanic rock, 3000 ft., March 11, 1934, Darrow 746. NEW MEXICO: Chaco Canyon, on sandstone, 6200 ft., Jan. 16 & 20, Feb. 28, 1952, Elmore. STAUROTHELE UMBRINA (Ach.) Tuck.—NEW MEXICO: Chaco Canyon, on sandstone, 6200 ft., Jan. 20, 1952, Elmore. DERMATOCARPON POLYPHYLLUM (Wulf.) Dalla Torre & Sarnth.—arizona: Pima Co., Rincon Mts., on shale, 4500 ft., March 4, 1934, Darrow 697. DERMATOCARPON RUPICOLA Zahlbr.—ARIZONA: Tucson Mts., on andesite, 3500 ft., March 15, 1934, Darrow боо. ENDOCARPON PUSILLUM Hedw.—ARIZONA: Tucson Mts., on caliche, 3000 ft., March 11, 1934, Darrow 745. RTHOPYRENIA HALODYTES (Nyl) Arn.—NEw MEXICO: Chaco Canyon, on sandstone, 6200 ft., Jan. 16, 1952, Elmore. PYRENULA NITIDA (Weig.) Ach.—Nrw MEXICO: Chaco Canyon, on grease- wood (Sarcobatus vermiculatus (Hook.) Torr.), 6200 ft., Jan. 20, 1952, Elmore. URCEOLARIA SCRUPOSA (Schreb.) Ach. var. BRYOPHILA (Ehrh.) Ach.— ARIZONA: Whitehouse Canyon, Santa Rita Mts., on Cladonia primary thalli, Dec. 24, 1933, Darrow 537. LEPTOGIUM APALACHENSE Nyl.—arizona: Pima Co., Rincon Mts., on shale, 4500 ft., March 4, 1934, Darrow 608. LEPTOGIUM BURGESSII (L.) Mont.—arizona: Whitehouse Canyon, Santa Rita Mts., on soil, Dec. 24, 1933, Darrow 533. PLACYNTHIUM MICROPHYLLIZUM (Nyl) Hasse—arizona: Tucson, on soil, 2500 ft., Nov. 26, 1933, Apr. 5, 1934, Darrow 527, 691; Pima Co., Cortaro, on soil, 3000 ft., Nov. 26, 1933, Darrow 528; Pima Co., Rincon Mts., on soil, 3500 ft., March 4, 1934, Darrow 740, 741. LECIDEA BRANDEGEI Tuck.—NEW MEXICO: Chaco Canyon, on sandstone, 6200 ft., Jan. 16, 1952, Elmore. LECIDEA PARASEMA Ach.—NEW MEXICO: Chaco Canyon, on sandstone, 6200 ft., Jan. 16, 1952, Elmore. LECIDEA PLANA (Lahm) Nyl.—new MEXICO: Chaco Canyon, on sandstone, 6200 ft., Feb. 28, 1952, Elmore. LECIDEA VORTICOSA (Floerke) Koerb.— NEw meExic о: Chaco Canyon, on sand- stone, 6200 ft., Feb. 28, 1952, Elmore. 1953] RUDOLPH—LICHENS ОЕ ARIZONA AND NEW MEXICO 65 RHIZOCARPON DISPORUM (Naeg.) Müll. Arg.—arizona: Mt. Lemmon, Santa Catalina Mts., on marble, 8500 ft., April 1, 1934, Darrow 721. ACAROSPORA AMERICANA Н. Magn.—ARIZONA: Tucson Mts., on volcanic rock, March 11, 1934, Darrow 747; NEW MEXICO: Chaco Canyon, on sandstone, 6200 ft., Feb. 28, 1952, Elmore. ACAROSPORA COLORADINA Н. Magn.—arizona: Tucson Mts., on volcanic rock, 3000 ft., March 11, 1934, Darrow 746. ACAROSPORA OXYTONA (Ach.) Mass.—NEw MEXICO: Chaco Canyon, on sand- stone, 6200 ft., Jan. 20, 1952, Elmore. ; ACAROSPORA PELTASTICA Zahlbr.—arizona: Tucson Mts., on volcanic rock, 3000 ft., March 11, 1934, Darrow 736; NEW MEXICO: Chaco Canyon, on sand- stone, 6200 ft., Jan. 20, 1952, Elmore. ACAROSPORA sTRIGATA (Nyl) Jatta—NrEw MExico: Chaco Canyon, on sand- stone, 6200 ft., Jan. 16 & 20, 1952, Elmore. ACAROSPORA TENEBRICA Н. Magn.—NEw MEXICO: Chaco Canyon, on sand- stone, 6200 ft., Feb. 28, 1952, Elmore. ACAROSPORA WASHINGTONENSIS Н. Magn.—ARIZONA: Рипа Co., Rincon Mts., on schist, 4500 ft., March 4, 1934, Darrow 766. PERTUSARIA FLAVICUNDA Tuck.—arIzoNa: Santa Rita Mts., on igneous rock, 7200 ft., Aug. 14, 1934, Darrow 820. LECANORA ATRA (Huds.) Ach.—arizona: Patagonia Mts., on granite, 5500 ft., Oct. 6, 1946, Darrow 4252. ECANORA CENISIA Ách.—ARIZONA: Santa Rita Mts., on igneous rock, 7000 ft., July 16, 1934, Darrow 796. LECANORA CONTORTA (Hoffm.) Stiz.—arizona: Mt. Lemmon, Santa Cata- lina Mts., on quartz, 8500 ft., April 1, 1934, Darrow 724. LECANORA DIFFRACTA Ach.—ARIzONA: Whitehouse Canyon, Santa Rita Mts., on granite, 6500 ft., Dec. 24, 1933, Darrow 470. LrEcANORA (ASPICILIA) elmorei E. Rud., var. nov.—TYPE: NEW MEXICO: Chaco Canyon, on sandstone, 6200 ft., Jan. 20, 1952, T. & Е. Elmore. Thallus determinatus, laevigatus, obscure olivaceus, areolatus, areolis crassis convexisque, 1-2 mm. diametro; cortex 31—41 p crassitudine, K—, fastigiatus, dimidia parte corticis crassitudine strato gelifacto tectus, hyphis circa 3 р diametro; algae ad Trebouxiam pertinentes, cellulis ad 38 p diametro, in strato non continuo, 100-170 р crassitudine sub cortice; medulla hyphis laxe contextis circa 4 p diametro. Apothecia in areolis crassioribus immersa, unum vel plura in quavis areola, amphithecio prominente, persistente, crasso, thallo concolore, disco concavo obscure olivaceo-alutaceo pulverulento; hypothecium indistinctum sed densum; thecium 150—171 р altitudine; paraphyses simplices, septatae, circa 2 p diametro, apicibus clavatis; asci 71.5-100.0 X 14.3-15.7 р, clavati, apicibus incrassatis; sporae uniseriate, 3—4-nae, sphaericae, 17.4—24.4 y diametro, granulosae, episporis tenuibus. Thallus determinate, smooth, deep olive, closely areolate, the areolae thick and convex, 1-2 mm. in diameter; cortex fastigiate, 31—41 p thick, covered by a 66 ANNALS OF THE MISSOURI BOTANICAL GARDEN BPS | 988009. GP coto o OE. 0° | > p 2999 9 Кк 8528 оъ we Vd ا‎ T RA N AS itu E = —— Y; Ee eS < c Fig. 1. Lecanora elmorei E. Rud. 1. Section of thallus. ni Enlarged section showing cortex. w . Section of an apothecium. 4. Азсиз with associated paraphyses. [Vor.- 40 1953] RUDOLPH—LICHENS OF ARIZONA AND NEW MEXICO 67 gelatinous layer about half the thickness of the cortex, KOH—, the hyphae ap- proximately 3 р thick; algae Trebouxia, forming a layer 100—170 p thick below the cortex, broken into vertically rectangular patches separated by fungous hyphae, the cells globular to slightly angular, 10-38 р in diameter; medulla of loosely woven hyphae about 4 и in diameter. Apothecia sunken, the margin formed of the remaining part of the aerolae, one to several per areolae, the disc very concave, dark olive-buff, appearing powdery; margin persistent, raised and thick, thalline, colored like the thallus; epithecium of gelatinous tips of paraphyses; hypothecium indistinct, slightly denser than the underlying tissues; thecium 150—171 p thick, the paraphyses simple, filiform, septate, about 2 д thick, the tips swollen. Asci clavate, 71.5-100.0 X 14.3-15.7 и wide, hyaline tips present in many; spores uniseriate, 3—4 per ascus, globose, 17.4—24.4 и in diameter, mostly coarsely granu- lar, the epispores thin. This new species is easily distinguished by its 3—4 globose spores per ascus and by having Trebouxia as its alga. The globose-spored species Lecanora praecrenata Nyl, with its diffuse, whitish, indeterminate thallus and sessile brown apothecia, contrasts easily with the thick, closely areolate, deep-olive thallus and sunken apothecia of the present species. LECANORA EPULOTICA (Ach.) Leighton—arizona: Santa Rita Mts., on ig- neous rock, 6500 ft., July 1, 1934, Darrow 614. LECANORA FRUSTULOSA (Dicks.) Ach.—arizona: Carr Canyon, Huachuca Mts., on marble, 5600 ft., June 12, 1945, Darrow 4236. LECANORA MELAENA (Hedlung) Fink—arizona: Tucson Mts., on calcareous rock, March 25, 1934, Darrow 682. LECANORA PARISENSIS Му|.— МЕХ МЕХ!СО: Chaco Canyon, on greasewood (Sarcobatus vermiculatus (Hook.) Torr.), 6200 ft., Feb. 28, 1952, Elmore. LECANORA POLYTROPA (Ehrh.) Rabh.—arizona: Santa Rita Mts., Pima te a оп igneous rock, 4500 ft., Aug. 14, 1934, Darrow 810. LECANORA THAMNOPLACA Tuck.—ARIZONA: Santa Кіса Mts., on igneous rock, 7000 ft., Aug. 14, 1934, Darrow 822; NEW MEXICO: Chaco Canyon, on sandstone, 6200 ft., Jan. 16, 1952, Elmore. PARMELIA Novo-MEXICANA Gyeln.—NEw MEXICO: Chaco Canyon, on sand- stone, 6200 ft., Feb. 28, 1952, Elmore. CaLoPLAcA arizonica E. Rud. sp. nov.—TYPE: ARIZONA: Cortaro, оп rhyo- lite, 3000 ft., Nov. 26, 1933, В. A. Darrow 521. Thallus areolatus, fuscus; areolae 0.5—1.0 mm., dispersae; cortex circa 18 po fastigiatus, strato gelifacto tectus, K obscure purpurascens; algae protococcoideae ad 10 р diametro, cellulis sphaericis vel subangulosis, in strato 90—130 p crassitudine sub cortice; medulla hyphis laxe intertextis crystallis saxi impleta. Apothecia 0.2— 0.6 mm. diametro, orbiculata, disco ferrigineo-aurantiaco, amphithecio distincto, Persistente, tenui, thallo concolore; epithecium crystallis runneo-luteis inspersum; hypothecium hyalinum, centro circa 70 и altitudine; thecium 63—87 р altitudine; Paraphyses filiformes, 1.7 р diametro, simplices, septatae; asci 24—42 X 7-11 p, [Vor. 40 Xx e WES SS ee Se eT om APA y; НИ A AMET n a N| VEU 1 MN. ^ Те СА d и (и М ub KU OYE ә КОЖО ж d T 9 ak. -T - sem Fig. 2. Caloplaca arizonica E. Rud. 1. Section of thallus. Enlarged section showing cortex. Section of an apothecium. Ascus with associated paraphyses. “ ЖУ и Ascospores, 1953] RUDOLPH—LICHENS ОЕ ARIZONA AND NEW MEXICO 69 clavati; sporae 8-nae, 8.7—12.2 X 3.5-5.2 y, polari-biloculares, isthmo circa 3 p longitudine, uni- vel biseriales. Thallus indeterminate, aerolate, fuscous, the areolae 0.5—1.0 mm., often quite far apart; cortex about 18 y thick, with a gelatinous covering, fastigiate, KOH dark purple; algae protococcoid, forming a layer 90-130 p thick below the cortex, the cells round to angular, 5-10 и in diameter; medulla of loosely woven hyphae about 3 p thick, containing numerous small rock crystals. Apothecia circular, 0.2—0.6 mm., disc brownish-orange, the margin persistent, thin, of the same color as the thallus; algae in the margin and in a layer below the hypothecium; epithecium incrusted with brownish-yellow crystals, about 13 u thick; hypothecium hyaline, about 70 p thick at center; thecium 63—87 p high; paraphyses filiform, septate, about 1.7 р thick, unbranched. Asci clavate, 24—42 X 7-11 и; spores 8, uni- or biseriate, polar-bilocular, 8.7—12.2 X 3.5-5.2 p, the isthmus about 3 y long. This species somewhat resembles Caloplaca cinnabarina (Ach.) Zahlbr., but can be distinguished by its relatively fewer and more dispersed apothecia, darker thallus, and narrower spores. CALOPLACA CERINA (Ehrh.) T. Fries—NEw MEXICO: Chaco Canyon, on sand- stone, 6200 ft., Jan. 16, 1952, Elmore. CALOPLACA CINNABARINA (Ach.) Zahlbr.—arizona: Whitehouse Canyon, Santa Rita Mts., on granite, 6500 ft., Dec. 24, 1933, Darrow 471. CALOPLACA ELEGANS (Link) T. Fries—NEWw MEXICO: Chaco Canyon, оп sand- stone, 6200 ft., Jan. 16 & 20, Feb. 28, 1952, Elmore. CALOPLACA Festiva (Ach.) Zwackh.—arizona: Miller Canyon, Huachuca Mts., on igneous rock, 6300 ft., June 14, 1945, Darrow 4243. CALOPLACA FULGENS (Swartz) Koerb.—arizona: Nogales, on calcareous rock, 4000 ft., Oct. 6, 1946, Darrow 4261; NEW MEXICO: Chaco Canyon, on sandstone, 6200 ft., Feb. 28, 1952, Elmore. CALOPLACA LOBULATA (Floerke) Hellb.—arizona: Рипа Co., Rincon Mts., on granite, 4000 ft., March 4, 1934, Darrow 785. CALOPLACA MODESTA (Zahlbr.) Fink—arizona: Tucson Mts., on calcareous rock, 3000 ft., March 19, 1934, Darrow 730; Pima Co., Coyote Mts., on granite, 3200 ft., Feb. 16, 1945, Darrow 4266. TELOSCHISTES PARIETINUS (L.) Norm.—NEW MEXICO: Chaco Canyon, оп sandstone, 6200 ft., Jan. 16, 1952, Elmore. BUELLIA BLUMERI Zahlbr.—arizona: Whitehouse Canyon, Santa Rita Mts., on igneous rock, 6500 ft., Dec. 24, 1933, Darrow 477. BUELLIA RETROVERTENS Tuck.—aARizoNA: Whitehouse Canyon, Santa Rita Mts., on granite, 6500 ft., Dec. 24, 1933, Darrow 489; Tucson Mts., on volcanic rock, 3000 ft., March 19, 1934, Darrow 736; NEw MEXICO: Chaco Canyon, on sandstone, 6200 ft., Feb. 28, 1952, Elmore. Rinopina darrovii E. Rud. sp. nov.—TYPE: ARIZONA: Santa Catalina Mts., on ground and moss, Nov. 12, 1933, R. A. Darrow 498; PARATYPE: Santa Cata- lina Mts., on soil, 8000 ft., Nov. 12, 1933, Darrow 503. ANNALS OF THE MISSOURI BOTANICAL GARDEN Fig. 3. Rinodina darrovii E. Rud. Section of thallus. — . ого» Enlarged section showing cortex. Section of an apothecium. Ascus with associated paraphyses. [Vor. 4 1953] RUDOLPH—LICHENS OF ARIZONA AND NEW MEXICO 71 Thallus ochraceus vel cinnabarinus, granulosus, K—, diffractus; cortex geli- factus, erodus; algae protococcoideae, cellulis sphaericis vel angulosis, 5—10 p diametro, in strato continuo 57-71 p crassitudine sub cortice; medulla densa, hyphis indistinctis circa 1.5 u diametro intertextis. Apothecia nigra, 0.5-1.0 mm. diametro, plana vel convexa; margine thalloideo, tenui, mox emarginato, thallo concolore; hypothecium hyalinum, in centro 80 џи altitudine; thecium 27-34 p altitudine; paraphyses 2 и diametro, septatae, simplices; asci 70-71 X 10-16 p, clavati; sporae 8-nae, biseriatae, 17.4—24.4 X 8.7-11.5 p, primo non-septatae, deinde tri-septatae, brunneae. Thallus ochraceous-tawny to cinnamon-brown, granulose, broken in places into irregular pieces, KOH-—; cortex gelatinous, eroded in parts; algae proto- coccoid, forming a continuous layer 57-71 р thick, the cells spherical to angular, 5—10 р in diameter; medulla of tightly woven, indistinct hyphae about 1.5 p in diameter. Apothecia black, 0.5—1.0 mm. in diameter; disc flat to convex; margin thalloid, thin, of same color as the thallus, soon disappearing; hypothecium hyaline, of closely woven hyphae, about 80 д thick at center; thecium 27—34 р high; paraphyses filiform, septate, about 2 p thick, swollen at tips and incrusted with yellow crystals, unbranched; asci clavate, 70—71 Ж 10—16 р; spores 8, biseriate, brown, at first 1-celled becoming 4-celled, 17.4—24.4 X 8.7-11.5 p This species has the general appearance of Rinodina оныг Dni, ) Vainio (R. nimbosa (E. Fr.) T. Fr.) but the four-celled spore places it in the subsection coNRADIA Malme as recognized by Zahlbruckner (1926). Its affinities seem to be with В. conradi Koerb. from which it can be separated by its larger black apothecia, its indistinct cortical hyphae, and its thicker hypothecium. The spores in this species have not been found to have more than four cells and also to be slightly smaller than those of R. conradi. ODINA EURYSPORA Zahlbr.—aAmizoNA: Santa Rita Mts., on igneous rock, 7000 ft., Aug. 14, 1934, Darrow 821. ODINA NOVOMEXICANA B. de Lesd.—ARIZONA: Santa Rita Mts., оп igneous rock, 4500 ft., June 5, 1934, Darrow 707; 7000 ft., Aug. 14, 1934, Darrow 823. BIBLIOGRAPHY — d iMac M (1932). topa de зи de New-Mexico (U.S.A.) recueillis par le Frére С. “т ids een У. na. Su im ent. da t Lichenol. 12:44—66. e nA . (1950). niim wd А fora of southeastern Arizona. Amer. Midl. Nat. Fink, Bı Mn EE и aati са Si The — ad of a desert кг -— y с "eti Distribution and movements of desert plants. Санга» Inst. Washington Publ изи. 3 EN The Lichen flora e the United States. Univ. ме an 426 he (1950). New Lichens from California, New Mexico and the Philippines. Bryologist 53:296-299, [Vor. 40, 1953] 72 ANNALS OF THE MISSOURI BOTANICAL GARDE кыл ча. AHO veo The Lichen genus Асатозротв in New Mexico. Meddelanden Gótesborgs tT. ved -72. Vien a ч Janoti of the genus Acarospora. Kungl, Svenska Vetenskapsakademiens $ Az , (193 ‘dona notes on ын и экы ЫА 1. Species from New Mexico. Meddelanden 2:87-9 7). Götesborgs Bot. Trad M Бар (1860). Den cial on North American and some other Lichenes, 1. Amer. cad. A 7007. з & Sci. Proc. 4 4 (1862) Part П. Ibid. 5:383—4 ‚ (1866). ie of California, cecus and the Rocky Mountains; as far as yet known. em › (18 25. | йай. In Rothrock's Report upon the botanical vore made in portions of Nevada; oni California, Colorado, New Mexico, and Arizona. Geol. Surv. W. 100d Meridian 6:3 > К ins) A synopsis of the North American lichens; Part I. 262 pp. Boston; art П. 176 pp. New Bedford, Zahlbruckner, Alerander (1808). ay меа American сакен Bull, Torr. Bot. Club 35:297-300. —————, (19 eue Flechte Mycol. 7:472-478. ————, а. Lichenes (Pacht): gue se Teil. ‘Ta Engler & Prantl’s Nat. Pflanzenfam. 8:61-270. THE ANALYSIS OF SUSPECTED HYBRIDS, AS ILLUSTRATED BY BERBERIS Х GLADWYNENSIS EDGAR ANDERSON During the last fifteen years a considerable portion of my time has been spent in attempting to measure the importance of hybridization in the evolution of natural populations. More than a score of genera have been studied intensively in the wild (and whenever feasible in the breeding plot as well) with students and associates, though as yet only a fraction of these studies has been published. Various techniques have been worked out for dealing with such problems (Ander- son, 1949, chap. 6), and as their validity has been more widely recognized (Stebbins, 1952) they are being more widely adopted. In the course of helping workers elsewhere to adapt these methods to their own problems I have become increasingly aware that in part they are based upon a scrutiny of the original material more exhaustive than is customary in many laboratories. It has accord- ingly seemed expedient to describe the procedure in more elementary detail than hitherto. Instead of discussing the complexities of hybridization in natural popu- lations I have chosen the relatively simple example of an apparent hybrid which arose spontaneously in the garden of Mrs. J. Norman Henry, of Gladwyne, Penn- sylvania. This seems a particularly happy choice since it also serves to demonstrate the usefulness of these methods in dealing with horticultural material of unknown ancestry. А vigorous young barberry was found coming up in the shelter of a large bush of Berberis verruculosa in Mrs. Henry's garden. Berberis verruculosa is a dense, evergreen-leaved plant, certainly the most distinctive of the hardy barberries. The seedling, so obviously related that it must have been a seedling of B. verruculosa, was nevertheless so different from it that Mrs. Henry had supposed it was prob- ably a spontaneous hybrid. It was different from any barberry known to her or to me and it exhibited the hybrid vigor which is the mark of so many hybrids. Furthermore, many garden hybrids of Berberis are known and Mrs. Henry at Various times had grown a number of other species of Berberis in her garden. With er permission a precise examination was made both of the putative hybrid and of Berberis verruculosa to establish the probable ancestry. The following is the procedure established in this laboratory for such examina- tions: l. Choose comparable material of parent and hybrid. 2. Examine, describe, and measure each feature item by item. Use great Care to work with truly comparable material. Do not attempt comparisons of a branch grown in the shade from one specimen and one grown in the sun from the other. If one is from a fruiting branch, then the other should be also. Use great саге in distinguishing between long shoots and short shoots. Many plants without such conspicuously heterogeneous shoot systems as the Ginkgo have a more or less well-defined short shoot-long shoot system which requires careful examination to Perceive, (73) [Vor. 40 74 ANNALS OF THE MISSOURI BOTANICAL GARDEN INTERNODE LENGTH S № о А e = n L T T T , 3 4 5 6 1 '8 9 NUMBER OF SUCCESSIVE INTERNODES - Fig. 1. Internode diagrams of comparable branchlets of two species of barberry and their putative hybrid. Solid line, Berberis verruculosa; long dashes, Berberis X gladwynensis; short dashes, Berberi апае. Аз explained in Anderson А made by plotting the lengths of successive internodes from the base to the tip an ^ i i im гу line. The first internode at the base of the branchlet in В. Julianae, tc. e o Е 5 e = 5 > =. 8 ts wi for instance, is 45 mm. long, the second is 28 mm., the third, 33 mm., e 3. (And, most important): Conduct the entire examination, if possible, against a plain neutral background. Cover the table top with clean wrapping Paper or some such material. Remove all distracting objects such as pencils, paper clips, erasers, etc., out of the immediate field of vision. 'The eye can work much more efficiently in the perception of resemblances and differences if it can con- centrate upon the problem in hand. Work in a good light and have a good dis- 1953] ANDERSON —HYBRID ANALYSIS IN BERBERIS 75 cussion, and yet in assisting other people to use my methods I nearly always find that these precautions have not been followed and that it is difficult to convince other scholars of their supreme importance. This is one of those ridiculously simple matters which are far more important than they seem and which are indicative of true scientific precision. Specific differences are frequently subtle and they can be apprehended much more readily if one uses comparable material, a neutral background, and a good light. 4. Make measurements with good steel calipers and a steel metric rule. 5. Pay particular attention to pubescence, pubescence pattern, internode pattern (Anderson and Schregardus, 1944), branching pattern. Careful analysis will show that nearly all closely related species of the higher plants differ sig- nificantly (and with enough study, definably) in their internode patterns. Comparable branches were accordingly chosen from Berberis verruculosa and the seedling. Both were selected from well-developed branches of flowering age, growing in almost full sun and borne upon the upper parts of their respective. plants. After these two branches had been examined carefully against a neutral background, comparable branchlets were chosen for further study as follows: Proceeding from the tip, the first strong secondary branch (all of the current season's growth) was selected for study. The internode lengths of this branchlet were measured to the nearest millimeter, the results being plotted as an internode diagram (fig. 1). Hybrids are usually intermediate between their two parents when due allow- ances are made for heterosis, for growth-pattern differences, and for possible dif- ferences in ploidy. From a study of fig. 1, it is possible to make several predictions as to the putative male parent of the seedling barberry. B. verruculosa is shown to have more and shorter internodes than the seedling. Its basal internode is only about twice as long as the other internodes and the remaining internodes are sub- equal. The seedling has a basal internode more than twice as long as the others and the remaining internodes are not at all uniform in length and seem to have a definite pattern of decreasing length toward the tip. We therefore predict that the putative parent (when comparable material is studied) should have conspicu- ously non-uniform internode lengths with a very long internode at the base of such branches and decreasingly short ones as the tip is approached, probably in a well-defined pattern. a Leaf comparisons were then made. Choosing comparable leaves proved to be difficult but not impossible. Leaf size and shape were uniform in B. verruculosa but variable in the seedling. Care had to be taken, therefore, to choose exactly comparable material. Most of the primary leaves in these barberries are spines. The leafy leaves are borne on short secondary branches in the axils of the spine- leaves, giving the appearance of little clusters up and down the branches. In B. verruculosa all the leaves on the plant were superficially similar in size, shape, and serration. In the seedling they differed markedly. Those in clusters arising from the older wood were variable. Among them were leaves which were narrower ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 40 PORE M 6.5 ; | 5.0 i | H | i | 5.5, I | : | 5.0] I j CF = Н = } Н / .0 H : 4.0] / “= Н Ё 3 H ; ^ 3.5. ; д Н / Н / |! / ^l Ly Н i № H H 2.5]; Н 2.0 3 ` 1.0 1.1 ; 1.3 ecu LEAF WIDTH CM. Relationships of size, shape, an of marginal setae in leaves from young and ies of estu their duin. hybrid in the previous figure, 2 hes, Ber x , h, Fig. 2. mature spurs for two spec solid line indi : young to ad. Ai rs hang matic, e. Th { size and shape are all to the same scale Perge represent ы values beris X gla | and seta number of e lines ien indicates the change pe and seta number Berberis ана ны es conspicuously i in size, $ hybrid is intermediate in degree of ; their ed t to the diagrams, while highly diga from actual measurem 1953] ANDERSON—HYBRID ANALYSIS IN BERBERIS 77 in shape, larger in size and with more strongly developed setae at the tips of the serrations along their edges. Careful examination then showed that B. verruculosa had a very slight tendency in the same direction though not strong enough to be readily apparent without examination. Measurements showed that while Berberis verruculosa and the seedling had about the same number of setae per leaf on the new wood, the seedling had nearly twice as many as verruculosa on the old wood. Figure 2 shows all these relationships. The putative parent for which we are searching should have leaves on its new wood as small as those of verruculosa, or even smaller, with greatly differentiated leaves on the older wood. Some of the leaves on the older wood should be very long, proportionately narrow, and with many setae along their edges. As, in this fashion, one proceeds feature by feature in the careful comparison of the one parent with the hybrid, he becomes progressively better acquainted with the general ground plan of the plant and is increasingly capable of making precise comparisons and valid inferences. Before long it is possible to draw up precise descriptions not only of B. verruculosa and its hybrid seedling but also of the putative male parent. Such a comparison is presented below in tabular form, including the predictions as to the male parent. verruculosa glad wynensis unknown (Julianae) Branches arching, Branches sub-arching, Branches straight, sub-horizontal some nearly erect erect to horizontal боса, short, Internodes medium, Internodes long, uni variable M Branchlets densely Branchlets E most Branchlets eglandular ew dular sub-glan ew growth scarcely New growth ча брон strongly Prem ated differentiated (bo iated Mature bark Mature Gá Mature Pan strongly barely ribbed definitely ribbed ribbed and furrowed Wood greenish Wood yellowish-green Wood bright yellow-green Spines up to 1.2 cm Spines up to 1.3 . Spines up to 1.5 cm. Leaves glossy Leaves somewhat glossy, Leaves not glossy, dark green above, -dark above, bright green above, glaucous bel sub-glaucous below light green bel Fall color a more or Fall color stronger on Brilliant fall color less general purple newest growth, occasional eaves о and d growth reels dior: ted Largest send on old spurs 7 mm. X 12 mm. 10 marginal setae at most Flowers borne singly in pairs Stigma sessile розаны н кы RN, Se DRIN T RING "o owth definitely аа Largest leaves on old spurs 43 mm. X 14 mm 20 ае setae at most Flowers in fascicles of 5 to 8 Stigma sub-sessile Leaves > new and old growt differen ciated Largest lave on ан spurs 65 mm. X 11 mm Over 30 ыа, setae Flowers in fascicles of 15 to 25 Stigma definitely not sessile, style of at least 1 mm. [Vor. 40, 1953] 78 ANNALS OF THE MISSOURI BOTANICAL GARDEN As the description began to resolve itself my knowledge of barberries was suf- cient to suggest that Berberis Julianae was probably the barberry we were looking or. As soon as I made this suggestion Mrs. Henry informed me that a large bush of this species had once stood just a few feet away from B. verruculosa but it had been damaged in a windstorm and removed. Many of the technical points in the description, however, involved matters which were completely outside my knowl- edge of barberries. Predictions as to the numbers of setae on the mature leaves, the number of flowers per fascicle, the presence of a style, etc., were drawn up with no knowledge of what they might be like in B. Julianae. The entire hypothetical description was then run down in Rehder's key (Rehder, 1940) to the barberries just as if one had the specimen actually in his hand. It led to B. Julianae and fitted the description of that species precisely, down to the most technical detail. There can be little doubt, therefore, that the seedling was a spontaneous hybrid, B. verruculosa X B. Julianae. I am accordingly describing it as: BERBERIS X gladwynensis hyb. nov. Intermediate between its parents B. verruculosa and B. Julianae. Flowers in fascicles of 5 to 8, stigma subsessile, leaves subglaucous below, nearly evergreen. Type: Henry, in the herbarium of the Philadelphia Academy of Natural Sciences; from the type bush, E. Anderson 5. 9., It is a ens to name this handsome barberry after the site of Mrs. Henry's garden which has long been a mecca to botanists and gardeners alike. LITERATURE CITED Anderson, е рн у: MSS aene Hybridization. = Wiley & Sons, Inc. New ا‎ d ork. پا‎ — 3 1944). : os thod for recording and analyzing ай of inter- node t. Gard. а Rehder, А "a 1940). ЫТ? кө of "Cultiv vat id pus | Shrubs, The Macmillan Co. York. New Stebbins, G. Г. (1952). The evolution of совет plants and weeds. рае 6:445—448. 1 BERBERIS быку nsis hyb. nov. inter parentes В culos, Flores in fasciculis 5-8, stigmate subsessile. Folia fere = » E Јанета, verruculosam semperviridia, subter subglauc VARIATION IN СОВ MORPHOLOGY AMONG CERTAIN ARCHAEOLOGICAL AND ETHNOLOGICAL RACES OF MAIZE NORTON Н. NICKERSON* INTRODUCTION Zea Mays L., one of the most highly evolved of all grasses, is still botanically much of an enigma. Commonly known as Indian corn or maize, it was a very important plant to the peoples of the New World long before the arrival of Columbus. It formed the basis of the highly developed Inca, Maya, and Aztec civilizations and was the staple crop from Canada to Chile for several thousand years. It is, of course, quite important to the present-day inhabitants of Central and South America, and agriculture and industry in the United States uses three and one-quarter billion bushels annually. Botanists have until recently followed almost exclusively Sturtevant's (1899) classification of contemporary races of maize by kernel types. Although this classification is an artificial one, it has proven to be of practical commercial worth. Archaeologists have long known the value of maize kernels found at excavation sites for determining types of maize grown and the uses to which they were put. Until the last few years, other parts of the maize plant have been quite generally neglected as a source of historical data, chiefly because of the complexities involved in determining and evaluating such evidence as they contain. Evidence from maize tassels has been used with promising results (Alava, 1952; Anderson, 1944b, 1944c, 1949a, 1951; Anderson and Brown, 1948; Anderson and Cutler, 1942; Brown ef al, 1952; Cutler, 1946; Wellhausen e£ al, 1951, 1952). Prat (1948) showed that in maize and other grasses hairs and other epidermal emergences can be used as a basis for identification and classification. Internode patterns have been studied by several workers (Anderson, 1943, 1949а; Anderson and Brown, 1948; Anderson and Schregardus, 1944; Stonor and Anderson, 1949; Wellhausen e£ al, 1951, 1952). Ear and tassel ontogeny have been studied by Bonnett (1940, 1948) and Kiesselbach (1949). Esau studied the ontogeny of the maize vascular bundle. Surprisingly little has been done to measure and evaluate those morphological structures which are present on a maize cob after the kernels have been removed. Weatherwax, a pioneer in the study of the maize cob, pointed out (1918) the need and value of accurate morphology in understanding this structure, and dem- onstrated (1920) the orderly spikelet behavior underlying changes in kernel row- ing. Fujita (1939) observed that an even number of pairs of kernels resulted in straight rows, and an odd number in spiral rows. Cutler (unpublished work) made many histological studies of the cobs of both North and South American maize, as well as of several closely related genera. Lenz (1948) indicated the types eer "Special Assistant in Botany, Henry Shaw School of Botany, Washington University. The stimulating advice of Dr. Edgar Anderson and the technical erai of Mr. Ding Hou throughout the course of this investigation are hereby gratefully acknowledge [Vor. 40 80 ANNALS OF THE MISSOURI BOTANICAL GARDEN of evidence available from a histological study of the maize cob. Alava (unpub- lished work) modeled in great detail small portions of the exterior surface of a maize cob. Mangelsdorf and Smith (1949) tabulated a number of external and internal maize ear characters, and their general procedure was used in character- izing 25 present-day races of maize in Mexico by Wellhausen e£ al (1951, 1952). These studies were based on the evaluation of many characters which did not in- volve exact measurements. In the present study, the value of several morphological characters used by these investigators has been increased by employing exact measurements. Other measureable morphological characters have been which are of significance in tracing historical influences of one race of maize upon another. : There are two reasons why evidence present on the maize cob is important. The first one is based on the generally accepted fact that distinguishing characters are not distributed at random among plants; with regard to their occurrence in the Maydeae, agrostologists have found the female inflorescence to be of particular significance. The second reason, already pointed out by Lenz (1948), is that there are more archaeological remains of maize cobs than of any other plant material. The importance of these remains can perhaps be appreciated when one considers that maize, unlike most cultivated crops, is completely dependent upon man for its preservation. This fact can only mean that it has had a constant associatio with man since its adoption as a food plant, and that in its manner of origin, in the course of its migrations, in its development into a myriad of races, and in its intimate association with ancient religious symbolism, the history of maize becomes intimately tied to the history of man. Increased insight into the one will certainly add to our present understanding of the other. For a reasonably complete investigation of the maize cob, the following three courses were deemed necessary: (1) to measure certain easily-recognized features of the maize cob and to interpret their morphological nature; (2) to set forth one possible scheme of analysis of the variation patterns of different races of maize based on these measurements; (3) to discuss the resulting indicated trends as to their validity and applicability in extending our present understanding of the history of maize. MORPHOLOGY oF THE EAR Bearing in mind that an adequate treatment of variation in maize ears would be more complete if the nature of those parts under discussion were understood, 2 detailed study of morphological structures found on the maize ear was first under- taken. However, since the results are of a technical nature quite different from those enum from a study of variation, they are treated elsewhere (Nickerson, in press). 1953] NICKERSON—VARIATION IN СОВ MORPHOLOGY 81 MORPHOLOGICAL STRUCTURES EXAMINED One of the purposes of this study in variation is to determine what morpho- logical structures expressed in the cob are of significance in differentiating between races of maize. To accomplish this, several characters had to be included which ultimately proved to be of little use. A total of 526 cobs was examined for each of the three external and five internal characters discussed below. EXTERNAL CHARACTERS Shape of Ear.—This character is the only one used in the investigations which is dependent upon a subjective grading. Ears were classified into one of four types: straight, cigar-shaped, tapered, enlarged butt. Two of these types—straight and tapered— were used in analysis of each sample. For reasons discussed else- where (Nickerson, in press), a straight-eared race exhibits less condensation than a tapered- or enlarged butt-eared race, and a race with cigar-shaped ears is inter- mediate between these extremes. Shank Diameter.—Yhe diameter of the shank was measured in millimeters at a point close to the base of the ear, above the last apparent husk node whenever possible. In specimens with elliptical shanks, two measurements were made and an average of these was used. Shank diameters are fairly consistent within any one kind of maize and are often markedly divergent between different kinds. Row Number.—Row number was determined by counting the number of vertical rows of glumes in the middle area of the cob. The middle area was used for this and all subsequent measurements because of its uniformity of size, stage of development, and freedom from growth irregularities common at both base and tip of the ear. Row number is a readily-observed character the importance of which is not yet fully understood. Anderson and Brown (1948) were able to correlate the condensation index of the tassel with kernel row number in the ear. Cutler (1952) used row number in his preliminary study of cobs from successive layers of cultural remains found in Tularosa Cave. His tables indicated the pres- ence of high row numbers in lower strata and progressively lower row numbers in successively higher strata. These data show the reverse of the situation reported for the stratified remains of Bat Cave by Mangelsdorf and Smith (1949), who stated that in general the older (and more pod-corn-like) the ear, the lower was its row number. INTERNAL CHARACTERS To study internal characters, the ear was broken in two at the approximate middle; the exposed ends were then observed under a dissecting microscope. Width of Lower Glume.—This character was measured in millimeters with the calipers held perpendicular to the widest point on the abaxial surface of the glume. measurement is not a direct expression of a particular gene such as Mangels- dorf and Smith (1949) reported for the relationship between lower glume length and various alleles of the Tu gene, but reflects the width of the basal portion of the kernel which it encloses. [Vor. 40 82 ANNALS OF THE MISSOURI BOTANICAL GARDEN Height of Rachis-flaps——This measurement in millimeters was made perpen- dicular to the axis of the cob. In instances where the rachis-flaps were of unequal height, an average measurement was used. The rachis-flap is a part of the cupule (Nickerson, in press) and its significance is somewhat minimized by its variability I in some otherwse homogeneous samples used in the present investigation. Mangels- dorf and Smith (1949) reported the same situation for maize cobs from Bat Cave, but Wellhausen e? al (1951, 1952) found rachis-flaps to be extremely useful in characterizing races of maize in Mexico. Deptb of Cupule.—This measurement in millimeters was made perpendicular to the cob axis, and was given a positive or a negative value, depending upon whether it was out from, or down into the rachis. It represents the extent of adnation of the cupule-forming prophyll (Nickerson, in presa) to the axis, and is well correlated with ear shape. Width of Cupule—This measurement was made by holding the calipers at right angles to the cob surface and measuring the distance in millimeters between iN See ри MN dnm A A NORTHERN FLINT 1 HUACA PRIETA HEIGHT OF RACHIS . FLAPS N DEPTH OF CUPULE - POSITIVE 3 DEPTH OF CUPULE - NEGATIVE > WIDTH OF CUPULE 5 KERNEL THICKNESS : Text-fig. 1. Drawings of two representative maize races ques distances in deter- mining four internal cob characters. А, transverse section of c have bes ч ch a transverse cut was m. i i n ings are to scale; arrows at 2 and 3 are each 1953] NICKERSON—VARIATION IN COB MORPHOLOGY 83 any pair of outside rachis-flap edges across their cupule. It is in part an indication of kernel width, in that wide cupules always mean wide kernels. Thickness of Kernel.—The distance along the cob axis from the base of one set of lower glumes to the upper edge of their cupule is exactly equivalent to the thickness of each kernel produced there. The measurement is useful in making comparisons between productivity of different cob samples, especially those in- volving fragmentary archaeological remains. These last four characters were recorded in outline drawings carefully made to scale (working with calipers under a dissecting microscope) for each of the 526 cobs examined (text-fig. 1). Other characters measured but not used include height and texture of the lower glume and longitudinal profile drawings of each cupule. Diameters of the pith, cob, and rachis were measured and the cob/ rachis index (Mangelsdorf and Smith, 1949), as well as the rachis/pith index, was com- puted for all specimens. However, neither the averages nor the individual vari- ations of these indices were found to be particularly characteristic among the maize samples studied in this investigation. NAMES, SOURCES, AND ANTHROPOLOGICAL BACKGROUND OF SAMPLES Maize samples examined were of diverse origin both geographically and temporally. Races of contemporary maize here included were chosen as typical for those parts of the world. Archaeological cob samples were largely from the southwestern United States, but included one sample from Lower California and two extensive collections from sites excavated by Junius Bird in Peru and northern Chile. In listing the samples they are grouped arbitrarily into geographical units; no further inferences should be drawn from this arrangement. Samples, except where otherwise noted, are from the maize collection of the Museum of Economic Botany of the Missouri Botanical Garden, St. Louis. NORTH AMERICA PRESENT-DAY RACES: Iroquois Sacred Flour.—This sample was from material originally grown by Professor Frank P. Bussell from seed collected by Erl Bates from the Iroquois Indian tribes of northern New York. The Indians grew it under conditions of Strict isolation for ceremonial use (Anderson, 1947b). Ears are straight, 20—28 cm. long, with heavy shanks and 8 rows of wide white or yellow (sometimes pale blue) kernels. Fourteen ears were examined. Northern Flint.—A composite sample, this group included flint corns from the Northeast (Parker's, Canada, Mammoth Yellow, Longfellow) and the - Great Plains (Winnebago, Bear Island, Brownell, Mandan, Tama, Golden Chip- Pewa). Brown and Anderson (1947) studied these and other representative flint Varieties and reported that the ears were long and slender with 8—10 rows of wide, crescent-shaped kernels, heavy shanks, and a tendency toward a large cob base. wenty-three ears were examin [Vor. 40 84 ANNALS OF THE MISSOURI BOTANICAL GARDEN Hopi White Flour.—This sample was furnished by Dr. William L. Brown, Pioneer Hi-Bred Corn Co., Johnston, Iowa, who collected it at the Hopi town of Hotevilla, Arizona. Carter and Anderson (1945) considered Hopi White Flour to be an example of Puebloan maize. Brown еЁ al (1952) reported that this variety is one of their main sources of meal. The straight ears are 18—24 cm. long, with wide shanks, no basal compression, and wide white kernels in 12—14 rows. Seven ears were examined. Hopi Blue Flour.—Also furnished by Dr. William L. Brown, this collection has substantially the same history as Hopi White Flour. However, Brown ef al (1952) reported that it is probably more important as a source of food. e cigar-shaped ears are 14—20 cm. long, with small shanks, a tendency toward basal compression, and 10—20 rows of narrow blue-aleuroned kernels with white endo- sperm. Eleven ears were examined. : Papago Flour.—This race of maize is grown by the desert-dwelling Papago Indians, and was reported by Carter and Anderson (1945) to be quite similar to prehistoric Basketmaker maize. 'The ground kernels make a meal of excellent quality. Ears are 20-25 cm. long, cigar-shaped, with narrow shanks and 12-14 rows of isodiametric yellow kernels with white endosperm. Twenty-five ears were examined. ARCHAEOLOGICAL REMAINS: Basketmaker.—Cobs included in this collection were recovered from Kinboko and White Dog Caves, both of which are located in Tsegi Canyon, Marsh Pass, northeastern Arizona. The samples were submitted by Mr. Harold S. Gladwin, who estimated their age to be circa 200-300 A.D. On the basis of there being no pottery associated with these cobs, they are believed to belong to the cultural level known as Basketmaker П. The cigar-shaped cobs are 6—12 cm. long, and bore somewhat isodiametric kernels in an average row number of 14. Six cobs were examined. Marsh Pass.—Cobs included in this collection were recovered in West Hack- berry Canyon and Turkey Cave, both of which are located in Tsegi Canyon, Marsh Pass, northeastern Arizona. These samples were also submitted by Mr. Н. S. Gladwin, who estimated their age to be from 300 A.D. to about 1000 A.D. The collection represents cultural levels from late Basketmaker П or later up through Pueblo П. The predominantly cigar-shaped cobs are 6—14 cm. long, and bore kernels wider than thick in an average row number of 10. Thirty-one cobs were examined. Turner Site.—This sample consisted of 50 cobs and fragments recovered from the Turner Site, near Cisco, Utah, by H. M. Wormington, Curator of Archaeology, Denver Museum of Natural History. Wormington estimates its age to be around 1000 A.D. (personal communication, 1953), and believes that the site was in- habited by peoples who were culturally a later manifestation of the Fremont Basketmakers. Remains of their culture were recovered from Castle Park in the 1953] NICKERSON— VARIATION IN COB MORPHOLOGY 85 Yampa River Valley, Utah, and are dated at 400-800 A.D. by Burgh and Scoggin (1948). There are nearly equal numbers of tapered and cigar-shaped cobs in this collection; some were not used in this study because of their fragmentary nature. Cobs are short, 3—8 cm. long, charred, with medium shank diameters and medium- wide kernels in an average number of 12 rows. Twenty-seven cobs were examined. Luster Cave.—Ten cobs were recovered by Dr. Robert Lister, Department of Anthropology, University of Colorado, Boulder, from a cave located in Utah, just west of the Colorado state line, in the Glade Park area. A preliminary report (Lister and Dick, 1952) described the archaeological finds of this cave and of other nearby sites. On the basis of pottery types, the maximum possible age assigned to Luster Cave by these workers is 900-1000 A.D., but it may be much more recent. The straight and tapered cobs are generally uncharred and fairly complete with small shank diameters and wide kernels in an average number of 10 rows. Nine cobs were examined. Tularosa Cave.—More than 30,000 maize cobs were recovered from this site in Pine Lawn Valley, northeast New Mexico, by archaeologists from the Chicago Natural History Museum. The Museum has submitted these cobs to Dr. Hugh C. Cutler, who has generously made available for inclusion in the present study samples representing three levels of Square 3R2. Maize cobs from the lowest levels of Tularosa Cave were dated at 2300 + 200 years by the Carbon 14 method. The cave is estimated to have been abandoned somewhere around 1000—1200 A.D., and thus was continuously occupied for about 1500 + 500 years (Martin ef al, 1952). Material from Square 3R2, Levels 3, 6, and 11 was studied as three separate sam- ples, each of which was chosen so that the cobs analyzed reflected the proportionate row numbers determined by Cutler (1952) from all cobs found in each of these layers. "There was a decrease in row number from the lower levels to the surface, average row numbers for Level 11 being 12—14, for Level 6, 10—12, and for Level 3, 8-10. Culturally, Level 11 is the oldest (400 В.С. + 200 years) and represents the Pre-pottery Phase. Level 6 dates back to 500—600 A.D., and represents the Georgetown Phase. Level 3, at 1000—1200 A.D., represents the San Francisco Phase. Twenty-five cobs from each of the three levels were examined. Point of Pines.—This sample consists of about 250 ears of charred maize. It Was recovered from Room 50 of a large ruin at Point of Pines, Arizona, numbered Ariz. W:10:50, and was submitted by Dr. Emil W. Haury, Director, Arizona State Museum, Tucson, Arizona. He believes its age to be 1250 A.D. (personal communication, 1953). It can be assigned to the Pueblo II cultural phase. The straight and cigar-shaped ears are 5-8 cm. in length, with small shanks and smooth kernels wider than thick in an average number of 10 rows. Seventy-five cobs Were examined. [Vor. 40 86 ANNALS OF THE MISSOURI BOTANICAL GARDEN MEXICO AND CENTRAL AMERICA PRESENT-DAY RACEs: Chapalote——This race of popcorn is considered by Wellhausen её al (1951, 1952) to be one of the ancient indigenous races of Mexico. It was mentioned by Anderson (1944a, 1946) as being allied to the primitive Maíz reventador. Well- hausen e£ al pointed out its resemblance to archaeological maize finds at Painted Cave (Haury, 1945) and Cottonwood Cave (Hurst, 1948; Hurst and Anderson, 1949). They considered it to be one of the most distinctive races of maize in Mexico. Mangelsdorf (1948) and Mangelsdorf and Smith (1949) regarded it as possessing the "weak" allele for pod corn. The cigar-shaped ears are 10—15 cm. long, with small shanks and smooth rounded chocolate-brown kernels in an average row number of 12. Ten ears were examined. Guatemala Flint.—A common variety of the Guatemala highlands, this race was described by Anderson (1947a), who considered it to represent one of the basic elements in Guatemalan maize. Ears are 10—25 cm. long, with a conspicu- ously enlarged and irregularly rowed butt, heavy shanks, a heavily sclerenchyma- tized rachis, and wide flinty kernels (some of which are capped with soft starch) in 8 (occasionally 10 or 12) straight rows. Eleven ears were examined. ARCHAEOLOGICAL REMAINS: Lower California.—This sample of 52 maize remains was recovered from Cave B.C. 100, located 8-10 miles east-southeast of Comondu, in central Baja Cali- fornia, and was submitted by Dr. William С. Massey, Dept. of Anthropology, University of Washington, Seattle. The cobs were recovered from a layer dated by Dr. Massey at 1697—1750 A.D. (personal communication, 1953). The cigar- shaped and straight cobs are 6-12 cm. long, with small shanks, and bore nearly isodiametric kernels set in 10-12 rows. Thirty-five cobs were examined. SOUTH AMERICA PRESENT-DAY RACES: Peru Flour.—This sample was collected by G. Edward Nicholson at Huancayo, Peru, where it was being offered for sale in one of the native markets under the name of "Maíz de Color.” It belongs to the race Cutler (1946) called “Valle Maize.” Ears are short, tapering, constricted at the base, with small shanks. The race exhibits little sclerenchymatization. Kernel color varies from ear to eat; brown, red, yellow, and various delicate striped combinations are most common. Kernels are pointed, as wide as thick, and arranged in prominent rows averaging 10 per ear. Twenty-three ears were examined. Coroico.—This race, described by Cutler (1946) as "the most unusual race so far known," has the odd characteristic of brick-like arrangement of “‘alicoles” (cupule plus pair of associated spikelets; Nickerson, in press). Cutler described the tapered ears as long, slender, and flexible (25-30 cm. lengths were common in the collection examined), with a light brown cob, small pith, and brown-orange 1953] NICKERSON—VARIATION IN COB MORPHOLOGY 87 kernels. He stated that the row number averaged 9, but he now believes (personal interview, 1953) that two successive rings of alicoles constitute a single diametral whorl; thus the row number of Coroico is generally high, averaging 18. Whether 9 or 18 rows are present is governed by the state of condensation of the cob axis. Twenty-five ears were examined. ARCHAEOLOGICAL REMAINS: Arica.—These samples were recovered from two midden sites, Playa Miller and Quiani, located on the coast in the vicinity of Arica, Chile (18° 30’ S. Latitude), by Junius B. Bird, Associate Curator of Archaeology, American Museum of Natural History, New York. Cobs from both sites were treated as one sample following Bird’s (1943, 1948) suggestion that Playa Miller represented a continu- ation of sequences started at Quiani. The oldest maize remains from this area are dated at about the time of Christ (Bird, personal communication, 1952). It is thus younger than material from the Huaca Prieta sites discussed below. The cigar-shaped and tapering cobs are 5-12 cm. in length and 1—2.5 cm. in width, with small shanks and an average row number of 14. Twenty-one cobs were examined. Huaca Prieta.—These samples, also from the American Museum, were recovered by Junius B. Bird from Huaca Prieta, a site on the Peruvian coast at the south of the Chicama Valley (app. 8° $. Latitude). These comprehensive samples consist of 604 cobs from Test 1 and 171 cobs from Test 5. The two groups were treated separately, since Bird regarded cobs from Test 5 as older than those from Test 1, and were designated Huaca Prieta 5 and Huaca Prieta 1, respectively. Cultures responsible for the Huaca Prieta middens began about 3000 B.C. (Bennett, 1948; Bird, 1948), but maize did not make its appearance there until 850 B.C. (Bird, personal communication, 1953). Implications as to archaeological and botanical Significance of these finds are further discussed by Anderson (1947b), Carter (1950), and Whitaker and Bird (1949). The well-preserved, tapered and cigar- shaped cobs are remarkably uniform, especially those found in the lower levels, which have small shank diameters, an average row number of 14, and are horny rather than brittle or bony in texture. Thirty-five cobs from Huaca Prieta 1, representing 7 layers, and 44 cobs from Huaca Prieta 5, representing 5 layers, were examined. ASIA Assam and Lushai.—The samples from Asia were contributed by Stonor, who collected them in the hills of Assam. The history and morphology of varieties in this collection have been described by Stonor and Anderson (1949). The cobs used here were those actually raised in Asia. For this investigation they were Separated into two groups: (1) Assam, which includes 21 cobs from tribes other than the Lushai; (2) Lushai, which includes 8 cobs representative of the maize Varieties grown by the Lushai tribe. Cobs of both groups are sufficiently alike to TABLE I AVERAGE VALUES OF NINE ca amm CHARACTERS AND TWO INDICES FOR EACH MAIZE COB SAMPLE STUDIED (M urements in mm. except where otherwise indicated % cobs Height of Kernel % cobs | Lower gl. | Cob/rachis | Rachis/pith Name at Sanpla straight | rachis flaps | thickness | tapered width index index Iroquois Sacred Flour 43 2.0 3.7 43 8.7 1.7 2.1 Northern Flin 35 2.0 4.4 32 7.0 1.8 2.1 Hopi White Flour 35 2.0 4.7 65 6.2 17 1.8 Hopi Blue 35 1.6 4.2 0 5.8 1.9 1.8 Papago Flou 0 1.2 4.0 28 5.4 1.8 1.9 Basketmaker 9 0 1.5 4.3 16 4.4 1.6 2.0 Marsh Р 1 16 1.5 4.1 20 5.5 1.5 2.1 Turner Site 8 12 1.0 3.0 50 3.7 1.3 1.7 Luster Cave = 33 1.6 4.5 33 3.4 1.8 1.8 Tularosa Cave—L. 3 9 0 1.5 4.1 4 5.9 1.8 2.4 Tularosa Cave—L. 6 7 4 1.3 4.7 48 5.5 17 2.3 Tularosa Cave—L. 11 8 0 1.0 4.0 4 5.5 1.7 2.3 oint of Pi 8 20 t2 3.6 3 5.5 1.7 2.0 apalote - 0 1.0 4.0 10 5.4 zl 2.5 Guatemala Flint 15 0 13 4.5 44 6.8 1.9 2.9 er California 8 37 1.3 5.8 27 5.9 1.8 2.1 Peru Flour 8 8 .9 4.0 92 5.0 1.9 24 roic 13 0 P 6.6 95 5.9 1.9 11 i 6 0 5 3.2 30 3.5 1.6 2.1 Huaca Prieta 1 4 5 .3 Ai 31 2.4 1.6 1.8 Huaca Prieta 5 6 9 2 3.0 13 а 1.9 1.9 Assam 12 0 .9 4.9 75 5.0 1 2.3 Lushai 10 0 s 4.0 75 6.0 1.9 2.0 МЯҤачу5 'IVOINV.LOdH IWüOSSIW AHL AO SIVNNV 0p тод] 1953] NICKERSON—VARIATION IN СОВ MORPHOLOGY 89 зр ne Ш 5 SACRED MOVE il tower QVATEMALA ном BLUE ном WHITE Е CAUFORNIA MNT nove nour P] N AUSMAI ASSAM TULAROSA PAPAGO TULAROSA MARSH LUSTER uvm + = МУН 2 PASS CAVE w P о $075 H BASKETMAKER eu, o se ive n % * v COROINCO CHAPALOTE шее эд Хы” 3P e 2 A Г a L A a n Г n Г B. —————— Sl 2.00 1.50 1.00 .50 0 .50 1.00 150 200 neg pos CUPULE DEPTH MM. PERCENT OF COBS 8- ROWED $ FY a 100 . 51 50.1 0 SHANK DIAMETER é 4 * . 16 15 . 10 9. PERCENT OF СОВЅ STRAIGHT 9 . . 30 29 . 10 9.0 HEIGHT ОҒ RACHIS - FLAPS v ә . 20 19.9 Ж. KERNEL THICKNESS » E . 244 41. 36 35. PERCENT OF COBS TAPERED » bd . 6l 60.31 30.0 LOWER GLUME WIDTH ~ ~ а „60 59.40 39 . sa б Further posal tion in скалой л. in ug 1, showing relation- of maize. Each dot in that с groupi horizontal‏ ا ters are diagrammed by rays,‏ NHGWVO 'IVOINV.LOS MNOSSIW AHL AO SIVNNV vw К y yy v w + м y » 3 м LUSHAI MODERN IROQUOIS SACRED FLOUR MODERN V L y $ es ON у w E dy % v% v% « + €, зу W N зуу ws Е" > ч ASSAM MODERN NORTHERN FLINT mere Ob TOA] 5 * ~ 7 № S v, | л, ws $$ * > 4 rs è . vee o,f E k^ ` e а r e Ы L HUACA PRIETA 1 850 ВС.- 1000 AD. ARICA 0 - 1250 SINGLE SPECIMEN SCORES NUMBER OF ROWS 4 $ ° 8 10 2 STRAIGHT СОВ « e PRESENT ABSENT М TAPERED СОВ ° E ° PRESENT ABSENT HUACA PRIETA 5 850 B.C.- 1000 AD. Text-fig. Te Pe 3, 4, 5. Twenty-three eee — diagrams dita. ie variation and relationships of nine measured maet erat TM each group of maize cobs stu — Each dot esents the actual measurements e ear. Horizontal and vertical axes and seven additional characters are scored as in text-fi xcept row number a nd roy erue which were ind. as indicated in the lower right-hand corner of text- ug E Dotted 8. 2, ray indicates no iita could be m A9SO'IOHdWONW 800. NI NOLLVINVA—NOSWHMOIN [£861 92 ANNALS OF THE MISSOURI BOTANICAL GARDEN > > »»»» НОРІ WHITE FLOUR MODERN ~ [Vor. HOP! BLUE FLOUR MODERN 40 PERU PLOUR MODERN GUATEMALA FUNT MODERN + * Ф v * + aww wes vv A. c ou» X + + » 9% Ф y + № wre & % + v Ne + ^ x € » Ф Ф «4. У % LOWER CALIFORNIA 1650 1700 MARSH PASS 300 1000 v à Ww Y we "nw Sew у $ wee ч аҹ <“ é ч чоч i в + % * * r ¥ » < COROICO MODERN TURNER SITE 1000 . 1100 Text-fig. 4 AS9O'IOHdWONW ЧОО NI NOLLVIWVA—NOSWAWOIN £6 [8661 є 7 v v X* ~ » v. WM у v + x چ‎ САЛЫ $ ~ > * LUSTER CAVE 900 - 1000 PAPAGO MODERN ¥ % * ¥ + 4 $ М 422972. v 4 69 à Ф é € - CHAPALOTE MODERN TULAROSA LEVEL 3 900 - 1000 r6 NHGWVO 'IVOINV.LOH INNOSSIW AHL ЧО SIVNNV 0p '10A] > Ф s% bé de ۳ xy ei. Per. $$ • * * © POINT OF PINES 1250 1300 e Y “i v Ф v vvv M $ LY ر‎ v Wey ¥ TULAROSA LEVEL 6 500.700 BASKETMAKEP 200 300 Text-fig. 5 TULAROSA LEVEL П 300 BC.- 150 BC. A9O'IOHdWONW 800 NI NOLLVIYVA—NOSYAMOIN S6 [£ $61 [Vor. 40 96 ANNALS OF THE MISSOURI BOTANICAL GARDEN conform to the same general description. They are tapered, 15—25 cm. in length, with a medium-sized shank and bore 14—16 rows of narrow kernels. That they are markedly lighter in weight than ears of North and Central American maize of comparable size is undoubtedly another expression of their lack of heavily sclerenchymatized tissues. MEASUREMENTS OF SAMPLES The number of cobs in each of the above-listed samples varied, but an attempt has been made to examine enough cobs to form consistent pictures of variation in the populations represented. Measurements recorded for each cob include, in addition to those of four external and five internal characters mentioned earlier. diameters of cob, rachis, and pith. From these last measurements, а cob/rachis index and a rachis/pith index were computed. Since plants are affected by their environment in numerous ways, several rep- resentatives of any population must be studied to formulate a complete picture of their range of variation. This idea applies quite as much to archaeological remains as to living plants, for, as Cutler (1952) has pointed out, a single speci- men of any archaeological plant remains is of little significance; it may represent TABLE II AVERAGE VALUES OF FIVE MORPHOLOGICAL CHARACTERS FOR ТУ/ЕМТУ-ЕТУЕ EXISTING RACES OF MAIZE IN wr ur (DATA FROM WELLHAUSEN ET AL, 1 5 х (Measurements іп mm. except where otherwise indicated.) Cupule Width of % cobs Shank Kernel Name of race width lower glume 8-rowed diameter thickness Palomero Toluqueüo 3.0 3.0 0 8 2.8 АгосШо Amarillo 3.5 3:5 0 8 2.5 Chapalote 5.5 55 20 9 4.1 Nal-Tel 5.0 9:9 20 Z. 3.9 Cacahuazintle 5.0 7.0 10 5.2 Harinoso de Ocho 8.0 9.5 100 14 4.4 Olotón 7.0 8.5 20 17 6.0 Maíz Dulce 7.0 6.0 0 11 4.0 ӧпісо 4.0 5.0 0 8 3.6 Reventador 5.0 5.5 20 8 3.6 Tabloncillo 9.0 10.0 90 11 4.3 hua 10.5 9.5 0 21 3.9 Tepecintle 9.0 8.0 0 10 3.7 iteco 8.0 9.0 0 22 4.5 8.0 7.0 0 34 4.6 Zapalote Chico 7.5 7.5 20 13 3.6 Zapalote Grande 7.5 7.0 0 18 3.8 Pepitilla 5.0 0 12 3.5 Olotillo 7.5 10.5 90 10 3.9 Tuxpefio 7.5 0 20 13 3.7 Vande 8.5 8.5 0 13 3.6 Chalquefio 6.0 6.5 0 10 3.9 aya 7.3 7.0 20 9 LY Cónico Norteño 6.0 3:5 0 11 3.5 lita 9.0 10.0 50 9 4.1 ا ا Шш‏ ت 1953] NICKERSON—VARIATION IN COB MORPHOLOGY 97 an accidental deposition, or it may have been buried in an old layer or brought to a recent one by rodents, pot hunters, or a recent occupant of the site. To mini- mize these effects, arithmetic averages of the measurements of each character have been computed for each sample. Alava (1952) pointed out that such a technique provides the only safe basis for studying variation between different varieties of plants. Table I lists the average values for each of the characters of each sample studied. Ear shape contributed two columns, one for straight ears and one for tapered ears, thus giving a total of nine characters in which to compare variation. A pictorialized scatter diagram of these same results is presented in text-fig. 2. зр 1 MARSH 9ASS 2 TULAROSA LEVEL Ф 3 PAPAGO — ^ OUR ne 4 TULAROSA t 5 OWER — CALIFORNIA Е Др». i нот BLUE nour a ; мот white пош rom Ó SOUTA т NORTHERN „>. TABLONCILLO \ You = E ane „ә Е ня a£ aped tyster > ГЕ B on m? b E di s № | v : 25 pe XA „бы "Sar d A me S 5 or hints pes conico 5 ВАЗ ЕТМАКЕЕ ENO CHALQUENO w ore у О araor - d a E] REVENTADOR d v 5 TURNER TULAROSA coroico CACAMUAZINTLE © SITE LEVEL п e das "0 b м comico те oS HUACA PALOMERO AROCILLO PRIETA $ TOLUQUENO AMARILLO r : А : x ^ a a A 2d —— EUN Y А т n n А ~ 8 9 10 n LOWER GLUME WIDTH MM. LJ PERCENT OF COBS 8 - ROWED E Ф 100 . 75 74 - 5l 50.0 SHANK DIAMETER E é Ф - lé 15 . 10 9. LJ KERNEL THICKNESS » ә „АЛ 41. 36 35 - Text-fig. 6. Pictorialized scatter diagram of material from Tables I vi II —Á — ships betw variation in five different cob characters among the twenty-five races tw * Each dot represents the average measurements of five characters for all maize 8rOup; horizontal axis, width of ees er glume; vertical axis, width of Lp - diagrammed by rays as sistas d on the figure. Further explanation in [Vor. 40 98 ANNALS OF THE MISSOURI BOTANICAL GARDEN Text-figs. 3, 4, and 5 show the variation and relationships of these same nine cob characters within each sample of cobs. The method of construction, as well as the general usefulness and reliability of these diagrams, is further discussed below. An attempt was made to compare the results of this investigation with those obtained by Wellhausen e£ al (1951, 1952) for races of maize in Mexico. Several of the major characters employed in the present study were not employed by Wellhausen and his co-workers. However, their paper contained excellent diagrams of cross-sections of ears of each race drawn to scale; it was possible to make certain measurements directly from these diagrams. Lower glume width in mm. was measured at the widest point indicated on the third concentric circle of each езг diagram. Cupule width in mm. was measured as a straight-line distance across the bases of any two consecutive kernels from the points where their boundaries intercepted the rachis diameter (innermost circle). Percentage of 8-rowed ears was obtained by comparing the number of rows in each drawing with the average values listed for each race in their Table 15. Average values for shank diameter and kernel thickness were obtained directly from their Table 15. The informa- tion obtained is summarized here in Table II; it is also presented as part of text- 8. 6, a pictorialized scatter diagram which compares related data from both Tables I and II Male spikelet variation has been studied by Alava (1952) for seven of the twenty-three samples used in this investigation. А crude index has been con- structed after the manner outlined by Anderson (1936) for both Alava's data and the appropriate data set forth in text-fig. 2. Text-fig. 7 shows the relation- ships of these index values. The significance of each of the above comparisons is discussed below. Discussion Whiting (1944) and Cutler (1951) pointed out that botanists alone probably cannot solve the problem of the origin of maize, because maize is so intimately linked with human cultures. This point is emphasized by the fact that maize is completely dependent upon man for its preservation. Whatever may have been the original factors leading up to this novel situation, the fact remains that maize and man have been associated years. In a study which involves the disciplines of ethnology, archaeology, and botany, caution must be exercised lest theories proposed in one science be regarded as proven facts by workers in another. Since the present investigation has been conducted from the standpoint of botany, such conclusions as can be drawn should depend in so far as possible upon botanical evidence for their support. To this end, botanical information is discussed first, and an attempt is then made to relate this discussion to known facts of ethnology and archaeology. 1953] NICKERSON—VARIATION IN COB MORPHOLOGY 99 BOTANICAL AsPECTS OF MAIZE Сов ANALYSIS Variation among races of maize is nowhere better illustrated nor more difficult to describe than that found among female inflorescences, or cobs. Within any race, many cobs possess characters the measurements of which lie well within the range of variation of the same measurements from several other races. This exasperating fact is true for all the characters used in this study. Using any one cob character at a time, it is nearly impossible to distinguish races of maize; it is for this reason that the empirical classification of maize by kernel types proposed by Sturtevant (1899) is not of much value to modern maize breeders. Mangels- dorf and Reeves (1939) used evidence from several sources to characterize maize from different parts of the New World. Anderson (1946, 1947a), Anderson and 15Р өс DE ФЕ 2 Б oF өс < eb ot < > of o 6 № ш өз ә ә < Гем 3 № ФА + r1 i L JL — НЮ 3 6 9 12 15 18 СОВ CHARACTERS INDEX VALUES MAIZE RACE СОВ CHARACTERS TASSEL CHARACTERS А ARICA 1 1 B ASSAM 9 5 C CHAPALOTE y 9 D PERU FLOUR n 8 Е GUATEMALA FLINT 13 8 F NORTHERN FLINT 18 12 7 15 С PAPAGO Text-fig. 7. Scatter diagram of index values of seven races of pis i in which both the x ага (male їпйогезсепс ce) and the cob (female inflorescence) have been inves gated. Horizonta к ex values red e" cob study; vertical axis, index values derived as tassel study, Further explanation’ in the [Vor. 40 100 ANNALS OF THE MISSOURI BOTANICAL GARDEN Cutler (1942), Mangelsdorf and Smith (1949) and Wellhausen e£ al (1951, 1952) ave used an increasing number of subtler characters in distinguishing both modern and prehistoric races of maize. Pictorialized scatter diagrams (Anderson, 1949b) are a means by which several such variable characters can be considered at once, thereby emphasizing consistent and typical differences as well as similarities among the particular entities involved. Methods by which a pictorialized scatter diagram is developed are explained by Anderson (1949b, 1952) and Anderson and Gage (1952). Anderson (in press) has discussed the biological and mathematical criteria upon which pictorialized scatter diagrams are based. The value of such diagrams has been emphasized by Stebbins (1952), who considered the method “by far the best yet devised for making the observer aware of a pattern of variation in respect to three or more characters which are varying simultaneously." He also stated (1952) that the method "has the advantage of being repeatable and of focusing the attention of the observer upon certain essential features of variation." The data listed in Table I have been plotted as a pictorialized scatter diagram (text-fig. 2) to show how the several cob characters are varying and to what extent their variation is correlated. Measurements of depth and width of cupule were chosen as abscissa and ordinate respectively because they varied consistently within each sample (text-figs. 3, 4, 5), could be ascertained from fragmentary cobs, and were susceptible to accurate measurement. The other seven characters were then indicated by rays from each dot of the scatter diagram. Extremes of each character tending to be associated with higher values of cupule width and depth were represented by long rays, while extremes tending to be associated with lower ordinate-abscissa values were represented by no rays. Several facts concerning the maize samples studied were immediately apparent. At one extreme (upper right-hand corner of the diagram) were predominantly straight 8-rowed cobs with wide, shallow cupules, large rachis-flaps, wide lower glumes, heavy shanks, and thick or moderately thick kernels. This complex of characters was found in Northern Flint and Iroquois Sacred Flour; occurrence of endosperm as flinty or floury is dependent upon a single gene and is therefore of no significance here. These results agreed with those found by Brown and Anderson (1947) for typical Northern Flint varieties; the same complex was identified by Carter and Anderson (1945) as the Eastern complex. At another extreme (lower left-hand corner of the diagram) were cigar-shaped cobs with high row numbers narrow deep cupules, small rachis-flaps, narrow lower glumes, small shank diam- eters, and thin kernels. "These characters were associated in prehistoric maize recovered at Huaca Prieta, Peru, and Arica, Chile. It is significant that the majority of cobs were essentially complete but only about 5 cm. long. As a whole, these samples were remarkably homogeneous (text-fig. 3). Existence of a thi extreme (middle left area of the diagram) is indicatd by tapered cobs with high row numbers, extremely deep cupules, little rachis-flap development, wide lowe! — glumes, medium shank diameters, and thick or moderately thick kernels. The two _ 1953] NICKERSON—VARIATION IN COB MORPHOLOGY 101 maize samples from Asia were responsible. Cobs studied were those grown by Stonor in Asia; varieties involved were described by Stonor and Anderson (1949). The most striking features of the Asian corn cobs were their extremely deep cupules and their particular form of sclerenchymatization. The rachis tissue and lower glumes of all Asian cobs studied were quite hard, but their hardness was not identical with that of “bony” rachis tissue described by Mangelsdorf and Smith (1949) in connection with teosinte introgression. Whatever its cause, the end product was an extremely stiff, light-weight bulky cob. Studies by Lenz (1948) did not indicate that cell wall thickness had much to do with degree of sclerenchy- matization among various races of maize. A deep cupule could possibly be in- terpreted as a condition which showed little if any “Тирзасо influence" (Cutler, 1946). An inference cannot be drawn that all Asian maize varieties are alike in this respect, but for both Assam and Lushai samples, this conclusion seems justified. The nine characters employed in constructing text-fig. 2 showed a strong tendency for association at both extremes. A strong overall association between all sample averages as well as within each sample is clearly shown by comparing the average of any sample in text-fig. 2 with its appropriate population diagram in text-figs. 3, 4, or 5. Individual cobs in the upper right-hand area of each population diagram consistently showed more and longer rays than did those from other parts of the same diagram. This similarity of separation patterns in the several diagrams emphasizes the fact that those limits employed in separating sample averages, while originally chosen in a subjective taxonomic fashion (Steb- bins, 1952), were also working within each sample in such a way that morpho- logical trends associated with a particular race of maize were associated with each other. The diagram as constituted contains samples of rather diverse races of maize; closeness of position on the diagram and similarity of glyphs indicate similar morphology but do not necessarily imply close relationship. Weatherwax (1936) stated that no sound basis exists for any attempt to trace maize origin by arranging Varieties from different localities in an evolutionary sequence. Introduction of the dimension of time in the form of archaeological material, however, allows the investigator to arrange maize varieties of both present and past according to cultural succession; the result may or may not be an evolutionary sequence, but Such an arrangement is the strongest means available for documenting maize history and has therefore been employed in this investigation whenever possible. Results obtained in this investigation are compared below with those obtained in three other independently conducted investigations. A high degree of correla- tion is found in each instance, a fact which indicates a basic soundness in methods employed for cob analysis in this investigation. NE Brown e£ al (1952) have discussed relationships among three — of Hop Р maize, two of which, Норі White Flour and Норі Blue Flour, were included in the present study. These workers found that the two varieties had certain re- [Vor. 40 102 ANNALS OF THE MISSOURI BOTANICAL GARDEN semblances to Basketmaker and Eastern maize as well as to each other. Unfor- tunately, it was not possible to obtain a large enough sample of Kokoma, the variety they reported as resembling Basketmaker most closely, to warrant its inclusion here. Text-fig. 2 indicated that Hopi Blue Flour was basically similar to Basketmaker, differing in cupule width, wide lower glumes, heavy shanks, and presence of 8-rowed cobs. It should be noted that there has been no change in cupule depth and predominance of cigar-shaped ears in development of Hopi Blue Flour from Basketmaker maize. Hopi White Flour was found by Brown e£ al (1952) to contain an admixture of both Eastern and Mexican germ plasm. The Mexican complex (Carter and Anderson, 1945), known to be present on the Mesa Central of Mexico, is char- acterized by strongly tapered cobs with high row numbers, thick kernels, and small shanks. This conclusion was borne out by text-fig. 2, which showed Hopi White to have all the characters associated with the Eastern complex with the exception of row number and cob shape; tapered cobs of high row number may be attributed to Mexican influence. Cobs of these two Hopi Flour maizes differed from each other in cupule depth (text-fig. 2 indicates this to be the strongest character con- tributed by the Eastern complex), cob shape, shank diameter, rachis-flap height, and row number. In the correlation between results obtained in the present investigation and those of Wellhausen e£ al (1951, 1952) which are listed in Table II, the number of characters used was less than the number employed to separate the samples listed in Table I; yet, a pictorialized scatter diagram (text-fig. 6), using five measurements, presented essentially the same variation pattern as did text-fig. 2, in which nine measurements were employed. It should be noted that the basic pattern of maize cob variation was not altered by using measurements of a different character for the abscissa of text-fig. 6 than that employed in text-fig. 2. If a sample on text-fig. 6 has any real affinity to races with which it may be allied on other grounds, the fact should be shown by its position on the diagram. Text-fig 6 should also lend further confirmation to the genealogies postulated by Wellhausen et al for several hybrid races of maize. That these confirmations are indeed pos- sible was shown quickly and easily. The four races classified by these workers as Ancient Indigenous Races (Palomero Toluquefio, Arocillo Amarillo, Chapalote, Nal-Tel) were closely grouped; moreover they were all placed in the lower left- hand corner of the diagram and all had few to no rays. Thus their characters 0 small cupules, narrow lower glumes, high row numbers, thin kernels, and small shanks, all believed to indicate primitiveness in relation to other races of maize here considered, were strikingly indicated by text-fig. 6. Three of the four races classified as Pre-Columbian Exotics (Cacahuazintle, Harinoso de Ocho, Olotón, Maíz Dulce) formed a consistent graded series along the lower edge of the upper region of text-fig. 6. They were placed in a separate class because of their antiquity and resemblance to South American maize. i 1953] NICKERSON—VARIATION IN COB MORPHOLOGY 103 the class was somewhat artificial, the positions of three of its four constituents on the diagram in a nearly perfect ascending order indicated a basic similarity in variation pattern. From these races a group of thirteen Prehistoric Mestizos and a group of four Modern Incipient races were thought to have arisen. For many of these, Well- hausen e£ al compiled a chart showing the two putative parents of the particular race under discussion. Out of a total of eleven genealogical relationships involving other races plotted on text-fig. 6, seven were found to be borne out by diagram position, being for the most part midway in both position and number of rays. These genealogical relationships as postulated by Wellhausen e al are listed below. Group 1 is comprised of those races whose positions on text-fig. 6 offered further proof of relationship to their putative parents. Group 2 is comprised of those races whose positions on text-fig. 6 did not approximate positions between their putative parent races. GROUP 1 Cónico. Palomero Toluqueüo X Cacahuazintle Zapalote Chico : al-Te X Tepecintle Chalquefio Cénico X Тахрейо ónico Norteño. Celaya х nico ; Bolita Zapalote Chico x Tabloncillo Tuxpefio. Tepecintle X Olotillo Comiteco T X Olotén GROUP 2 Zapalote Grande Teh x Zapalote Chico andeno. Тох; x Zapalote Grande Celaya Tabloncillo х Tuxpeño Mhao Tabloncillo x miteco The fact that а majority of such genealogies can be substantiated by arranging data obtained from the cobs alone is added proof that races may not only be recog- nized by cob characters, but that their histories can also be ascertained in the same manner with a high degree of accuracy. It should also be borne in mind that data for the Mexican samples represented averages determined from 3—5 cobs of each race; the accuracy might well have been increased had more cobs been used in determining average values. Text-fig. 6 further showed that variation in Mexican maize followed the same general pattern of character association as did Variation among the samples employed in this investigation, and that prehistoric North and South American samples were close to the primitive Mexican races. - Spikelet variation in Zea Mays was studied by Alava (1952), using methods Which, while applicable to male inflorescences, were of no use in the investigation of maize cobs. Thus if any correlation between results of two independent y conducted investigations should exist, it would not only be further substantiation for validity of methods employed and results obtained, but also would indicate that cob and tassel analyses yield comparable results. Similar material from seven different races was examined in each investigation. А numerical index (Anderson, [Vor. 40 104 ANNALS OF THE MISSOURI BOTANICAL GARDEN 1936; Stebbins, 1952) for both cob and tassel characters was computed for each race. ile admittedly a crude indicator of morphological characters, numerical indices are valuable for summarizing diverse data in a manner which permits direct comparisons to be made. "Text-fig. 7 shows the extent of this correlation. Arica had the smallest value in both indices, and there was a general progression toward the upper right-hand corner of the graph occupied by Northern Flint. Papago appeared somewhat aberrant as far as tassel index was concerned, but came closest to Chapalote, a related race, in both values. Peru Flour and Guatemala Flint were close, a further expression of their presumed relationship to one an- other. Guatemala Flint was closer to Northern Flint for cob-index value. Although a considerable number of units away, Asian material was closest to Arica in both cob and tassel values. Thus with the possible exception of Papago, in which variation is unique but consistent with its past history, the samples showed high correlation between tassel and cob morphology in distinguishing races of maize. Maize Сов ANALYSIS IN RELATION TO ETHNOLOGY AND ARCHAEOLOGY Knowledge of ancient peoples is in large part built upon a detailed study of their refuse heaps; rubbish being what it is, a place is assured for the botanist in archaeological analysis. Analysis of maize remains is especially helpful, since few other plants have become so closely associated with man to the extent that they are reliable indicators of his early history in many parts of the New World. SOUTH AMERICA AND ASIA Lowie (1940) pointed out that it was a common mistake to identify all early agriculture with maize. Sauer (1952) contended that seed-crop agriculture was developed after a root-crop agriculture which involved vegetative propagation only. Excavations in northern Chile (Bird, 1943), and more especially in Peru (Bennett, 1948; Bird, 1948), showed that there were indeed agricultural com- munities without maize. Layers of middens at Huaca Prieta, Peru, beginning about 4000 years ago, contained squash, beans (4 types), chili peppers, Canna, cotton, and bottle gourds (Lagenaria siceraria, Whitaker and Bird, 1949). Maize did not appear until 850 В.С. Whitaker (1948) pointed out that it is extremely difficult to account for the bi-hemispheric distribution of Lagenaria, as it is pre- sumed to be of Asiatic origin. Partly on such distributional patterns, Carter (1950) postulated pre-Columbian contacts between the Old and New Worlds, a theory also put forward by Stonor and Anderson (1949) and Sauer (1952). That it was quite possible to sail in either direction across the broad expanse of the Pacific has been amply documented by Buck (1938). Another odd fact about prehistoric plants of west-coast South America is that they are presumed to be native to Central America. Either these plants reached the coast as articles of trade among established peoples, or they were taken there by the first settlers of the area, who either came from or passed through the Cen- 1953] NICKERSON—VARIATION IN COB MORPHOLOGY 105 tral American region. It does not seem possible that maize would have been over- looked by these peoples as a crop plant had they passed through an area in which it was growing or being grown. Yet evidence on the possible antiquity of maize in central Mexico was recently reported by Barghoorn (1952). Maize pollen has apparently been recovered from sedimentary deposits under Mexico City which may antedate human occupation. Recently, however, the date of human occupa- tion has been pushed back to 9000 years in the same area (Richards, 1953). A type of maize hitherto not mentioned in this investigation is that which is generally referred to as popcorn. True pops are generally of quite primitive mor- phology (Mangelsdorf and Smith, 1949). Prehistoric Arica and Huaca Prieta maize remains may quite possibly be those of popcorns; Wissler (1945) showed that this is true for some whole maize ears excavated at Arica. In ear shape and size and in tassel characters (Alava, 1952), these prehistoric remains were com- parable to a modern pop variety collected by Parodi in Argentina. They may also be related rather closely to Asian types, since they had deep cupules, high row numbers, and some tapered cobs. Tassel analysis (Alava, 1952) also showed an ' Asian similarity. The above evidence is suggestive, but it also indicates the neces- sity for more data before the history of maize in both South America and Asia can be considered to be completely known. THE AMERICAN SOUTHWEST Maize in the American Southwest has had a long and complicated history. Man was present in the area 10,000 years ago (Wormington, 1949; Johnson, 1951) but there was little evidence that the area was under continuous occupation from that time until the advent of agriculture. Randolph (1952) stated that an ap- parently humid and subtropical climate existed there 5,000—10,000 years ago, and that the area itself could be the original home of maize. Carter (1945) likewise postulated the Southwest as a center of agricultural dispersal. All known archae- ological remains which contain maize are younger than 5,000 years, so the problem at present is to evaluate the evidence at hand. It should be kept in mind that whatever statements are made below are subject to revision in the light of further discoveries, Carter and Anderson (1945) noted three major cultural provinces in the Southwest: Hohokam, Mogollon, and Anasazi (Puebloan). At the time of their survey, little was known of the Mogollon civilization, but recent excavations at Tularosa and Cordova Caves (Martin et al, 1952) have produced evidence estab- ishing this cultural province as one of equal importance with Hohokam and Anasazi. Archaeological maize is now known for each of the three cultural Provinces. Hohokam maize was found in Ventana Cave, Arizona (Haury, 1950). It is similar to but not identical with modern Pima-Papago varieties (Carter and Anderson, 1945). Chapalote, an ancient Mexican race, is today found in north- Western Mexico, an area which was also part of the Hohokam cultural province (Amsden, 1949). Relationships between Papago and Chapalote have been dis- [Vor. 40 106 ANNALS OF THE MISSOURI BOTANICAL GARDEN cussed above. Since the Hohokam area was relatively isolated from influences of neighboring cultural patterns, the maize found there today is much the same as it was in ancient times. It is reasonable to assume that prehistoric maize in this cultural province was similar but not identical to that of Basketmaker and Mogollon cultures, and that modern Papago and Chapalote were derived from it. It was shown (text-fig. 2) that maize remains recovered from Tularosa Cave Level 11 were quite similar to both Chapalote and Basketmaker. Similarity between Chapa- lote and Basketmaker П maize from Cottonwood Cave (Hurst, 1948; Hurst and Anderson, 1949) was pointed out by Wellhausen e£ al (1951, 1952). Amsden (1949) noted the similarity of modern Papago to Basketmaker, as did Carter and Anderson (1945). The present investigation showed that Basketmaker resembled Chapalote much closer than it did Papago, but the important point is that a three cultural areas, Hohokam, Anasazi, and Mogollon, had at an early date maize which was variable but essentially similar. Maize of the Hohokam area remained _ relatively unchanged; that of the other two areas was strongly influenced by maize from more remote places. Text-fig. 5 illustrates the variable nature of early southwestern samples. In contrast to maize of the Southwest, remains from Arica, Chile, and Huaca Prieta, Peru, are very homogeneous; furthermore, maize of these two areas (American Southwest and South American West Coast) were con- temporary. Explanation of the variability in the one location and homogeneity in the other at the same time is but one of the many problems of maize history. Basketmaker maize was widespread, having been reported from several sites in the Anasazi area. Some of the better known sites are Cottonwood Cave, western Colorado (Hurst, 1948; Hurst and Anderson, 1949), Mummy Cave, Сайоп del Muerto, Arizona (Anderson and Blanchard, 1942), and Painted Cave, northeastern Arizona (Haury, 1945). Basketmaker-like maize was reported outside the Anasazi area by Brown and Anderson (1947) in prehistoric remains of rock shelters and caves from the Ozarks to southern Ohio. Two combinations of maize characters, one peculiar to the Mexican Mesa Cen- tral and the other to the eastern United States, were superimposed in varying amounts on southwestern maize. The spread of each influence may be traced by analysis of successive samples; in such work, the importance of dated maize remains can hardly be over-emphasized. Mexican influence is characterized by strongly tapered cobs with high row numbers (Carter and Anderson, 1945), thick kernels, and small shanks. This type entered the Southwest from the Northeast rather than from Mexico directly; its greatest influence has been found in remains of the Fremont Basketmakers of Yampa Canyon, Utah, dated at 400-800 A.D. (Burgh and Scoggin, 1948). At Luster Cave, in extreme eastern Utah, maize was subjected to strong Eastern influence prior to 1000 A.D. Text-fig. 2 showed this sample to be closely related to Hopi White Flour among ethnological races examined in this study. Нор. | White Flour differed, however, in being more extreme for certain Mexican (tapered Ad 1953] NICKERSON—VARIATION IN COB MORPHOLOGY 107 ears, high row numbers) and Eastern characters (large shanks, wide lower glumes). Maize from the Davis Site (Newell and Krieger, 1949) was analyzed by Jones (1949) and found to be entirely Eastern. These maize remains were dated at about 700 A.D. by Jones and at about 400 A.D. by Johnson (1951). The question of origin of Eastern influence has long been a puzzling one. Anderson (1947a) and Brown and Anderson (1947) presented evidence that Eastern maize is related to Guatemalan maize. Jones (1949) reported that he and Krieger independently reached the conclusion that both maize and pottery types recovered at the Davis Site bore a resemblance to those of Guatemala. Carter (1946) suggested that maize could have been carried up the east coast of Mexico or across the Gulf to southern United States. Regardless of the method employed, this movement must have taken place in time for the influence to have been carried to the Southwest by 1200 A.D., a date which Carter and Anderson (1945) recognized as closely approximating its first appearance in that region. Judson (1951) called attention to the fact that a great drought occurred in the Southwest in 1276-1299 A.D., and that many Indian tribal migrations took place around that time. The direction of movement of maize into the Southwest was from east to west; Carter and Anderson (1945) noted that among the present-day Puebloan tribes, those of eastern pueblos have more eastern-like maize than do those of western pueblos. It is generally accepted that there is influence of Plains cultures in the southwestern area. This influence also is stronger in eastern pueblos than in western ones, and is another indication of how well maize history is cor- related with the history of the peoples who grew it. THE AMERICAN SOUTHEAST AND EAST Wherever the Eastern complex came from, it was probably carried up the river valleys of the Mississippi and its tributaries by Indian tribal migrations. A date of 900 A.D. was assigned to the earliest known Burial Mound I culture by Ford and Willey (1941), who also identified the beginnings of horticulture in eastern United States at this time. Burial Mound I peoples were in turn supplanted by Burial Mound П and later (1200-1400 A.D.) by Temple Mound I and Temple Mound II cultures. Ford and Willey derived the Iroquois culture of New York from a welding of these four intrusions onto an archaic hunting-gathering popula- tion; they considered that this and other allied upper Mississippian cultures reached a peak after 1500 A.D., and lasted until historic times. Although other evidence has been presented in support of the idea that maize was a late arrival юса woodland cultures (Linton, 1924; Kroeber, 1939), from a botanical point of view the postulated time of arrival of maize in this area appears too recent. This impression is strengthened by consideration of the length of time the same type of maize was present in the Southwest. Recent radiocarbon dating (Johnson, 1951 and other eastern cultures were of greater antiquity ) indicated that Mississippi valley than had previously been [Vor. 40 108 ANNALS OF THE MISSOURI BOTANICAL GARDEN thought. Johnson (1951) reported that Hopewell sites from Ohio and Illinois were about 1000 years older than had previously been estimated. The whole chronology of the Southeast has become unsettled because of radiocarbon dating results, but the presence of cultures in this area at earlier dates further strengthens the idea that maize is of greater antiquity than had been thought possible, and opens a way to minimize a long-standing inconsistency between archaeological and botanical evidence. Since there was, as Ford and Willey (1941) and Waring and Holder (1945) pointed out, a suggestion of strong Mexican influence in Temple Mound I and П cultures, the idea would be greatly substantiated should remains of Mexican maize ultimately be found in excavations. Botanical evidence exists that maize of the Southeast is related to a Mexican Gulf Coast race (Wellhausen e£ al, 1951, 1952), and hybridization of this Mexican maize with eastern flint maize already present in the Southeast probably resulted in the variable forms called Southern Dents by Brown and Anderson (1948). Northern Flint and Southern Dent maizes were subsequently brought together to form modern hybrid corn-belt maize (Anderson and Brown, 1952a, 1952b). A significantly large collection of Southern Dents was not available for inclusion in the present study, but the few available cobs which were studied showed a rather close resemblance to Northern Flints in cob characters. Evidence that many Southern Dents were intermixed with Northern Flints was presented by Brown (1949), who studied chromosome knob numbers in United States maize. He found Northern Flints to have the lowest knob numbers, Southern Dents the highest, and corn-belt forms intermediate between these extremes. Longley (1938) had previously surveyed Indian maize varieties from the United States and northern Mexico. He reported the same general rise in knob number among southeastern varieties. Не also found that maize of Arizona and New Mexico was high in knob number. Carter (1949) used knob numbers on chromosomes of Indian maize to indicate tribal affinities and differences. The fact that such an investigation yielded results substantially the same as those of different approaches is another example of how closely maize mirrors the history of those with whom it is associated. Ап important point regarding use of Northern Flint and Southern Dent maizes in corn-belt maize production is that the hybrid vigor manifest in this cross is based upon small differences between two already-intermixed races. This situation may possibly indicate that improvement of maize by hybridization techniques has barely begun. Since identification and classification of races are becoming à necessary part of maize breeding, a knowledge of cob variation should be of prac- tical value in the development of new strains of hybrid maize. The principles of cob analysis set forth above, even though they employed small morphological differences, may be considered fundamentally sound, because results obtained from their analysis were in agreement with ethnological and archaeological as well as botanical data of other workers. 1953] NICKERSON—VARIATION IN COB MORPHOLOGY 109 SUMMARY Morphological characters present in the female inflorescence, or cob, of Zea Mays L. have been measured on over 500 cobs representing both modern and archaeological varieties. External characters measured included row number, shank diameter, cob diameter, and over-all cob shape. Internal characters measured in- cluded cupule width, cupule depth, height of rachis-flaps, kernel thickness, lower glume width, rachis diameter, and pith diameter. Analysis of pictorialized scatter diagrams of averaged measurements showed a high degree of association of these characters. These results agreed closely with those of previous investigations employing other methods. The results of archaeo- logical maize analysis were in harmony with previous conclusions based on purely archaeological data. Such agreements indicate the validity of cob analysis for characterizing variation in races of maize. LITERATURE CITED Alava, Reino O. (1952). Spikele variation in Zea Mays L. Ann. Mo. Bot. Gard. —96. Am > en, a чуч pui ү ку 28 ehistoric southwesterners Кон basketmaker to етты 163 рр. uthw. Anderson, bani. (1956). x m жнын in American Tradescantias. Ann. Mo. Bot. Gard. 23:511- ‚ (1943). A variety of maize from the ча Loa. Ibid. 30:467—475. ‚ (1944a). Maiz Asiae Ibid. 31:301-315 , (1944b). Homolog the ear and tas ats Zea Mays. Ibid. 31:325—344. (1944c). Two pt saria of prehistoric corn tassels from southern Utah. Ibid. 31:345- > 354. — — —— , (1946). Maize in Mexico—a preliminary survey. Ibid. 33: و بر‎ — — (1947a). Field suiii rà E die a aize. Ibid. 34:433—4 —————, (1947b). Corn before Columbus. 24 pp. Pioneer Hi-Bred я Co; Des Moines. oft —— ——, (1949b). Introgressive hybridization. 109 pp. John Wiley & Sons, New York — — ——, (1951). The sacred plume. 24 pp. Pioneer Bre r , Des аем 19323 nts, man and life. 245 pp. Little, Bro ., Boston. » (In € s). Efficient and inefficient methods of measuring specific differen 3 O. Bla aes (1942). Prehistoric maize from Cafion del Монев Ка: Jour. Bot. 29: 832-835. , and William L. Babe (1948). A morphological analysis of row number i Mo. Bot. Gard. di п maize. Ann. ae MEE ini. y of corn belt maize and its genetic significance. Heterosis. РР. 124-148. 1 Y. Gowen, Ed. Iowa State Coll. Press, Ames. "o. The history of the common але varieties of the United States corn bele Авт H EER RE ugh c. “Cuter (en. Races of Zea Mays: Ann. n dm Gard. 29 , ор ст Gage MT Introgressive hybridizati I. Their recognition and classification. on in Phlox bifida. Amer. Jour. Bot. 39:399—4 SEMIS and D po. кона (1944). А method for recording and analyzing variation in в internode pattern t. Gard. 31:241-247. "PE em BSS. (19 52). PME "No нт er e U.S. Thé а em mes d С. (1948). А resp raisal of Peruvian carin А Атег. Агс Eon Antiquity 13). 128 pp. í Bird, ns i 1943). oer! in оя Chile. Amer. Mus. Nat. Hist., Anthrop. Papers а 94): America’s oldest farmers. Nat. Hist. 57:296— ^ Bonnett, "dd T. (1940). Dept of the staminate je td inflorescences of sweet corn. Jour. Agr. Res. 60:2 ; ТГ» S MB. Ear 2 ыз development in maize. Апп. Мо. Bot. Gard. 35:269-287. [Vor. 40 110 ANNALS OF THE MISSOURI BOTANICAL GARDEN Brown, William L. (1949). Sene and distribution of chromosome knobs in United States maize. Genetics 34:524—53 and Edgar AER т. 1947). The Northern Flint o Ann. Mo. Bot. Gard. 34:1-28, — — ——, (1948). im Southern Dent corns Md e 255—268. А E: G. ARA, and R Lo Je. (1952). Ви on three varieties of Hopi maize. Amer. Jour. Bot. p 597—6 Buck, Peter Н. (1938). Vikings of the sunrise. 33 tokes Co., New York. Burgh, weis F., and Chas. R. Scoggin (1948). The d of = Park, Dut National Mon ment. Univ. Colo. Studies, Ser. in Anthrop. 102 p s Gate F. (1945). Plant geography and СҢ Aie in eg American Southwest. Viking und Pub 14 ork. nthrop., No. 5. ; (1946). Origins of CSS Di. pee ге. Ап гор. 48 ‚ (1949). Чуку: implications of chromosome A of Nock kes Indian maize. Southwest Jour. Anthrop. 5:199—207. ‚ (1950). Plant evidente for early contacts with America. Ibid. 6:161—18 ; and Edgar mno dea А pre ED ci survey of maize in the usas United Ann. Mo. 297— Cutler E а (1946), Ёле of maize in South America. Leafl. Bot. Mus. Harvard Univ. "8 (т The geographic origin of maize. Chron. Bot. 12:167—169. ( А preliminary соге = аван remains of Tularosa Cave. In Fieldiana: Anthro- pology ( chess Nat. Hist. Mus.) 4 Esau, АЯ (1943). Ontogeny of ха е bundle in Zea Mays. Hilgardia 15:327-356. Ford, a ; and Gordon Willey 1941). An аена ир е of the prehistory of the eastern United States. Amer. Anthrop. 43:325— neta т. (1939). Über die Organstellung bei M aiskolben. ас Jour. Bot. 10:113-140. Hau mil W. (1945). Painted Cave, northeastern Arizona. 7 pp. Amerind Foundation, Inc. jos n, Ariz. x 1950). The stratigraphy and еее of Ventana Cave, Arizona. Univ. of Ariz. Tucson; Univ. of New Mex Albuquer Hurst, C. T. (1948). The Cotto ied Cum expedition, 1947— cave and pueblo site. South- wes — — ———, and Edgar ‘Anderson (1949). A corn cache from western Colorado. Amer. Antiquity 14:1 161- Johnson, F. O. os). Radiocarbon dating. Mem. Soc. Amer. Arch., No. 8. (Suppl. Amer. An- tiquity 17). Jones, ier H. 19), Maize from the Davis Site: its nature and Run ceqpoaterion, In: Newell and r, The George C. Davis S MN scias County, Texas. . Soc. Amer. Arch., No. 5 e Т Amer. Antiquity 14) :239 tole Sheldon А 19 p Arroyos. s te Amer. 1879:7 Kiesselbach, . (1949). The structure and undue. of corn. Univ. Neb. Coll. Agr. Bull. А уз Kroeber, А. Je (1939). Cultural and natural areas of native North America. Univ. Calif. Publ. in Amer. Arch. and Ethnol., No. 3 Lenz, m Wayne Rs 8). d histology of the female inflorescence of Zea Mays L. Ann. t. Gard. 35:353—376. Yide, Ralph (1924). beg significance of certain traits in North American maize cultures. Amer. nthrop Lister, Robert H., and He rbert ыы ох (1952). Archaeology of the Glade Park area—a progress : S Mon of maize from North American Indians. Jour. Agric. Res. wie; К obert H. (1940). American Се" history. Amer. Anthrop. 42:409—428 m . Gard 3 en 196. The role of pod corn іп the or rigin and evolution с maize, Ann. ar 1377. — de noe The origin of Indian corn and its relatives. Texas Agr. Exp- Sta. Bull. 574. 315 ‚ and C. Earle Sith; Jr. (1949). мы archaeological evidence on evolution in maize . Bot. Mus . Harva Leal Univ. 13:213—24 Mattia, Раш $ ; John B Rinaldo, Elaine ME uS Hugh C. Cutler, and Roger Grange, Jr. os Mogollon cultural саол and change. Fieldiana: Anthr opology (Chicago Nat. Hist. Mus.) 40:1— — H s and Alex D. Krieger ( 1949). The Geor А rge C. Davis ах (dero County, Texas. m. Soc. Amer. Arch., No. 5. (S uppl. Amer. Antiquity 14). 256 Ў 1953] ; NICKERSON—VARIATION IN COB MORPHOLOGY 111 Nickerson, Norton H. (In press). Morphological analysis of the maize Prat, nt (1948). General features of the epidermis in Zea Mays. pet "мо. Bot. Gard. 35:341- Randolph, І. Е. (1952). New evidence on the origin of maize. Amer. Nat. 86:193—202, Richards, Annette H. (1953). The myst ed of the first Midian. Nat. Hist. 62:168—174. Sauer, Sek = мне ааа origins and dispersals. Bowman Memorial Lectures. Ser. II. New Yor fub. G. are erg kx (1952). “The кыр of cultivated plants and weeds, Evolution 6:445- Stonor, С. К., апа га Anderson (1949). Maize among the hill peoples of Assam. Ann. Mo. Bot. Gard. 36:355—4 ошл» E. L. (18 WI Varieties of corn, U. S. D r. Off. Exp. Sta, Bull, 57:1 Р and Preston sen СЭ А уча, poiius complex in de: ive dudo nited States. Amer. Ant 7:1— Vidua. Paul (1918). The evolution г maize. Bull. €— Bot. Club 45:309-342. — ———2, (1920). А misconception as to the structure of the of maize. Ibid. 47:359-362. ‚ (1936). The | ра of the maize poe and maize и in ancient America. Univ. New Mex., pte p. Ser. I, No. 5:11-18. Wellhausen, Е. J., L m rts, and E. нгаб X. (1951). Razas y^ maiz en México. Ofic. tud. E 299ге. 9 с" Agric, y iut. Foll. Tech. 5. 237 pp. Mexico ———, (1 952). Races of maize in Mexico. т М, Harvard Univ. Whicker, "t. W. (1948). Lees a pre-Columbian cultivated plant in the Americas. South- ern Jour. Anthrop. 4:49—6 id Junius Bird —— Identification and (à: of the cucurbit materials from a Prieta, Peru. us. t. 1426. 15 New York, Whine Alfred F. ur hs ОНЫН a corn: n evaluation А p and theory. Amer. Anthrop. Wiser, Cark, Саму Corn and early American civilization. Nat. Hist. 54:5 Worm fà: M. (1949). Ancient man in North America. Denver Mus. КС Не, Рор. $ег. vus i" THE CYTOLOGY, MORPHOLOGY, AND SYSTEMATIC RELATIONSHIPS OF DELPHINIUM x; BELLADONNA HORT. EX BERGM.* MARILYN AMY GAGE I INTRODUCTION Among the cultivated perennial Delphiniums the problem of the origin of the forms variously termed "garland larkspurs" (Bailey, 1930) or Belladonna types has been raised more than once. The group as it is known at present includes not only Delphinium X Belladonna Hort. ex Bergm., but at least the 45 variants listed by the Royal Horticultural Society (1949), and probably certain other hybrids of unknown origin some of which are no longer in cultivation but were recorded in the literature of the last century. Two main theories have been proposed regarding the nature of the assemblage. (1) L. H. Bailey, following the example of Huth (1895), the last monographer of the genus, considered these plants to be forms of D. cheilanthum Fisch. var. formosum; his grounds are wholly morphological (2) Lawrence (1936), on the other hand, has suggested that D. Belladomna originated in gardens as a triploid hybrid between a tetraploid species and a diploid species, and that subsequent doubling of the chromosome number resulted in the fertile line now in cultiva- tion. His hypothesis is in line with the known facts, for until early in the present century Belladonna was quite sterile and was propagated wholly vegetatively; moreover, the present fertile strain is a hexaploid. In studying the origin of cultigens the historical method alone is obviously in- adequate. Students of cultivated plants are well acquainted with the hazards of attempting to place even well-known species into conventional taxonomic cate- gories. Even with such a relatively recent variety as D. Belladonna, caution is necessary in considering historical references, for it often happens that opinions given in early descriptions are contradictory or suggest some parentage which, in the light of present knowledge of cytology and genetics, is far from the truth. One can accept, with caution, species listed in cultivation at a certain period, but there must always be the reservation that morphologically similar species are sub- ject to confusion, and that any of these species may, and probably do, differ from their wild relatives as a result of horticultural “amelioration.” Morphological comparisons with known species have often been used in estab- lishing the possible parentage of plants of unknown origin. This method is most useful in groups where there are clear-cut characters whose extreme and inter- mediate expressions are easily recognized. The genus Delphinium is, however, notoriously variable, and specific distinctions are based to a very great extent upon quantitative differences, such as width of leaf segments, degree of dissection of чел ap i ; haw School of Botany «t al tt Ter erg ا‎ варце. ко of the requirements for the degree of Doctor of Philosophy. During the last year the study was carried out under a pre- doctoral fellowship of the National Science Foundation. (113) [Vor. 40 114 ANNALS OF THE MISSOURI BOTANICAL GARDEN the leaves, extent and type of axial branching, degree of incision of the "bee", length of spur, and the like. Every one of these characters is subject not only to genetical, but also, to a high degree, to environmental variability. Therefore, before employing such characters, it is important to understand to what extent each character may be modified by the environment. A third method of general use in establishing relationships between species, that of cytological analysis, is often more reliable. А number of investigators have shown that degree of relationship between species may often be estimated by studying the morphology and especially the pairing affinities of the chromosomes (see Stebbins, 1950, for a review of the literature). However, where the chromo- somes of the parental genomes are not strongly differentiated there must often remain doubt as to the sources of the chromosomes which are pairing in the hybrid. Especially in polyploids above the tetraploid level such problems are intensified by the presence in the same individual of several genomes, so that autopolyploidy and allopolyploidy may coexist. The relationships of D. Belladonna must therefore be considered from a com- bination of several standpoints, and the solution must be based upon that hypothesis which best makes use of the information obtained by each method. Of the materials employed in this study, families of various species and hybrids were grown in the experimental greenhouses of the Missouri Botanical Garden and of Washington University, in St. Louis. Herbarium specimens were made available by the Bailey Hortorium of Cornell University and from the collections of the Missouri Botanical Garden. I wish to acknowledge my indebtedness to Dr. G. A. L. Mehl quist, of the University of Connecticut, formerly Research Horticulturist at the Missouri Botanical Garden, who suggested this problem and who stimulated my interest in it by his own enthusiasm. I am also indebted to Dr. Edgar Anderson and other members of the staff of the Missouri Botanical Garden who have aided me in many ways. IL. TAXONOMY AND MORPHOLOGY A. NATURAL SPECIES History of the Genus.— Since the establishment of the genus in its modern form Бу Tournefort (1700), Delpbinium has received various treatments. In his 'Species Plantarum' Linnaeus listed the six species, D. consolida, D. ajacis, D. peregrinum, D. grandiflorum, D. elatum, and D. Staphisagria. By the close of the eighteenth century only six more species had been added, but in the century and a half since then, the number of species described has increased to more than 300. A. P. DeCandolle was the author of 23 species, and in his ‘Prodromus (1824) he grouped 53 species into the four sections, CONSOLIDA, DELPHINELLUM, DELPHINASTRUM, and STAPHISAGRIA. Prantl, in Engler and Prantl's *Die Natürlichen Pflanzenfamilien’ (1888) merged DeCandolle's srAPHIsAGRIA and DELPHINASTRUM into one section, STAPHISAGRIA; 1953] GAGE—DELPHINIUM Х BELLADONNA HORT. 115 and later, modifications were made by others—notably Asa Gray (1887), Robin- son and Fernald (1908), and Huth (1895). The most recent treatment of the American species is that of Ewan (1945). Huth published in 1895 his "Monographie der Gattung Delphinium,” in which he attempted a natural classification of the 200 odd species treated. Не recognizes two subgenera: CONSOLIDA, in which there is a single carpel, the petals are fused into one, and seeds are scaled and three-cornered; and EUDELPHINIUM, in which there is more than one carpel and the four petals are free. Within the second sub- genus he distinguishes three sections: ELATOPsis—petals deep violet or black, lower petals bifid and bearded. DIEDROPETALA—petals light, the same color as the sepals or dirty yellow, deeply ifid, lobes acute. KOLOBOPETALA—color of petals as in DIEDROPETALA; limb of lower petals round or rectangular, entire or bilobed, lobes round or truncate at tip. Of the species considered here, D. elatum L. belongs to the series КАСЕМОЗА of section ELATOPSIS, and D. grandiflorum L., D. cheilanthum Fisch. ex DC., and D. tatsienense Franch. are included in the series CHEILANTHOIDEA of section KOLOBOPETALA. While D. Belladonna Hort. was not known to Huth, he in- cluded as varieties of D. cheilanthum several cultigens which are certainly very similar to it. Further consideration of the taxonomic categories will be restricted to D. cheilanthum, since it alone has been confused with D. Belladonna. Тһе other species (D. elatum, D. grandiflorum, and D. tatsienense) are taxonomically well- known, and while they include many natural and horticultural variants, there is no problem in determining the affinity of such variants. Delphinium cheilanthum Fisch. ex DC.— Apart from the question of any direct relationship between D. Belladonna and D. cheilanthum, a problem which will be discussed later, the two species are con- used in herbaria and in gardens, where identifications must be made largely upon a morphological basis. The magnitude of the problem is amply brought out by L. H. Bailey in ‘The Garden of Larkspurs’ (1939), in which he attempts to trace the history of D. Belladonna as a derivative of D. cheilanthum. He bases his con- clusions upon a comparison of certain historic plates and descriptions of D. cheil- anthum and of certain horticultural varieties (D. formosum Hort. and D. Hender- soni Hort.). The last two will be taken up in the next section, and I shall discuss here only D. cheilanthum in order to clarify its position, since it is apparently infrequent in herbaria in this country. Delphinium cheilanthum Fisch. ex DC. is based upon a specimen sent to A. P. De Candolle (1818) by Fischer from the region of Doroninsk in Dahuria (in тч терт 8А „Йи: hat cheilanthum does not appear as a native plant in the great А ыгы is ek I "aee эе ана Сы шей cited by Huth mostly are plants considered by him to be botanical varieties of D. cheilanthum.” [Vor. 40 116 ANNALS OF THE MISSOURI BOTANICAL GARDEN trans-Baikal Siberia). In 1819 von Schrank figured and described the same plant (under the name of D. cheilanthes) in ‘Plantae Rariores Horti Academici Mona- censis’ (pl. 7). Earlier, іп 1769, J. С. Gmelin had described and illustrated in his ‘Flora Sibirica’ a plant which, according to De Candolle, Fischer had indicated was the same as his D. cheilanthum. In 1820 a plant grown near London from seed sent by Fischer was figured in the ‘Botanical Register.’ In addition to the sources cited by Bailey, a number of accounts of the species given by collectors in Siberia and northern China (Ledebour, 1842; Regel, 1861; Trautvetter, 1847, 1877а, 1877b; Turczaninow, 1842; Glehn, 1876; Brühl, 1896; and Komarov and Schischkin, 1937) delimit D. cheilanthum Fisch. as a species or species complex, with well-defined characters and a distinct geographic distribu- tion. From these accounts, and the Huth monograph, the species may be character- ized as follows: DELPHINIUM CHEILANTHUM Fisch., in DC. Prodr. 1:53. 1824.—Stem tall, simple or branched; leaves glabrous or pubescent (especially on the under surfaces); 5-parted, parts oblong to narrowly lanceolate-acuminate; lower bracts many- parted. Inflorescence racemose to subcorymbose, peduncles bibracteolate; upper bractlets equal to spur. Flowers blue, rarely white; spur straight or slightly curved; sepals ovate, pubescent or puberulent externally; upper petals glabrous, pale yellow or blue, lower petals large, with ovate or subrotund limb, entire or rarely emarginate-bilobed, with yellow pubescence above. Carpels 3, glabrous or pubescent, tips curved at maturity, to 20—25 mm. long; seeds 3-angled, with winged margins, faces scarcely squamate. Flowers in June and July (pl. 9). A number of varieties have been described, but undoubtedly any future monographer must rework the alliance. The exact limits of the species are still somewhat in question, since Huth describes Delphinium Middendorffii Trautv. as a distinct species, although in his natural key to the genus he lists it as а variety of D. cbeilantbum, distinguished by having the apices of the sepals lanceolate. Moreover, he notes: “D. Middendorffii scheint eine verkümmerte Form des Dd- phinium cbeilantbum mit verlängerten Kelchblattern zu sein." By 1877 Traut- vetter, who originally described D. Middendorffii (1847), conceded that it was but | another variety of cheilanthum, and added (translation) : with бы, Tli tects рона konali «low wem (oet Middendorfi) eid rome times a taller stem; and this latter form has flowers either larger (var. typica) or smaller | Ая parviflora). Indeed, sometimes var. Middendorffi has larger, and sometim imes smaller flowers, so that sometimes it approaches var. Руріса and sometimes var . parviflora. An American species, D. chamissonis Pritz. (from the Bering Sea region and the Yukon valley) is also apparently a close ally of D. cheilanthum. Comparing D. chamissonis and D. Middendorffii, Ewan (1945) says: “All of the more than 4 score of collections examined show so much variation in spur length and, indeed, flower size as a whole, as to render it difficult to draw any distinction between two species from these characters." 1953] GAGE—DELPHINIUM Х BELLADONNA HORT. li7 B. HYBRIDS OF THE BELLADONNA TYPE The history of the cultivated Delphiniums is extremely complex (see Wilde, 1931; Bailey, 1939). As early as 1778 D. elatum L. and another form listed as D. azureum (not D. azureum Michx., which was not described until 1803) were offered in the trade, and at least by 1824 D. grandiflorum L. (two varieties) was offered. Other species were rapidly introduced and by the middle of the century hybridization was being carried on by a number of French growers, chief among whom was Victor Lemoine, originator of most of the first varieties of perennial garden Delphiniums. English plantsmen were somewhat later in the field, but particularly since the 1870’s, when the first varieties of James Kelway were intro- duced, a large part of the development of new hybrids has been carried out in Britain. As nearly as can be ascertained, most of the present forms were derived from relatively few species. D. grandiflorum and D. elatum have chiefly been involved, although hybrids with other species, including D. tatsienense Franch. (Lemoine, 1914), D. cardinale Hook. (Wilde, 1931), and D. mudicaule Torr. and Gray (Lawrence, 1936) have enjoyed some popularity. Horticultural literature has frequently not been reliable regarding species in cultivation, and allies of D. grandiflorum or D. elatum could have been confused with these species and so have been employed in crosses at a time antedating any existing records of their use. As has been mentioned above, D. Belladonna first appeared some time after the middle of the last century; the earliest known mention of its name is in 'Flori- cultural Cabinet and Florist’s Magazine’ for 1857, where it is listed as а desirable form with large pale lavender-blue flowers. In the 1865 catalogue of James Backhouse & Son, of York, England, Belladonna is listed and described as "a lovely turquoise, perfectly hardy." There are sporadic references to the plant in various horticultural publications before 1880, in which year the English firm of Kelway introduced it to a wide market. For a long time Belladonna was reported to be sterile. However, in 1902 or 1903 a plant grown by G. Gibson, of Leeming Bar, yielded three pods, whose seeds produced five plants, two of which became the named varieties “Mrs. С. Gibson," and "Grandiflora" (Phillips, 1949); and in 1905 Sutton & Sons, of Reading, England, obtained seed from two plants. Thus it appears that a chromosome-doubling occurred at least twice within a few year's time. The fertile strain has always been true-breeding, in the sense that there has been no recovery of anything which resembles any diploid or tetraploid species Which might be assumed to have been the original parents (although it will be demonstrated later that weak but measurable associations of characters present in such species are still to be found). It it true, nonetheless, that heterozygosity in these original amphidiploid forms must have been fairly high, judging from the large number of Belladonna varieties introduced within the next few years. It is of interest that the ‘Journal of the Royal Horticultural Society’ (1907) says of Sutton's first seedlings: "Some . . . resembled Delphinium Belladonna very closely, but some more nearly approached Delphinium formosum, and one bore [Vor. 40 118 ANNALS OF THE MISSOURI BOTANICAL GARDEN flowers of a very beautiful deep blue tint." D. formosum referred to here is the horticultural form and not the species of Boissier and Huet. Like Belladonna, it was a garden hybrid of unknown origin, and, also like Belladonna, it is said to be a hexaploid (Propach, 1939, 1940; Mehlquist, unpublished). It was first offered in 1855, and according to Van Houtte (1857) it was already fertile. This, coupled with its superficial resemblance to the diploid species D. cheilanthum, caused Huth to list it as a variety of that species in his monograph, although it has never been reported in nature (pl 8). It is suggested that formosum may have originated from an early and unrecorded doubling of the chromosome complement of Bella- he We Bee Se se ми. at at t-fig. 1. Leaf types im Delphinium species and hybrids (X и). D. чл аы var. chinensis Blue Butterfly. D. — var. chinensis 20 ег х м ^ D. tatsiene D. cbeilantbum ‚ from toners псн Ьу Каго in Dahur D. бейин. from plant grown from seed at the Macer Botanical Garden. 6-9. D. elatum рч range a vs Variation. 10-15. Belladonna types: 10, Lamartine; 11, Smith's Belladonna 16-20. Experimentally produced triploids: 16, 50-15-3; 17, 50-40-2; 18, 48-27-4; 19, 48-27-63 1953] GAGE—DELPHINIUM X BELLADONNA HORT. 119 donna, or at any rate that it originated in the same manner as Belladonna. Cer- tainly, there are no consistent differences between the two. The following is a general description of D. Belladonna: DELPHINIUM X BELLADONNA Hort. ex Bergm.?—Racemose open-flowered perennial with simple or branched stem up to 1.5-2.0 m. tall; herbage pubescent or glabrous; leaves palmate or deeply 3-parted and strongly ribbed, the parts again divided or lobed, the degree of dissection of the parts gradually increasing from rosette leaves to bracts, the main division of the stem leaves narrow-cuneate to narrow-oblong, 0.2—2.0 cm. broad at base, the petioles nearly or quite as long as the blade. Inflorescences of relatively short and open (5- to 20-flowered) racemes, the peduncles with parted bracts at base and the pedicels usually long, with basal simple bracts and paired short bracteoles immediately subtending the flowers. Flowers large, single, occasionally semi-double, mostly light to dark blue and somewhat declined; sepals thinly pubescent outside, blunt; spur about equalling the sepals and straight or somewhat curved at the end; petals mostly white or light-colored, bearded with yellow, making a large "bee" which fills the throat of the flower. Follicles 3, pubescent or glabrous, up to 2.0 cm. long, the apex curved; seeds 3-angled with somewhat winged margins, not squamate. Flowering period prolonged by the production of successive shoots during the growing season (pl. 8). Another variety of interest, Lamartine, originated with the French house of Lemoine & Son in 1903. Like Belladonna, it was at first sterile, and the present fertile strain had its origin in 1916-17 when doubling of the chromosome number apparently occurred among some plants growing at the Royal Moerheim Nurseries in Holland (Lawrence, 1936). The parentage of this variety is also obscure. In answer to my inquiry, the present head of the firm replied that they had no record of its antecedents, but that experimental work had suggested that the parents might have been an elatum type (sic, Pacific Giant) and D. grandiflorum. In habit, Lamartine is quite similar to Belladonna, being of the same height, with a loose, tapering spike, and fairly numerous lateral branch spikes. The leaves are similarly cut (text-fig. 1). The flowers are long-spurred, with sepals ultramarine suffused with purple, and the petals are dark-edged, shading to white, with a yellow beard. Like Belladonna, it has an extended blooming period. There are also grown at present two triploid varieties, Moerbeimi and Capri, which, while not directly related to Belladonna, have a close morphological simi- larity to that variety. The most complete and authoritative account of triploids is that given by B. Ruys of the Royal Moerheim Nurseries (1911): *D. Belladonna Hort. is equivalent to D. Belladonna Bergm., listed in ‘Index кин. Me mans, who published his binomial in 1925, is among those who believe the plant to have originat from a cross between elatum and cheilanthum. Не lists no sources for his opinion. [Уот. 40 120 ANNALS OF THE MISSOURI BOTANICAL GARDEN all the light-colored novelties which I could obtain, with a view to crossing them with large- ligh str itution . .. ew years ago] I found in a batch of seedlings one plant with five spikes. Of these, two spikes bore pure white flowers, two bore blue flowers, and some parti-colored, half-blue, half-white flowers. Next year, when the five divided plants flowered, I noticed that two had only white flowers, two only blue flowers, and one plant had some flowers white and others blue, whilst still others were half white and half blue." Both the white and the blue variety were propagated and marketed, the former under the name of Moerbeimi, and the latter as Capri. Aside from the color, there is no single definitive character to separate these and Belladonna, although, as mentioned above, these varieties are triploid. In addition to the present-day varieties of the Belladonna type, a study of the literature reveals that other hybrids of similar appearance were known in the last century, though they have now passed from cultivation. Because of their bearing upon the possible constitution of D. Belladonna, some of these early hybrids are described below: The earliest recorded Delpbinium hybrid was Barlowii, figured by Lindley in the “Botanical Register’ of 1837. The plant, from a nursery in Tooting, had been received from an establishment in Manchester, and was believed to be the result of a chance cross between D. grandiflorum L. and D. elatum L. It was reported to be a continuous bloomer. ‘Paxton’s Magazine’ described it the following year: “Plant perennial, growing usually from 4—5 feet high. Leaves with 5 principal divisions, deeply lobed and jagged; serrations acute; deep green on surface, light green beneath, smooth. Flowers semi-double, of a most intense blue color, produced very numerously in terminal spikes. Sepals of the calyx greenish externally.” Neither the description nor the Lindley plate, which shows a rather dense spike of semi-double blooms and, in the background, part of a leaf in outline, is com- plete enough to indicate the affinities of this plant. Moreover, except that the flowers are semi-double in each case, the Lindley plate is quite different from the Paxton plate, which indicates that the plant had a brownish bee, and leaves more deeply incised than those in the earlier plate. Since both plates were drawn from living plants, one is confronted with the alternatives that Barlowii was a highly variable variety, or that the two artists had available two quite different varieties. The plant in the earlier illustration might easily belong to the group of hybrids which includes Belladonna; the later plate appears, rather, to be merely that of an elatum variety. Herbarium specimens of plants sold as Barlowii within the last 20 years or so are quite different from either of these, being single dark-blue types, and not different morphologically from D. Belladonna. The next form of interest, D. Hendersonii Hort. (pl. 7) is of extremely dubious background. It was first mentioned by Moore and Ayres in the ‘Gardeners’ Magazine of Botany’ (1850) and was described under the name D. cheilanthum var. Hendersonii A. Henfr.: 1953] GAGE—DELPHINIUM Х BELLADONNA HORT. 121 “Hardy herbaceous perennial with large showy rather distant flowers. Leaves five- parted, lobes oblong or acuminated, trifid or obscurely bi-trifid, 4 in diameter, on long cemes inches el broadly linear, acuminated, simple. Racem axillary and terminal, lax; the pedicels larger than the Flowers large and showy, ns a little deeper; two lower petals with a roundish ovate limb, whitish in the middle, bearing a yellow beard, slightly irregular on * the margins, with a few ciliary hairs at the apex. Ovaries 3, glabrous, green, veined with h lines. Wood, from seedlings purchased from M. Chauviére of Paris var. Cbauvieri. Lemaire regarded it as merely a variant of the species, whose dis- tribution and habitat he considered briefly. ite in opposition to Lemaire, Harrison (1853), describing the same plant, stated: “It was raised by M. Chauviére, nurseryman of Paris, from D. chinense, impregnated by D. elatum-splendens.” This author also noted that Hendersonii was a profuse bloomer, of bushy habit, flowering over a long period of time. Hybrids of known parentage.— During the past eight years there have been produced at the Missouri Botanical Garden a number of hybrid Delphiniums of particular interest. Several hundre pollinations were made between D. elatum and different strains of D. cheilanthum and D. grandiflorum var. chinense, in an effort to find out which, if either, of these hybrids more closely resembled the Belladonna assemblage. Up to this time there have been no successful pollinations of D. elatum by D. cheilanthum. How- ever, the other cross (that between elatum and grandiflorum) has been successful four times, yielding two families of 3 plants each, one family of 2 plants, and one family of 13 plants. 1. 47-98-3 X D. grandiflorum var. chinense “Blue Butterfly”.—Hybrid Nos. 48-27-1 through 13. (This and the following cross were made by Dr. С. A. L. Mehlquist.) The seed parent was one of a line of white semi-double elatum forms developed by Dr. G. A. L. Mehlquist. The group is exceptional in being the only line of Whites in which there has been an approach to homozygosity without concomitant Weakening and loss of the line. 47-98-3 had dense spikes borne on the rather short main axes, and leaves of medium size, pubescent, and palmately divided, with relatively narrow segments ( pl 10). The pollen parent, on the other hand, was a low-growing plant with a branched, very open inflorescence of fairly large rich deep blue single flowers in few-flowered panicles. The leaves were relatively small, dull green, glabrous, and finely dissected. This form, too, was true-breeding, and the variety has been grown at least since 1910 when it was among the perennials planted at Wisley (Jour. Roy. Hort. Soc., 1910) (pl 9 and text-fig. 1). The Progeny of this cross included 13 individuals which were generally of intermediate bit and varied very little among themselves. The following description serves to characterize them and point out their chief differences from the parents. [Vor. 40 122 ANNALS OF THE MISSOURI BOTANICAL GARDEN SUMMER SKIES X 47031-1 50-15-4 X SMITH'S BELLADONNA ا ~. - SMITH. X se к X SELF 51- 213-2 X SELF изз X SELF 51-217 51-216 51-218 52-224 48-2-3 X BLUE BUTTERFLY 41-98-3 X BLUE BUTTERFLY 48-6-3 X SMITH'S BELLADONNA 48-27-5 X SMITH'S BELLADONNA SMITH X 50-20-1 X SELF SMITH X 51-209-1 X SELF 51-201-1,2 T 202- Y TO 14 51-223 51-222 SMITH х 51-202-5 X SELF 51-202-1 X 50-20-1 (51-220 51-219 51-221 Text-fig. 2. Pedigrees of major lines of triploids and hexaploids. 48-27- Series.—Racemose open-flowered perennials, with simple or usually branched stem up to 1.0—1.25 meters tall; herbage pubescent. Leaves palmate and deeply 5- or 3-parted and strongly ribbed, the parts again divided or lobed, the degree of dissection increasing progressively from the rosette to the bracts, the main divisions chiefly narrow-cuneate, up to about 1.0 cm. broad, the petioles 45 long as or slightly longer than the blade. Inflorescences not spiciform, the terminal one a 15- to 20-flowered raceme, the lateral of fewer flowers, the peduncles with 3-partite bracts at base and the pedicels with basal simple bracts, and pair bracteoles 0.8—1.0 cm. long subtending the flower. Flowers medium-sized, single, violet and cobalt, on long pedicels and somewhat declined; sepals thinly pubescent outside, blunt, the spur slightly longer than the limb and usually somewhat curved at the end; petals the same color as the sepals, bearded with short golden yellow hairs. Follicles 3, pubescent, not developing into fruits, apex curved. Flowering period prolonged by the production of successive shoots during the growing season (pl. 11 and text-fig. 1). 2. 48-2-3 X Blue Butterfly.—Hybrid Nos. 48-6-1 through 48-6-3. 48-2-3, the seed parent of this group, was a selection from Vetterle and Reinelt's 1953] GAGE—DELPHINIUM Х BELLADONNA HORT. 123 Galahad Series of their Pacific Giant Hybrids. The plant, a white semi-double, was not available to me. The three hybrid individuals in this group were quite similar to the hybrids described above, the chief differences being the regular production of one or several additional petals of either sort, and the generally larger, less open flowers. 3. Summer Skies X D. grandiflorum (47031-1).—Hybrid Nos. 50-15-1,-3,-4. The seed parent of this cross is a selected seedling from the Pacific Giant series, Summer Skies, an elatum type with semi-double flowers with sky-blue sepals and white petals in tapering, somewhat dense spikes (pl. 12). It has been found to be true-breeding for color (Mehlquist, unpublished). The pollen parent was from seed obtained from the Royal Botanic Gardens at Kew as D. tatsienense. How- ever, determination of the plant, using Huth's key, showed that it must have been, rather, a corymbose variety of D. grandiflorum. Numerous attempts were made to obtain seed, but 47031-1 was completely seed-sterile. The three hybrid individuals of this cross differ in a number of particulars from those of the other crosses, being taller, with a strong main axis which bears lateral spikes in the manner of D. elatum, but these more profuse and more strongly developed than in that species. Leaves of 50-15-3 are much like those of e 48-27- series, whereas those of the other two plants are considerably larger and have the ultimate divisions rounded rather than acute. The flowers are again mixed blue and violet (although the blue of these plants is somewhat more brilliant than in the other crosses) and are more or less semi-double, with spurs about № longer than the standard of the sepal (pl. 12, and text-fig. 1). 4. 47-98-3 X 47031-I.—Hybrid Nos. 50-40-1 and 2. The seed parent was the white elatum used in producing the triploids of line 48-27; the pollen parent was also used to produce the triploids of line 50-15. The two individuals of this cross resembled those of 48-27, except that the leaves were more finely dissected, and the flower was the same brilliant cobalt noted in the 50-15 individuals (text-fig. 1). 5. Derived bexaploid bybrids.— Since the fertile strain of D. Belladonna apparently originated from spikes Whose chromosome numbers had been fortuitously doubled, it was hoped that it might be possible to produce doubling of the chromosomes of the hybrids described above. With this in view the developing rosettes were treated each year for three Years with aqueous colchicine in various concentrations for varying lengths of time. In at least one case, doubling of the somatic chromosome number of the Spike was achieved, judging not only from morphological features but from greatly increased amounts of "viable" pollen produced (Table I), but the 14 seeds ob- tained did not germinate. The other method employed, that of repeated selfing сыны hether delayed planting or some undetermi genetic factors may have been wn whether delayed planting | е ` anger ын ш Gi de Н Lac ТЫР hans ры may not germinate at all after six months, even in cold storage. [Vor. 40 124 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE I POLLEN SIZE AND VIABILITY IN SPECIES AND HYBRIDS OF DELPHINIUM: LIVING PLANTS. s Pollen diameter in д Plant % bius ble | S | Range Median Diploids D. grandiflorum Blue Butterfly 97.0 18.8—37.6 233 47031-1 65.0 18.8—28.2 23.5 Tom Thumb 91.0 23.5—37.5 28.2 D. cheilanthum 98.0 23.5—32.9 28.2 Tetraploids 47-98-3 64.6 23.5—37.5 30.6 Summer Skies 32.0 23.5—37.6 30.6 Triploids 48-6-1 p Sr d 28.2—37.5 32.9 48-6-2 7.0 23.5—40.0 32.9 48-27-1 16.0 25.9—40.0 30.6 48-27-3 14.0 23.5—37.6 324.9 48-27 21.2—47.0 33,3 48-27-5 21.0 23.5—42, 32:9 48-27-6 E 23.5—37.6 35.3 48-27-7 9.0 23.5—36.5 Е 48-27-8 4.0 23.5—42.3 2 48-27-10 16.3 21.2—51.7 37.6 48-27-11 5.0 23.5—37 0. 48-27-12 23 23.5—47.0 32.9 50-15-1 1:5 50-15-3 0.0 50-15-4 2,5 5 - treated) 2.0 21.2—37.6 30.6 50-40-2 (treated) 52.0 25.9-37.6 30.6 Moerheimi 0.5 те = ЫЕ Hexaploids OERE E E EE Smith's Belladonna 41.0—95.0 23.5—42.5 32.9 Cliveden Beauty -1 68.2 £ Б 37.6 -5 1 80.5 25.9—37.6 32:9 Lamartine 82.5 23.5—32.9 30.6 ел of the triploids, proved even less satisfactory, only two inviable seeds being produced from upwards of a thousand individual pollinations made in several seasons. However, pollinations of the triploids by Belladonna proved more suc- cessful: eventually ten plants, representing five crosses, were obtained, and of these, five plants (representing three crosses) were used in establishing new hexa- ploid lines, both by selfing and by backcrossing to Belladonna. A detailed descrip- . tion of the cytology of these hybrids follows (chapter IV), but it is to be no that in morphological characters and cytological details they do not deviate from 1953] GAGE—DELPHINIUM Х BELLADONNA HORT. 125 the Belladonna category except in being more variable; and in some cases fertility is already as high as in Belladonna. C. DISCUSSION Two major points are to be brought out here: First, while it has been demon- strated that there is indeed a resemblance between the Asiatic diploid species D. cheilanthum Fisch. ex DC. and the hexaploid cultigen, D. Belladonna, there are also important differences. The botanical species and its allies are restricted to Siberia, Mongolia, and North China, and to the islands of the Bering Sea and the Yukon, and it has not often been planted in gardens. Moreover, when one comes to compare the two entities, the first impression of similarity is modified by an appreciation of their subtle differences. Thus, the leaf of D. cheilanthum may often be incised so that the major divisions are of about the same width and degree of segmentation as in the Belladonnas, yet the latter will invariably have the ultimate divisions more acuminate, the segments often intricate, and the main lateral veins divergent in their departure, features rarely present in D. cheilanthum, and certainly never present in combination in that species. The flowers of D. cheilanthum are small but always have an extremely conspicuous bee; while the bee of Belladonna may be of the same absolute size, it is always smaller in propor- tion to the size of the flower. Lastly (and it is realized that the botanist will not often have access to the quantity of material necessary to make this distinction), if one has on hand a number of plants of both sorts, it is a relatively simple matter to separate the Belladonnas from D. cheilanthum, since the latter holds together by virtue of concordant variation, and the former is less cohesive on account of discordant variation (Anderson, 1951). The second point brought out in this section is that Belladonna types have originated a number of times in the past, and always they have first been noted in gardens. From the evidence, it seems clear that Hendersoni, Lamartine, Moer- beimi, and the others, while not directly related, are quite similar in leaf shape, flowers, branching habit, and prolonged blooming periods. Several of them re- produce by seeds; others have been completely sterile. In no case has the exact origin of the plant been unquestioned. However, all of them have been found to be very much like certain hybrids produced by crossing Del phinium elatum and D. grandiflorum varieties. In the absence of direct evidence one cannot 1mme- diately exclude D. cheilanthum as one of the parents, but evidence against such a supposition will be given in the following pages. Ш. CHARACTER ASSOCIATION AS AN AID IN EsTABLISHING RELATIONSHIPS _ At least since the time of Gmelin, botanists faced with the necessity of classify- ing species of Delphinium have realized the difficulty of finding reliable specific characters. This is the result not so much of lack of genetic differentiation within the genus, as of the presence of numerous characters of a quantitative nature, [Vor. 40 126 ANNALS OF THE MISSOURI BOTANICAL GARDEN whose genetic basis may be modified or simulated by environmental factors. On account of this "inability to use one or two differentiae to distinguish species or subspecies," Ewan (1945) has based his treatment of the North American species on "maximum correlation of characters existing in combinations." A species characterized in such terms could be visualized, then, as a plexus of "typical" individuals, possessing the most pronounced correlation of the characters of the species, and a large number of individuals possessing these characters in lesser degree, but with no sharp discontinuities within the group. Such discontinuities, on the other hand, would be expected between species; and populations of hybri origin might be expected to be largely coherent, but with some segregation if the parental genomes have sufficient segments in common that occasional heterogenetic pairing occurs (chap. IV). The study of hybrid ancestry in such a group as this would not long ago have been considered a very difficult one from which to arrive at any very definite conclusions. However, the student of variation has been given a valuable tool in the graphic methods developed by Anderson (Anderson, 1948, '51, '52; Anderson and Gage, 1952), which permit the study of complex patterns of variation on the two-dimensional level of the scatter diagram and ideograph. Аз Stebbins (1952) has recently pointed out, the method of extrapolated correlates involves more than the mere random choice of characters; rather, the choice of a measure is determined by how well it expresses a feature of the plant, and how subject that feature is to environmental variation. While the methods are thus a compromise between pure statistical analysis and subjective intuition, their success in a number of cases indicates their great usefulness. Before employing these methods in Delphinium, it was necessary to study | thoroughly within-a-plant variation. Good material for such a study was avail- able in pressed material, and in living plants of several vegetatively propagated individuals, chiefly hybrids of the 48-27 and 50-40 series and Smith’s Belladonna. Preliminary measurements of a number of characters demonstrated what was already suspected, that only a few features of the leaf could be used. (Leaf variation in Delphinium has been the subject of various investigations: Brown, 1944; Lewis, 1947; Ashby, 1948.) Ву trial and error it was eventually possible to score a number of characters which exhibited low variation even on divisions of a plant grown in different years. In addition to the few leaf characters, those characters most effective for measurement were those of the flowers and inflore- - cence. (Certain definitive characters, such as internode pattern, mode of branch- ing, etc. could not be employed because of the scanty amount of herbarium material) With these measurable characters as guides, the three natural specie (D. elatum, D. grandiflorum, and D. cheilanthum) and Belladonna were studi and those features which could be used alone or in combinations as differential characters were selected for use in separating the units on а multiple-charactet basis. The characters used were the following: 1953] GAGE—DELPHINIUM X BELLADONNA HORT. 127 W Text-fig. 3. Floral and leaf characters used in constructing scatter diagrams. Leaf.— Length of median segment/length of C (text-fig. 3). à; аге about equally dependent on absolute size of leaf, their ratio expresses fairly well the depth of cutting of the segments. Width of median segment.— This was found to be the best measure expressing what may degree of dissection of the segments. Inflore М TECE Number of flowers. The absolute number of flowers on the main axis moe ari Density. This was computed as the length of the inflorescence di noe (measured from the attachment of the lowest flower on the main axis) divided Ы; the number of flowers. Margin of Бее. Scored in three grades: bifid, notched, and entire or emarginate. Length of standard divided by length of limb of the bee (text-fig. 3). Since both measures be termed the [Vor. 40 128 ANNALS OF THE MISSOURI BOTANICAL GARDEN In laying out the scatter diagram (text-fig. 4), the ratios used as ordinate and abscissa were selected because, varying widely among the plants measured and expressed as individual grades, they permit a wide scattering of the points on the diagram. Moreover, the use of measures of one floral character and one leaf char- acter lessens the possibility that they might be merely different manifestations of the same factor. The scoring of the other characters was then adjusted so that each could be expressed in three grades, to which ray lengths and positions were assigned (further details accompany text-fig. 5). The "population" represented by the scatter diagram consists of herbarium specimens from the Missouri Botanical Garden and from the large collection of cultivated Delphiniums at the Bailey Hortorium of Cornell University. Extreme types such as double varieties and dwarf forms were excluded. АП other measur- able sheets were included in making the diagrams. Since pollen measurements made from herbarium specimens were found to be reliable indicators of approximate polyploid level (chap. IV), an additional measure, determined by pollen grain size, is expressed by the size of the dots on the scatter- diagram, the smallest dots representing diploids, the intermediate dots, tetraploids, and the large dots, hexaploids. A survey of the individuals thus diagrammed shows two distinct entities, D. grandiflorum in the lower left-hand corner, and D. elatum in the upper right- hand corner. Occupying the area between these species is the Belladonna complex, more variable than either but generally intermediate between the diploid and the tetraploid species. The position of D. cheilanthum just off the spindle but adjacent to Belladonna points out its notorious superficial similarity to Belladonna, but its actual physical distinctness is shown first by its diploid nature, and second by its almost "all or none" character with respect to the rays. It is seen to be not ай intermediate between grandiflorum and elatum, but, rather, it possesses some char- acters in common with grandiflorum and some with elatum. (The plants of cheilanthum scored included two collections of two plants each, as well as tw? plants growing in the greenhouse at the Missouri Botanical Garden, and so shows more grouping of the dots than would a random selection.) A graphic analysis of another sort, the internode diagram (text-fig. 6); points out clearly the differences in habit of the entities studied. Both D. elatum and D. cheilanthum have a relatively large number of nodes before the axis begins ® branch; these increase somewhat in length from the highly shortened internodes of the rosette (not illustrated) but do not increase in a regular pattern and afè generally not more than 7—8 cm. long. For D. elatum this was true not only of the cultigens, but of the wild species as well, where even dwarf alpine plants have a high node number, only the length of the internodes being shortened. In branch- ing pattern these species may also be compared. On the whole (the exception being some garden types which are bred for a diffuse habit) side branches are few and much shorter than the main raceme which they subtend, generally decreasing in length from the lowest to the uppermost. The most evident distinction betwee” 1953] GAGE—DELPHINIUM X BELLADONNA HORT. 129 у y з * | 3 d al Y ё ade Y n ees % ol o 5 bb 18 74 6 Y Уу 6 "T ee iv vf x w| у € 3 4. YOE BED 11.9.49 €- € $9 Text-fig. 4. Scatter diagram showing character correlations in Delphinium grandiflorum (small dots, rayless), D. cheilanthum (small dots, rayed), D. elatum (intermediate dots), and D. X Bella- donna Hort. (large dots). 18 Bee 36 @entire - emarginate M notch 2 bifid z 30 Width of median segment T ei- z” 3-7 9 2 $e v ш 2 B За Density of inflorescence = 1 @10-70 д” 649-40 = 0 ф39-10 £N 9-1 ә 48 Number of flowers Ew ei4 or 9 - 95-25 и. №26 ог more 12 10 зи оса LENGTH MEDIAN SEGMENT/LENGTH C lati ext-fig. 5. Scatter diagram showing character correlations in members of the 4 triploid popu- ations of this study. Distribution of D. Belladonna in outline. these species, on the basis of habit, is that the inflorescences of D. elatum are rela- ing successively shorter through the inflorescence. may branch very little, characteristically the plant is much-branched, the branches Sometimes being placed at such an angle and being of such a length as to have the superficial appearance of dichotomies. (This is especially true in the closely [Vor. 40 130 ANNALS OF THE MISSOURI BOTANICAL GARDEN related species, D. tatsienense.) The branches may again branch in a similar manner (not figured). The maximum internode length in these relatively dwarf plants ranges from 5 to 18 cm., the higher numbers being quite common. Variability in habit is characteristic of hybrid populations, as Anderson has repeatedly pointed out (see, for example, Anderson and Schregardus, 1944; An- derson and Gage, 1952). In allopolyploids such variability should be small if the variety is of the type termed by Stebbins a true allopolyploid (1947, 1950). If however, it is a segmental allopolyploid, more frequent pairing between members of the different genomes will result in a higher amount of segregation. Belladonna does, in fact, vary somewhat in habit, although the general tendency is for these plants to possess a more or less regular internode pattern, with branching common below the main inflorescence. In extreme cases the node number may be low, as in D. grandiflorum, and the aspect of pseudodichotomies may be found in the branching pattern; or they may be tall plants with short branches and internode lengths irregular below the inflorescence. On the whole, the direction of variation in this respect tends to be more toward D. grandiflorum than toward D. elatum. TABLE II POLLEN SIZE AND VIABILITY IN SPECIES AND HYBRIDS OF DELPHINIUM: HERBARIUM SPECIMENS.* UN % Viable Pollen diameter in ш pollen Range Median Diploids = naciais, Vilmorin-Andrieux 98 20.0—25.0 22.5 Blue Butterfly 96 17.5—25.0 22.5 D. сБейат ит, coll. Karo 92 20.0—25.0 20.0 Tetraploids — оо D. elatum, Bot. Gard., Cambridge 98 22.5—25.0 25.0 D. elatum, King of the Blues 60 22.5—25.0 25.0 Hexaploids D. formosum Hort., N 7186 60 25.0—32.5 30.0 D. formosum Hort., Helms. 17/6/25 25 25.0—32.5 27.5 D. formosum Hort., Dreer, July 1924 50 22.5—32.5 30.0 ormosum Hort., For 80 22.5-32.5 30.0 D. Barlowii Hort.t 77 22.5—35.0 30.0 D. Barlowii, Aug. 1t 31 22.5-35 30.0 Lamartine, Aug. 20 79 25.0—32.5 30.0 о о "Specimens listed are in the herbarium of the Bailey Hortorium, Cornell University. : e confusion as to the identity of D. Barlowii has been discussed on page 120. The sper mens cited here are morphologically of the Belladonna sort. 1953] GAGE—DELPHINIUM Х BELLADONNA HORT. 131 ELATUM- HORTICUL TURAL 0.909.999 7.0500 оо оро РҮҮ PM ELATUM- ALPINE [ BELLADONNA CHEILANTHUM А А МАМА, GRANDIF ORUM TRIPLOID- 48-27-3 Text-fig. 6. Internode diagrams to scale of diploid, tetraploid, and — Рон Ordinate, length of internode; abscissa, number of internode from rosett pright. unes р" le branches and are drawn to a scale 16 that employed for internodes. The black dots represent sing ower Text-fig. 5 is a scatter diagram for such of the triploids as were adequately represented by living or herbarium material, and an internode diagram of a typical triploid is shown in text-fig. 6. From the superimposed silhouette of the Bella- donna distribution (fig. 5) all these plants are seen to be within the limits of that group, ог to be intermediate between D. Belladonna and D. elatum. The same observation is borne out by the internode diagram, which shows the internode pattern of Belladonna, but with the high node number more characteristic of D. elatum, Summary.—Through the medium of the scatter diagram it has been demon- strated that Belladonna may be visualized as a more-or-less variable group, inter- mediate between the diploid D. grandiflorum (vel aff.) and the tetraploid D. elatum, and having as the source of its variation the segregation of characters derived from these species. Further, the experimentally produced hybrids between D. elatum and D. grandiflorum have been shown to be very similar to Belladonna, deviating from it only in characters for which their D. elatum parents were ex- treme. The scatter diagram also amplifies the position of D. cheilanthum: it is instantly differentiated from D. Belladonna by its diploid constitution, and by Constant morphological characters. The same conclusion is supported by the internode diagrams: amount of variability in D. Belladonna (extreme types were diagra there is a greater mmed, but are [Vor. 40 132 ANNALS OF THE MISSOURI BOTANICAL GARDEN not included in text-fig. 6) than in the three natural species, but on the whole it is intermediate between D. grandiflorum and D. elatum. In this respect D. cheilantbum is closer to D. elatum than to any other entity, differing chiefly in the development of its inflorescence. On the basis of character association, then, it is to be concluded that D. grandi- florum and D. elatum fulfill the morphological requirements for the parents of Belladonna; but that D. cheilanthum could not have been the diploid parent, given D. elatum аз the tetraploid parent. One cannot thereby rule out the pos- sibility that a tetraploid species other than D. elatum was involved, but no other known tetraploid species possesses the required combination of characters, nor is any other tetraploid so commonly grown. ГУ. CYTOLOGY or DELPHINIUM HISTORY The chromosome complement of Delphinium has been studied by a number of workers (Hocquette, 1922; Tjebbes, 1927; Langlet, 1927, 1932; Tischler, 1927; Beckman, 1928; Lewitsky, 1931; Lawrence, 1936; Propach, 1939, 1940; Gregory, 1941; Mehlquist её al, 1943, and unpublished; Lewis et al, 1947; and Lewis and Epling, 1951), and it has now been generally established that the basic complement of the genus consists of one long chromosome with a median centromere, one long chromosome with a sub-median centromere, five medium-length chromosomes with sub-terminal centromeres, and one short chromosome with a sub-terminal centro- mere. Langlet (1932) reported that satellites occurred in the genus, but his results were not conclusive regarding their number and location. Lewitsky (1931) published some data regarding chromosome differentiation within the genus, ай furnished ideograms for a number of species. Noting that in certain instance the four “homologous” chromosomes are not identical, he postulated amphidiploid origin for two tetraploid species, D. azureum* and D. elatum. Moreover, he pointed out that two species belonging to two distinct sections of the genus D. cardiopetalum and D. Staphisagria, have distinct types of ideograms from the other species he investigated. Моге recently, Lewis, Epling, Mehlquist and Wyckof (1951), reported that the chromosomes of the California species of Delphinium (which according to Ewan (1945) include representatives of several of the old World lines as well as a majority of endemics of uncertain affinities) 2 morphologically quite similar. They figure a composite ideogram from species included in Ewan’s SPICIFORM series, which shows one satellited chromosome (their E, which they report is otherwise very close in length to C, D, and F). That situation is not 80 simple as this and that greater possibilities exist for identifying different chromosomes in certain cases have been made clear from the present study in which, apparently for the first time in Delphinium, Feulgen and ace lacmoid squashes have been employed for somatic material. * D. azurem of Lewitsky was probably D. elatum; see footnote p. 164 1953] GAGE—DELPHINIUM Ж BELLADONNA HORT. 133 Numbers of chromosomes published so far in Delphinium are listed in the Ap- pendix. Diploid species are by far the most numerous, tetraploids are occasional, and hexaploids are known only in cultivated forms. Meiotic studies of Delpbinium have been largely confined to the garden forms, the major exception being the work reported by Lewis, Epling, Mehlquist, and Wyckoff on the California species, where both diploids and tetraploids were analyzed. For the diploid species, metaphase bivalents fall into several easily recognized types, since terminalization regularly occurs and chiasma frequency for a given pair shows little variation, the large chromosomes having usually two chiasmata at metaphase (there may be three or rarely four at diplotene and diakinesis) and the others having regularly one chiasma only, and that in the longer arm (Mehlquist e£ al, 1943; Lewis e£ al, 1951). In the tetraploids, the type of pairing varies, depending upon the origin of the tetraploid condition. Thus, in the California tetraploids, which are designated as races of the diploid species, D. Hanseni, D. gypsophilum, and D. variegatum, quadrivalents are formed by about 70 per cent of the long chromosomes with the submedian centromere, and by about 45 per cent of the long chromosomes with the subterminal centromere. (Likelihood of multivalent formation by the other chromosomes is precluded by their known low chiasma frequency.) Anaphase separation is regular, and the plants are fully fertile. However, it was demonstrated that the chromosomes are not completely homologous, since anaphase bridges indicative of inversions were common. Of other tetraploids studied, Lawrence (1936) recorded one to four univalents and an undetermined number of multivalents in Delphinium Ruysii, a garden hybrid possessing one D. elatum (n — 16) and two D. mudicaule (n — 8) genomes. А similar condition was found to exist in a group of tetraploid hybrids Produced by Mehlquist (unpublished) from crossing D. elatum with artificially doubled D. cardinale, in which meiotic pairing was found to be primarily by bivalents and univalents, although segregation in succeeding generations indicated that some multivalent pairing occasionally occurred. Particularly interesting in so far as this study is concerned are the varieties bred by К. Foerster of Bornim, Germany, and studied by Propach (1939, 1940). According to Propach, Foerster produced his hybrids by intercrossing D. Bella- donna varieties and D. grandiflorum var. chinensis, obtaining in the first generation fully fertile tetraploids. Later, some of these hybrids were again crossed with D. grandiflorum to form sterile triploids (Appendix). Unfortunately, full cyto- logical details were not given. Propach interpreted his results as indicating that there is little genomic differentiation in Delphinium and that autopolyploidy is involved.5 5 i . i ieties in all cases, d О (909) dee n Fr аши Тури and кире dtum X reno va chinense hybrids. Further, the accom i lates confirm his statements; нане thorough е companying р nfirm а cytological study of these forms is desirable before Propach's conclusions can be accepte f [Vor. 40 134 ANNALS OF THE MISSOURI BOTANICAL GARDEN The first object of this study was to discover whether any morphological differentiation of the chromosomes could be made among species of Delphinium, so that, if possible, the component genomes of the polyploid species might be identified. The second phase of the cytological work has dealt with the meiotic behavior of the chromosomes in the species and hybrids. Data obtained from chromosome studies have been supplemented by pollen studies which, to some ex- tent, serve to indicate the degree of fertility of a plant; and pollen measurements have been used, especially of herbarium specimens, to give information (which could not be obtained by other means) on the probable polyploid level of these plants. METHODS 1. Somatic Tissues—Chromosomes: Although the paraffin method has been the one used by previous workers, it was seldom employed here. This was not only on account of the length of time required to prepare the mounts, but also because root tips from sectioned material did not show the structural features evident in acetic squashes, nor could counts of chromosome number be made so readily. For squash techniques, it was necessary to adapt the method to the particular requirements of the material. Pre-treatment with a spindle inhibitor was found to be essential since the chromosomes of meta- phase plates were otherwise obtained in face-view only with the greatest difficulty; further, aceto-carmine did not give the degree of differentiation obtained with aceto-lacmoid. The schedule follows: а. Actively growing root-tips were removed and treated with a saturated solution of paradichlorobenzene (Meyer, 1945) for 3 to 4 hours. b. Tips were transferred to 7:3 acetic alcohol (Darlington & La Cour, 1947) for 1 to 5 days. Preparations from material preserved longer than this are not clear enough for studies of details of chromosome morphology, but are adequate for counts. Tips were hydrolyzed in a mixture of approximately 2 drops 0.1 normal НСІ and 5—6 drops acetic-lacmoid stain (Darlington & La Cour, 1947), heating on a glass until vapor could be seen, and then letting the tips remain in the mixture for 10 to 20 minutes, or until they became soft. (Sometimes it was necessary ® increase the proportion of HCl in the staining mixture in order to soften the tips sufficiently, but generally as little acid as possible was used since an excess decrease staining.) d. Root tips were rinsed in 45 per cent acetic acid for 5 minutes to about a? hour. e. The deeply stained portion of a tip was transferred to a clean slide, and stain and cover-glass were added. The mount was heated and squashed and then sealed. After the cells were dispersed by gentle tapping, considerable pressurt could be applied in order to spread out their contents. The preparations will keep for about a week if stored in a cool place, but they are best within a few hours after being made. 1953] GAGE—DELPHINIUM X BELLADONNA HORT. 135 2. Somatic Tissues—Nucleoli: . The Feulgen-fast green squash technique (Darlington & La Cour, 1947) was used. Hydrolysis time was about 20 to 40 minutes. 3. Meiotic Chromosomes: The exact method used depended upon the age and condition of the material. In general, the following schedule was employed: Anthers were killed and fixed in acetic alcohol for several hours to a week or more. (Best results were obtained if mounts were made in about one to four b. A whole anther was transferred to a slide, flattened with the needle, and stained. 1. Fresh material: acetic lacmoid gave the best results. | 2. Older material: a mixture of aceto-lacmoid and aceto-carmine was employed, the relative amount of carmine being increased with increasing age of the material. 3. If the material was not stored for more than two weeks, it could be transferred to 70 per cent ethanol, and thereafter stained with carmine, in which case mordanting was found to be necessary. с. A cover slip was added and the slide was squashed gently and sealed. After the slide had been sealed it could be heated and again squashed if necessary. 4. Pollen Studies: Preparations were made with either aceto-carmine or acid fuchsin, at least three slides being made for each plant of which the pollen was available. The anthers were taken on different days and from different flowers so that the effect of environmental variation could be observed. A total of 200 pollen grains was counted on each slide, only the deeply staining, normally shaped grains being considered "viable." Where possible, 100 g grains of each individual were measured, but in the triploids the number is understandably considerably less. Pollen preparations from herbarium specimens were made employing the fast- green glycerine jelly method of Wodehouse (1935). 5. Illustrations: Camera-lucida drawings were made with a Spencer Camera Lucida on a Spen- сег microscope with 15 X ocular and 95 X М.А. 1.40 apochromatic oil-immersion objective, giving a magnification of 2180 X at table height. Reductions are given with each figure. The photomicrographs were taken with a 1 X Bausch & Lomb fixed bellows camera; magnifications accompany the plates and text-figures. DETERMINATION OF THE LEVEL OF PLOIDY While the determination of polyploid level in the living plants wae always made by chromosome counts from root-tip preparations, the impossibility of applying direct methods to plants represented only by herbarium material made necessary the use of the indirect method of pollen measurement. In studying the Percentage of "viable" pollen produced by hybrid individuals, it had previously [Уот. 40 136 ANNALS OF THE MISSOURI BOTANICAL GARDEN been noted that diameter of the pollen grain was a good indicator of the polyploid level of the plant (Table I). For the plants measured there was good separation of diploids and tetraploids on this basis. The hexaploids and triploids generally produced the largest pollen, although there was some overlapping between these and D. elatum. It is to be noted that pollen production by the triploids is very low (except in the colchicine-treated spike of 50-40-2, which is presumed to have been doubled); the similarity in size of their few pollen grains to that of the hexaploids is ап indication that the pollen from these plants is unreduced, and hence triploid. For pollen grains of the herbarium specimens, size of pollen grain was again found to be a differential character for the polyploid level, although the actual sizes obtained were somewhat smaller in all classes, as might be expected on the basis of the difference in the suspending medium of the stain and from long drying of the material (Table II). CYTOLOGY OF NATURAL SPECIES, DELPHINIUM BELLADONNA, AND THE TRIPLOIDS Morphology of the Somatic Chromosomes.— n view of the general similarities in size of the chromosomes already noted as characteristic of the genus, ideograms are presented here for two species only: D. grandiflorum var. chinensis and D. cheilanthum. ‘The A and B chromosomes of these species probably are to be compared with similarly lettered chromosomes in the ideograms of Lewitsky (1931) and Lewis e£ al (1951), but because of the similar sizes of the smaller chromosomes it is not to be assumed that these are com- parable to similarly lettered chromosomes in the other ideograms. The major distinguishing features of the genomes figured here (pl. 13) have been the satellite-bearing chromosomes. The presence of nucleolar-organizing satellited chromosomes in the complement of many plants has been often noted since the classic papers of Heitz (1931), Navashin (1934), and McClintoc (1934). Later, the discovery of polyploid species in which the satellite number was comparably increased led certain workers to assume that often, if not always apparently diploid species which possessed several satellites in the gametic comple- ments are actually of polyploid derivation. This was the view held by Gates who published a comprehensive review on this subject in 1942. However, since many polyploids never have more than two satellites, and since, on the other hand, in such plants as Leontodon leysseri three of the four chromosomes of the haploid complement bear satellites (Elliot, 1950), inferences as to polyploid level based on satellite or nucleolar number alone must be regarded as unsound. Another problem related to the existence of satellites must be considered № order to appreciate the variations in satellite number observed in the Delphinium hybrids, and that is differential amphiplasty. This phenomenon was first noted by Navashin (1927, quoted by him in 1934), who applied the term to the apparent loss of satellites which sometimes occurs in hybrids. The following year, Mc- Clintock, in her paper on the nucleolar organizing chromosome of Zea Mays 1953] GAGE—DELPHINIUM Х BELLADONNA HORT. 137 discussed in detail Navashin’s work, interpreting the loss of satellites in hybrid forms as being due to the inability of the chromosome to "organize" a nucleolus in the presence of another more "active" chromosome. Amphiplasty has recently been reported in the genus Leontodon (Elliot, 1950), and had been occasionally reported in other genera (Meurman and Therman, 1939; Levan, 1937). While our present understanding of differential amphiplasty is by no means perfect, the cases cited above seem to show that under certain conditions such as obtain in some hybrid cells, the apparent morphology of nucleolar organizing chromosomes may be modified. Since the presence of a satellite at metaphase is usually directly related to the formation of a nucleolus in that region in the preceding telophase, then the causative factor is apparently related to nucleolar metabolism. Nevertheless, the actual physiological differences between individual nucleolar "organizers," which are the essence of the problem, are not known at all, and it is therefore the morphological aspect of the phenomenon which concerns us ere. Diploids.—In pl. 13 is given the ideogram of D. grandiflorum L. var. chinensis Fisch. hort. var. Blue Butterfly, pollen parent of the 48-27 and 48-6 series of triploids. The actual length of metaphase chromosomes ranges from 3.1 to 10.1 p. (These measurements are based on chromosomes not artificially contracted by PDB, although the chromosomes measured were subjected to minimum treatment in order that the spindle would not interfere with their being well-flattened.) Of the shorter, subterminally constricted chromosomes, chromosome H is readily identified by its length as well as by the satellite; of the rest, chromosome C may be identified by its satellite and sometimes by the secondary constriction of the long arm. Chromosomes D through G are individually distinguishable only with difficulty; the lengths of their short arms, on the whole, were found to be more diagnostic than the long arms, which may be variously extended, being slower to complete their coiling cycle than the short arms. А difference in the activity of the satellites was noted here. While usually both C's were satellited, no cells were ever observed in which at least one C did not bear a satellite; the satellite of the Н chromosome, on the other hand, was commonly present in unicate, and sometimes was even missing altogether. This Was generally true for all plants of D. grandiflorum and was statistically deter- mined for two individuals (Table II). Related plants, including 47031-1 from the Royal Botanic Gardens at Kew, D. grandiflorum L. var. chinense Fisch. from Basel, and numerous plants of horti- Cultural varieties such as Tom Thumb and White Butterfly were examined and found to have a similarly differentiated complement (pl. 13). 47031-1 was found to possess a telocentric fragment, about the length of the long arm of the H “Ju, the conres of numerous observations, in one plan Оп one of the oi chromosomes, but as it was not present in an omitted from the ideog this ies a third satellite was noted die ui er case it has been [Vor. 40 138 ANNALS OF THE MISSOURI BOTANICAL GARDEN chromosome. Though the numbers of fragments present in different cells of an individual have been found to vary somewhat, the fragment is commonly trans- mitted through the gametes, being found in plants three and four generations removed from 47031-1. Unfortunately, the original plant and the fragment- bearing triploids were lost before pairing affinities of the fragment could be determined, and there are sufficient irregularities in the descendant hexaploids to make its identification in PMC's doubtful. Plants of D. tatsienense Franch., a species closely related to D. grandiflorum І, were found to possess a genome similarly differentiated except that in the indi- viduals surveyed the satellite is as often developed on the H chromosome as on the C chromosome (see 51-311-9, Table III). D. cheilanthum.—By contrast with D. grandiflorum, D. cheilanthum bears its satellites on the A and C chromosomes, that on the A being not only larger but also the more "active" of the two (Table III, 51-307-6 and 51-307-1). $e far as can be ascertained, the chromosomes of the two species are otherwise similar, except that the satellite-bearing arm of A is slightly longer here than in D. grandiflorum. SATELLITE EXPRESSION IN DELPHINIUM SPECIES AND HYBRIDS Fraction of chromosomes bearin : | sstellite* B Number Theoretic 1 Actual average Plant of satellite | . lite number | C | H | А | се number Diploids grandiflorum 51-312-6 0.94 0.48 = 100 2:2:0 1.8:1.0:0 51-315-6 0.91 0.62 d. 100 2:2:0 1.8:1.2:0 tatsienense 51-311-9 0.85 0.80 = 100 2:2:0 1.7:1.6:0 cheilanthum 51-307-6 0.47 = 0.95 100 2:0:2 1.0:0:1.9 51-307-1 0.49 a 0.98 100 2:0:2 1.0:0:1.9 Tetraploids Е. D. elatum 51-318-2 0.56 0.12 0.85 21 4:2:2 2.2:0.25:1.6 51-318-4 0.50 0.07 0.92 71 4:2:2 2.0:0.15:1.8 Triploids 48-27-3 0.61 0.33 0.05 100 3:2:1 1.8:0.67:0.05 50-15-3 0.52 0.34 0.63 100 3:2:1 1.5:0.68:0.63 Hexaploids mede ete Smith's Belladonna 0.63 0.10 0.84 25 6:4:2 Ms 51-232-4 0.45 0.25 0.13 60 6:4:2 2.7:1.0:0-3 wur p лш т 7 *Fraction is computed on the basis of the actual observed maximum number of satellites for that chromosome. 1953] GAGE—DELPHINIUM Х BELLADONNA HORT. 139 Other diploids.—A third situation with respect to satellite number and position was found in D. cardinale (pl. 13) and D. nudicaule from California, and a fourth in D. Zalil Aitch. and Hemsl. (native to Persia and Turkestan) and D. sulpbureum Boiss. and Hausskn. (from north Syria). In the California species the satellite number is 5, and the satellites occur on the H and on two pairs of intermediate chromosomes, one of which may or may not correspond to C, the lengths of the chromosomes in these species being not exactly comparable to those of D. cheil- antbum and D. grandiflorum. ‘The third satellite is on the second shortest chromosome. Despite the high potential satellite number, the low number of nucleoli usually present at telophase and interphase (generally no more than 4) indicates that all are rarely active in a particular mitotic cycle. In D. Zalil and D. sulphureum satellites occur on A and on one of the intermediate chromosomes. However, the genome differs from that of D. cheilanthum in the arm lengths of the intermediate chromosomes, so that direct comparisons cannot be made. In studying satellite development in the above diploid species a situation was noted which has apparently not been described before, namely, that even within old diploid species differential "activity" of satellited chromosomes may be com- mon. With this in mind further variations in satellite development which are found in polyploid species may not be so unexpected. In order to avoid ambiguity here, I shall use the term “differential activity” to apply to the differences in satellite development such as characterize diploid species, and "amphiplasty" to refer to any pronounced alteration in differential activity which follows hybridiza- tion. Actually, however, it is not to be assumed on such scanty visual evidence that any physiological distinction is being made. Tetraploids: D. elatum.—1n D. elatum, the only tetraploid species observed, distinctions among the chromosomes are again difficult to make. There are 2 Pairs of long, medianly constricted A chromosomes, 2 pairs of long, sub-medianly constricted B chromosomes, 10 pairs of intermediate chromosomes with subtermi- nal constrictions, and two pairs of short chromosomes with subterminal constric- tions. Amphiplasty obscures the satellite situation, the more active A an satellites usually being the only ones present (Table Ш), but a study of а large number of cells, both in the species (accessions from Géteborg Botanic Garden, and the Royal Botanic Gardens at Kew) and of horticultural variants, indicates the presence of satellites on 1 А, 2 C's and 1 Н of the haploid complement. Although the number of satellites in the tetraploids parental to the triploids could not be determined statistically, camera-lucida drawings of the complements of 47-98-3 and Summer Skies indicate that these plants were typical in this respect (pl. 14). Asin D. cheilanthum, the satellited A chromosome stands out from its counterpart in the other genome by virtue of its longer satellited arm. Lewitsky (1931) noted this difference in length, but did not observe the satellite. The nature of the chromosome differentiation in D. elatum is thus strongly suggestive of an allopolyploid origin by hybridization and subsequent doubling of the chromosome number of diploid forms whose chromosome complements were of [Vor. 40 140 ANNALS OF THE MISSOURI BOTANICAL GARDEN the types represented in present-day species D. grandiflorum and D. cheilanthum. That the chromosomes designated at C are truly comparable in the two genomes and are not different members of the medium-sized group is corroborated by the fact that both possess a secondary constriction of the longer arm. Aneuploids are occasionally found, even among the plants from the botanic gardens. However, they do not possess the abnormal characteristics which ordi- narily distinguish aneuploids on the diploid level, and fertility, as judged by amount of seed produced, is apparently not greatly impaired. Triploids.—All plants of the four groups of triploid hybrids were examined and found to be essentially similar in so far as the morphology of the chromosomes is concerned (pl. 14). The maximum satellite number in each is 6: 1 A, 3 С% and 2 H's, as would be expected on the basis of the generally regular segregation of the parents. The precise situation as to satellite development has been thoroughly worked out only for two individuals, 48-27-3 and 50-15-3, most of the other plants having died before this part of the study was undertaken. An interesting case is 48-27-3, in which the satellited A chromosome, so active in D. elatum, is rarely developed. (Study of camera-lucida drawings of other plants of this line and of the very similar 48-6 line indicates that this is generally true of both groups.) However, in 50-15-3, the other triploid analyzed, the A satellite was observed in more than 80 per cent of the cells examined. Thus, in crosses of similar parentage, amphiplasty has been demonstrated in one case, but did not occur in another. With regard to the centric fragment of their pollen parent 47031-1, triploids 50-15-1 and 50-15-4 were found to possess the fragment in most cells; 50-15-3, 50-40-1, and 50-40-2 lack the fragment. Hexaploids.—The tendency already noted for partial suppression of the poten- tial satellite number was especially evident in the hexaploids, of which Smith’s Belladonna,’ Cliveden Beauty, Bellamosum, and Lamartine were observed in some detail, as well as in the derived hexaploids which are to be considered in chap. ү. In most cases, cumulative records on these plants indicated a possible satellite number of 2 A's, 6 C's, and 4 H's of the diploid complement. However, the normal satellite number in individual cells is often only half this number (pl. 15). When interplant variation in degree and type of amphiplasty was studied Smith's Belladonna was found to possess very active А SATs, but among its progeny obtained by selfing, all degrees of activity were found; this was also true of the Fı obtained by crossing this plant with Bellamosum. At the other е treme are certain individuals of Bellamosum, in which the satellited A can rarely be distinguished (Table III). Meiosis and Fertility.— Meiotic studies have been made of the triploid hybrids and their parents, where these have been available, as well as of Belladonna varieties, in an effort to compat? "The plant termed "Smith's Belladonna” th : А : the horti- cultural variety "Cliveden Beauty." throughout this study is a selection from TABLE ТУ. MEIOSIS OF THE TETRAPLOIDS eu ا‎ HRS ا‎ eS ah д ана | an Eno E р TE TET а 4 Ly м та € “ K][t1naonzas == Е ТЕ I Е $22 sequin Sede 3 jewsouge EZET nnd 2s Á[[t1n32n235 +“ epo 29qumN. TEL TEE Т e الا‎ * е NN ч ч a urirouqt д, єз + НЕЕ ILI p qt 25 iz а у. 2 IIL ‘ds I [jn =m" own E mer Aa == LASS a РАНА I6 TI 9 Hr IN ¥ n m~ = jewsou gá E < Á]Je1n35n21$ aa АНИ — 3 2 $122 39qumN Z & тартка о ENS jewsouge z ipe Á][tan3on2as 3 = $122 1241 М < = + I zz Ija zi T $22 jo 19qumN о + < 7 colon ee А РОН Cee x а ед < I O1 ‘II It on Е Павии n = ЕСЕСІ ЕКСЕ g IS T dg ә = Z pcr n Е. I91 чу 9 e- e| „Ше ec A I9'"Uu «| >> B Ф £ її ПП <| ^ 9 ^ кары o » тия | 9^ - IZI JI 9| = Ke ie _—_—_ a 5 IZ ‘IE SI ык : z тот“ Lj te || CS II 91 "m п IS I$ SEM - о © Е ү Е ут Бони a 75 = ыыы то лм м а 5 Е чөдө $777 “Е oo oo оо o0 оо || 00 1 ooo es "Bot cb ob ocBdd TT Olli i. a "LHOH VNNOGVTIJH X WüINIHdTIdG—H5Vv9 [#561 [Vor. 40 142 ANNALS ОЕ THE MISSOURI BOTANICAL GARDEN chromosome homologies in the triploids and Belladonna. Early prophases are usually difficult to stain by squash methods, and while fair preparations can be made by paraffin sections, using gentian violet stain, the high number of chromo- somes present makes prophases generally of less use than later stages, which can readily be stained by conventional squash methods. For the most part, configura- tions at diakinesis and metaphase were analyzed, and anaphase I and II were scored for structural abnormalities. However, counts of chromosome segregation were not made, since the frequent occurrence of aneuploids in all polyploid populations is sufficient evidence that hyperploid and hypoploid gametes are produced. The approximate percentage of "viable" pollen produced by certain key plants is given in Table 1. However, since samples from the same plant taken on dif- ferent days may show a wide range of viability, the values given can scarcely be considered absolute indications of fertility. Moreover, apparently good grains may not survive until the pollen tube has reached the ovule, not only on account of some deficiency of the grain, but also, perhaps, because of incompatibility with the stylar tissue. Certainly, the particular environmental conditions prevailing at the time of pollen formation and shedding are not the least important factor in the production of functional pollen and its further development. It has repeatedly happened that a plant which produced a large amount of well-filled, staining pollen grains and set a high percentage of viable seeds at one time, produced little pollen and set few seeds at another time, or again, produced apparently good pollen but set no seed. The immediate cause of impaired fertility of this sort is often 2 prolonged hot spell, but other factors may have a similar effect. Diploids.—Blue Butterfly is usually quite regular in meiosis. In almost every one of numerous cells examined at diakinesis and MI, pairing was in 8 bivalents (pl 15). Chiasma frequencies ranged from 2 to 3 for chromosomes А and B, to 1 or occasionally 2 for the others. Anaphase separation is quite regular, bridges occurring in less than 1.0 per cent of the cells examined. A regular second division is followed by the simultaneous cleavage of the pollen mother cell into tetrads. Unfortunately, metaphase pairing of 47031-1 was not observed during the first year and the plant was subsequently lost. In anaphase I, however, laggards (probably the fragment) were noted in 9 per cent of the cells, and bridges in 4 per cent of the cells. The amount of “viable” pollen was about 65 per cent. The presence of the fragment in two of the five triploid offspring of this plant is satis- factory indication that pollen fertility was not prevented by the possession of the fragment. However, whether the failure of the plant to set any seed when self- or cross-pollinated may indicate some deleterious effect of the fragment in the ovule cannot now be shown. This case thus appears to be quite different from that of the centric fragment studied by Rhoades in Zea (1940), where trans mission of the fragment was largely through the egg, fragment-bearing pollen grains usually not competing successfully with normal pollen grains. Tetraploids.—Meiosis of Summer Skies is not entirely regular, the occurrence of 2 to 4 univalents at metaphase I being rather frequent. Further evidence of 1953] GAGE—DELPHINIUM X BELLADONNA HORT. 143 structural hybridity is given by anaphase I and second-division configurations, bridges, fragments, and lagging chromosomes being present in 11.3 per cent of the А I and 23.3 per cent of the A II cells examined (Table IV). Pollen viability appeared to be rather low (Table I), but that the plant is highly fertile has been indicated by the amount of seed produced. 47-98-3 exhibited fairly regular pairing; only two of the twenty diakinesis and MI plates showed each 2 univalents. However, a higher proportion of the large chromosomes had reduced chiasma frequency, with sometimes only 1 chiasma per bivalent. Further evidence of chromosomal hybridity is seen in the rather high number of cells at anaphase I which showed visible abnormalities (44.0 per cent). Abnormalities of the second division were less frequent, being present in only four of twenty cells examined. Pollen viability was about 65 per cent. 47-2-3, the tetraploid parent of the 48-6 group, was not available for study. Triploids.—Representatives of the several crosses have been analyzed. Since the number of PMC's examined in each case have not been large enough to permit separate discussion, the data have been somewhat grouped. 48-27-: Numbers 2, 3, 5, 6, and 12 have been studied (Tables V and VI, pl. 15). Metaphase pairing is largely in bivalents and univalents, only 34 trivalents having been found in the 258 cells analyzed. Usually, the long chromosomes (А and B) were involved in the trivalent configurations, but trivalents formed by chromosomes of intermediate size were found at least twice. The low number of trivalents involving the intermediate chromosomes may be due in part to their observed low-chiasma frequency, but also to the presence of differentiated seg- ments. The amount of even the A and B chromosomes present in trivalents is only 7.2 per cent. In two individuals (48-27-3 and 48-27-5) bivalent formation is near the maximum number expected if normal chiasma frequencies prevail—that is, if those chromosomes which pair form the normal number of chiasmata. However, whether such pairing is ordinarily between the 8 chromosomes contributed by D. grandi- florum and 8 chromosomes of D. elatum, or whether pairing is primarily within the elatum complement only can not be determined precisely. The above-men- tioned occurrence of occasional trivalents does indicate that some mixture of the two sorts prevails. That intragenomic pairing may occur in D. elatum is already known from the work of Lawrence (1936) and Mehlquist (unpublished) ; but the morphological similarity of one D. elatum genome to that of D. grandiflorum I5 equally strong indication that pairing between these chromosomes cannot be discounted, : Despite the rather high proportion of bivalents which may be formed in cer- tain individuals, the frequency of univalents, which may sometimes number З Тапу as 22, is indicative of structural differentiation of the chromosomes. This is further substantiated by the high percentage of structural abnormalities observed at anaphases I and П. At least one bridge-fragment was observed, on the average, їп about 50 per cent of the cells examined, both at A I and at A П. Such config- [Vor. 40 144 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE VI MEIOTIC ANAPHASES OF THE TRIPLOIDS Anaphase I Anaphase II 2 "a 8 ч E: Я Я 3 я g я Е 5 E Е 5 Е ob a 5 50 t=] E Plant be. = 2 + = С] B an E э B a = - | Е B 8$ 21 * B E 50 & 9 “Ж о on о е $13.24 ов орна Be ET 39 @ - ед a = بم‎ a - рд a Е = 48-27-2 i i i rog 64 | 67.2 7 17 30 | 433 48-27-3 4 61 4192] 20451 5% i4 4 151 Hi 7 | 627 48-27-5 29 33 12 33 107 69.2 12 7 5 35 59 | 407 48-27-6 10 | 21 23 $5. 60: |. — | — do bo = е, 48-27-12 44 25 16 39 124 68 — == pa: 48-6-1 i ОТС Zii p |. 3 1 3 14 | 500 48-6-3 13 | 10 6 173 34 ОЛУ Wit ГИ 5 | 10 39 | 74.3 50-15-1 ZU] HEI E R4 ov Зее Бос] се. ЕЕ 1:19 Wr 50-15-3 14 1 4 | 28 47 | 40.5 || 11 B 4 1.23 41 | 46. 50-15-4 19 2 45 62 0.3 | — | — | — = ran Bees urations have been shown by others to be caused by crossing-over within inversions, by crossing-over involving certain types of duplication (McClintock, 1941) о as suggested by Emsweller and Jones (1938), by pairing between chromosomes whose insertions are not opposite. The number and behavior of the lagging chromosomes are of interest. The average number of laggards was less than one per cell in A I, and even lower in A IL As a result, except for the numerical imbalance of most of the gametes formed by these plants in natural consequence of their triploid constitution, the loss of univalent chromosomes, which is often characteristic of such plants, W?$ not pronounced. This point will be considered further in the treatment of the hexaploids. The number of microspores per tetrad is variable, 2 and 3 being the more com- mon variants, but a fifth micrograin is occasionally formed. Micronuclei - extremely common, but are usually included with one of the other nuclei within a common wall. The amount of apparently good pollen has been found to be quite small, and that which is formed is within the range of pollen size of the hexaploids, 518565008 that such pollen is actually unreduced (Table I). -6-: Numbers 1 and 3 of this line were briefly analyzed. There was арр?” ently a greater tendency to bivalent formation in 48-6-1, 7 II having been not in several cases, while 2-6 II were found in 48-6-3. One trivalent was noted i? 48-6-1. In anaphase I and anaphase II somewhat higher frequencies of abnor- malities were noted in 48-6-1 than in 48-6-3, 1953] GAGE—DELPHINIUM Х BELLADONNA HORT. 145 Viable (unreduced) pollen grains were formed in 2.8 per cent of cases in 48-6-3 and 5.7 per cent in 48-6-1 (pl. 15). 50-15-: Numbers 1 and 4 have been examined and seem to differ somewhat in pairing relationships. Single trivalents involving the large chromosomes occurred in 5 of 18 figures in 50-15-1, whereas no trivalents were noted in the 15 cells of 50-15-4 analyzed. Structural abnormalities at A I were present in about 40 per cent of the cells of each, and there were many irregularities at А П. Pollen viability was determined at 1.5 per cent and 2.5 per cent for numbers 1 and 4 respectively. 50-40-: A survey of meiosis in 50-40-2 indicated that pairing was usually in 5 to 7 bivalents, with an occasional trivalent. Reduced chiasma frequency was evident in the А and B chromosomes, and abnormalities of later stages were frequent. Moerbeimi was lost early in the study, but preliminary observations indicated that meiosis was comparable to the three hybrid groups. Pollen viability was ex- tremely low, less than 0.5 per cent unreduced pollen grains having been found. Hexaploids.—Belladonna: Lawrence (1936) has reported that meiosis in this form is largely regular. In Cliveden Beauty, a light-blue Belladonna variety, 2 to 4 univalents (of small or intermediate chromosomes) were noted in 10 out of 16 MI cells, but the chiasma frequency of the А and B chromosomes was only slightly reduced. At anaphase I structural abnormalities were visible in 12.3 per cent of the cells. Anaphase II was quite regular, only 7 of 121 РМС being visibly abnormal. Individuals of this variety were found to have from 63.2 to 80.5 per cent of good pollen. Preparations of Smith’s Belladonna showed somewhat less regular pairing at metaphase, with up to 8 univalents and the A and B chromosomes often with but one chiasma per pair. The actual amount of pairing was somewhat higher, how- ever, for in diakinesis univalents were uncommon and the large chromosomes had mostly two or three chiasmata in each bivalent (Table VII, pl. 15). Table VII gives some indication of the irregularities which are found in D. Belladonna varieties, although the plants cited by no means represent a random sample. A certain number of univalents are almost always present, but they assort at random and do not commonly lag at the plate. Multivalents are rare, but were noted in 52-325-11 (Smith’s Belladonna Х self), where both the A chromosomes and intermediate chromosomes were observed in such configurations. However, although one A chromosome was lacking from this plant, the monosome Progeny of various crosses. Thus, in 51-224 (Smith's Belladonna X Bellamosum, a dark blue Belladonna segregate) 8 of the 15 plants in the sample possessed 48 chromosomes, and in 52-311, produced by selfing а 48-chromosome individual of ` [Vor. 40 146 ANNALS OF THE MISSOURI BOTANICAL GARDEN MEIOSIS IN BELLADONNA TYPES S96 abnormal cells Metaphase I Anaphase I Anaphase II E 2 THERE 2 E Plant 3 3 d 3 E х 3 E i E ER 13 31434 з Е, 8| | =| 4% Е 2 ГЕЗ : 1] © ы в 5 aj 3 «| Х| > 1 Е E E Z| 8s! 3| 8| 3 a: я 8 |] аа ОЕ 1 е Ж 1 =i zi 61 dai a | z | 81434 Smith's Belladonna* | 48 | 55 | 24—20 0 0 23 || 12 2 2 33.3 || Mostly regular Cliveden Beauty-5 48 | 15 | 24-22 0 0 23 || 81 3 3 7.4 || 121 4 3 52-325-11 47 (SA). 32 | 23-19 | 4 2 22 || 48 3 9 25.0 || — | — " 51-224-16 48 | 21 ] 24-22 0 0 23$ 129 |. A4 7 35.9 14 1 0 $1-232-47 49 | 45 | 24-19 0 0 23 ||100 9 | 34 43.0 A9 rtine 48 | 46 | 24-21 0 0 23 || 67 6 4 14.9 || 104 7 6 * Of 21 late prophases observed, 17 showed 24 II, and 3 had each 23 II and 2 I. One cell had 6 univalents. this population, chromosome numbers of 25 plants ranged from 41 to 49, only 7 individuals having the "normal" number. Again, among 25 individuals of cross 51-232 (Bellamosum X self) chromosome numbers ranged from 45 to 49, 11 plants having the 48 chromosomes expected. In a sample of 10 plants obtained by selfing Smith's Belladonna (not figured), numbers ranged from 46 to 48. In most of the above cases the additional or missing chromosomes were of the smaller classes, but one plant with 48 chromosomes (51-232-1) lacked an A; and 51-224- 13, with 45 chromosomes, lacked a B. None of the aneuploids were visibly ab- correlation between chromosome number and pollen viability. Evidently, like many other polyploids, D. Belladonna can withstand the loss of a certain amount of genetic material without suffering the consequent weaknesses and abnormal development common when such losses occur at the diploid level. Discussion and Summary.— The cytological information given above bears on the problem of the identity 1953] GAGE—DELPHINIUM X BELLADONNA HORT. 147 E ЕЕ ER Text-fig. 7. Frequency distributions of chromosome numbers in populations of D. Belladonna re of the derived hexaploids. Ordinates represent per cent of population; abscissa, number of chromosomes, A—51-224 D-—51-216 G—51-219 J—51-222 B—52-311 E—51-217 H—51-220 K—51-223 C—51-232 F—51-218 I—51-221 not equivalent to morphological loss of the satellite was made evident by analyses of certain plants in which the rare occurrence of the satellite proved that its usual absence does not represent permanent loss. This condition is thus different from the actual morphological differences in satellite disposition between the chromo- somes of the two diploids D. grandiflorum and D. cheilanthum. Recognition of this difference is important for the interpretation of the satellite Pattern in D. elatum, the various triploids, and D. Belladonna. Thus, D. elatum might be a derivative of two diploid species whose genomes were differentiated along the lines of present-day D. grandiflorum and D. cheilanthum. And, given D. elatum as the tetraploid parent of D. Belladonna, D. grandiflorum might be the diploid parent; but D. cheilanthum could not have been, unless drastic mor- Phological alterations of the chromosomes have ensued since the time of origin of Belladonna. If the alternative supposition is made, that D. cheilanthum was the diploid parent of Belladonna, again, as with the morphological characters, there is the problem of finding a tetraploid other than D. elatum in the gardens of Europe— а tetraploid, moreover, both of whose genomes of the haploid complement are . similarly differentiated. [Vor. 40 148 ANNALS OF THE MISSOURI BOTANICAL GARDEN Analysis of pairing homologies of the chromosomes has proved illuminating in some respects, though not so useful in others. This was not unexpected, since in polyploids above the tetraploid level all combinations of autopolyploidy, segmental allopolyploidy, and amphiploidy may exist within a single individual (Stebbins, 1947, 1950). The triploid hybrids were found to exhibit neither the complete absence of pairing characteristic of the progeny of crosses between very distantly related species, nor the relatively complete pairing of some segmental allopolyploids, or of autopolyploids. True, as has been pointed out, the low metaphase chiasma fre- quency of the chromosomes would ordinarily prevent trivalent formation by all except the large A and B chromosomes, but the actual number of univalents was ordinarily higher than the eight expected if heterogenetic pairing either between grandiflorum and elatum chromosomes or within the elatum complement were of regular occurrence. Actually, the formation of occasional trivalents by the large chromosomes indicates that such pairing as does occur is a mixture of the two sorts. It has repeatedly been noted that pairing of chromosomes in the F, of such a species cross cannot be taken to indicate that a polyploid derivative of the cross (here a hexaploid) would necessarily be characterized by corresponding multi- valent formation (Darlington, 1937; Goodspeed and Bradley, 1942; Clausen, Keck and Hiesey, 1945; Stebbins, 1947, 1950). Rather, after doubling has oc- curred, differential affinity of the chromosomes may result in regular bivalent formation, not only in allopolyploids (Clausen and Goodspeed, 1925; Buxton and Newton, 1928; Kagawa and Nakajima, 1933) but even in known autopolyploids (Skirm, 1942). Since each case must be considered individually and the direct hexaploid deriva- tive has in this instance not been produced, the triploids can best be described as being the sort of plants which might well give rise to segmental allopolyploids, the low chiasma frequency of whose chromosomes would, on the whole, lead to more-or-less regular bivalent formation in the hybrid. One should expect, then, relatively little segregation in future generations, and the plants would continue to possess the more-or-less intermediate appearance of the triploids. The cytology of Delphinium Belladonna itself suggests that this species behaves rather like a segmental allopolyploid, but with perhaps fewer aberrations than аге generally found in polyploids of this type. Thus, while at first metaphase, pairing is primarily by bivalents, still the formation of univalents is not uncommon. Multivalents are rarely foun А consequence of univalent formation in gametogenesis is the production of numerous aneuploids. 'The mere occurrence at the polyploid level of viable aneuploids of good fertility is not unknown (Sears, 1944; Clausen and Cameron, 1944), although their high frequency here was not anticipated. In Triticum aestivum and Nicotiana tabacum, most of the nomosomics and nullisomics are individually recognizable, and this has been found to be true in certain other poly- ploid hybrids where aneuploids have been found (Primula kewensis, Newton 1953] GAGE—DELPHINIUM Х BELLADONNA HORT. 149 Pellew, 1929; Nicotiana tabacum, Clausen and Goodspeed, 1924; Nicotiana rustica, Lammerts, 1932; Nicotiana glauca X М. Langsdorffii, Kostoff, 1939). However, while a part of the variability within Belladonna may be owing to aneuploidy, still it has not been possible to isolate any characteristic monosomic phenotypes, even for the large chromosomes, suggesting that there must be a considerable amount: of duplicated chromosomal and genetic material in D. Belladonna. This conclusion is corroborated by a survey of the extent of hypoploidy which Belladonna tolerates. Plants having from 43 to 44 chromosomes, while apparently infertile, are not themselves weaker than their euploid sibs. In most allopolyploids with fully differentiated genomes, such aneuploids as occur are usually deficient for no more than one or two chromosomes (Newton and Pellew, 1929; Müntzing, 1937). It is in plants in which the genomes contain many homologous chromo- somes or chromosomal segments that extreme aneuploids have generally been found (Müntzing, 1937; Myers and Hill, 1940; Myers, 1947; Love and Sunneson, 1945). Since the univalent chromosomes are usually included in the daughter nuclei, the prevalence of hypoploids among the plants of all the Belladonna populations studied can be explained on the assumption that in these relatively balanced plants, a gamete deficient for one or several chromosomes is at an advantage over gametes containing chromosomes in excess. Therefore, those gametes which function would then include 24 or fewer chromosomes. It is not, however, to be concluded that Belladonna is insensitive to chromosome losses of indefinite extent: on the Whole, aneuploids with 46 or 47 chromosomes are the most frequent, so that functional gametes must not, as a rule, lack more than one or two chromosomes. Further, though intervarietal crosses are easily made, the germinability of seeds Produced is higher in selfs than in intervarietal crosses (Table XI). It seems, therefore, that such chromosomal differentiation as exists between varieties may result in sufficient deficiencies either in the embryo or in the endosperm that in consequence the hybrid seed is less viable. Finally, since fertility tends to be gen- erally higher in the more nearly euploid individuals, and the progeny of the aneuploids do not vary more in chromosome number than do the progeny of normal individuals, D. Belladonna Hort. maintains itself approximately at hexaploid level. Of the many allopolyploids which have been analyzed (Stebbins, 1947, 19 50) the one perhaps most similar to D. Belladonna is the segmental allopolyploid Anemone Jancewskii (= A. sylvestris X А. multifida var. magellanica), first Synthesized by Jancewski before 1892, resynthesized by Gajewski (1946), and turther, occasionally spontaneous in botanic gardens. Morphologically, the hybrid 1$ intermediate between the parents and is quite constant, even aneuploids not being of abnormal appearance. A cytological study of the parents and the hybrids by Gajewski revealed that in the triploid (24 chromosomes) at MI from 3 to 8 bivalents are formed, the remaining chromosomes being unpaired. In the F2 8neration produced by the functioning of unreduced gametes (Jancewski re- Ported two instances of somatic doubling but Gajewski did not obtain any hexa- [Vor. 40 150 . ANNALS OF THE MISSOURI BOTANICAL GARDEN ploids by this means), in addition to cells in which 24 bivalents occur, in other cells there were some univalents and from 1 to 3 multivalents. Seed and pollen fertility in different individuals of this generation ranged from rather high to rather low, and in the Ез there was a general increase in fertility. The chromo- some numbers were uniformly about 48 in Gajewski's plants, but in the accessions from botanic gardens they ranged from 42 to 48. The chief respect in which A. Jancewskii seems to differ from D. Belladonna and from the hybrids synthesized here is in the complete absence of trivalents in the triploid and in the higher frequency of multivalents in the hexaploid. The first may be due to greater dif- ferentiation of the parental genomes; the second may well be the result of the much higher chiasma frequency of the chromosomes of this hybrid. To summarize, then, although D. Belladonna has previously been considered an amphidiploid on the basis of its largely regular meiosis and its relatively constant appearance, intermediate between certain diploid and tetraploid species of cultiva- tion, the garden hexaploid has been shown to have somewhat less regular meiosis than is characteristic of true amphidiploids. "This is indicated both by the common occurrence of univalents and other irregularities in meiosis, and in the higher production of viable, fertile aneuploids than is characteristic of such plants. Rather, D. Belladonna may better be classed as a segmental allopolyploid, in which the low chiasma frequencies of the chromosomes tend to enforce bivalent formation, and morphological constancy is further enhanced by the quantitative mode of inheri- tance of many of the differentiating characters. DERIVED HEXAPLOIDS For descriptive purposes, all of those plants having in their ancestry one of the triploids and Belladonna have been termed "derived hexaploids.” The pedigrees of the major lines are given in text-fig. 2. First- generation Hybrids: Triploids< Belladonna Varieties. — e dangers of making decisions regarding the nature of polyploidy in 4 hybrid, whether it be a segmental allopolyploid or an amphiploid, have been point out by Stebbins (1947, 1950). Neither external morphology of the chromosomes nor the nature of chromosome association can be used as a single criterion 1 of the triploids has not been achieved, a compromise has been made in the produc- tion and study of hybrids between the triploids and Smith's Belladonna. By using the triploids as female parents and pollinating them with Belladonna 23 F, seeds were obtained, of which 7 yielded plants which reached maturity. Unfortunately, these hexaploids, which are of the greatest interest in this study» are not known in as great detail as is desirable, since all but one were lost in the 1953] 51 GAGE—DELPHINIUM X BELLADONNA HORT. TABLE VIII MEIOSIS IN THE DERIVED HEXAPLOIDS Metaphase I Anaphase I Anaphase II Е * 5 i- Ф £ " & к: i3 i 3 4 Chromosome] ,, 3 ы ue A я 3 ч Мыс мш ра | 8 аера ва аа ЕЗ Б дш €) 2| Зач ||| за 3 Ф во Е ча s „д ud E 5 „д8 Е 50 3 E м © я - ad = g Da ы ы g g ная e hi D] з ч © ы "- Б ч © aa a Ф.н O d wa E в E Я 8 в Е я Е i з | 238 $ у ате) n аа * Z| 28 | Ro eee а ow Of Os] €^ | x1 01 Ба Ж 51-213-3 48, 2F 45 24-19 3 1 23 1139 6 | 13 13.7 60 9 | 2 18.3 51-216-6 49, 2Е 6 | 24—20 2 0 — | — Largely regular سے ا — اس‎ 51-216-10 50 (7А) |29 | 24-22 8 2 | 24| 45 1 4 1.1 || About 15% irreg. 51-216-16 50 (7А) 31 24—20 4 0 24 65 9 9 27.7 — — — — 51-219-4 46 55 22-17 0 0 21 50 8 2 20.0 37 2 2 11.2 51-220-4 48 10 23—20 1 0 — 25 3 2 20.0 — — — — 51-221-11 48 20 24—21 0 0 23 82 8 8 19.5 70 5 13 25.7 51-222-21 47 33 23—20 1 0 21 83 17 4 25.0 45 7 5 26.6 51-222-38 48 23 4—21 0 0 23 32 7 5 37.3 22 7 3 45.4 51-223-5 47 33 23-21.. 2 1 23 — — — — — — — — 51-223-31 44 7 22—1 0 0 61 8 2 16.7 36 7 5 25.0 [Vor. 40 152 ANNALS OF THE MISSOURI BOTANICAL GARDEN unusually severe winter of 1951-52 following their first flowering, and most of the meiotic material available for study was a few buds embedded in paraffin. Of the plants which were surveyed, 50-20-1 (48-6-2 X Smith's Belladonna) formed mostly bivalents at MI, but 6-12 univalents were commonly present, and reduced chiasma frequency in the long chromosomes was observed. At first ana- phase, abnormalities were observed in 28.0 per cent of the cells studied, but in only 10.0 per cent of cells at anaphase П. Pollen viability was 41.4 рег cent, and а fair amount of seed was set both by selfing and by backcrossing to Belladonna. 51-209-1 (48-27-5 X Smith's Belladonna) exhibited fairly regular pairing at MI. Usually no more than 4 univalents were found, and frequently 24 bivalents were present. Anaphase I was visibly normal in about 45 per cent of all cases, and AII was somewhat more regular. Pollen viability was in the neighborhood of 85 per cent. This plant produced a heavy set of viable seed, whether selfed or further backcrossed to Belladonna. 51-213-1 (50-15-4 Ж Smith's Belladonna) was studied in sectioned material, in which only approximate counts could be obtained. There was apparently rela- tively good pairing at MI—many PMC's appeared to be quite regular, and usually no more than 4 univalents were evident. Several times, a single trivalent involving one of the larger chromosomes was observed. АГз were regular in 60 per cent of the cases, as were most of the AII's. Pollen viability was as high as 91 per cent. Again, both selfs and backcrosses to Belladonna proved successful. А sister seed- ling, 51-213-2, showed metaphase pairing with generally 2—4 univalents, the re- maining chromosomes being in bivalent configurations. AI segregation in this plant appeared to be about 55 per cent normal. 51-213-3, studied in some detail (Table VIII), had at least 25 bivalents in two-thirds of the cells at MI, and both anaphase divisions were highly regular. Little trivalent formation was recorded, however. Two other hexaploids, 51-211-1 and 51-211-2 (50-15-1 X Cliveden Beauty); could not be analyzed; of these, the former yielded a good set of seed upon selling: The other plant was not pollinated. The advanced generations among the derived hexaploids had their origin, 3 may be seen in text-fig. 2, from five Е; individuals. Of these, two of the parental triploids had different white elatum parents but the same blue grandiflorum parent, while the triploid parent of the third line was the offspring of an azure elatum variety and the fragment-bearing cobalt diploid, 47031-1. In general, as will be seen, the lines derived from the last cross are distinct in cytological behavior from the other two. Somatic Chromosome Morphology.— Because of the number of plants represented by these populations, studies of somatic chromosomes have been restricted to root-tip counts, The total number of each of the easily identified A and B chromosomes was counted, gross morpho- logical alterations of the chromosomes were noted, and in instances where staining was good enough, average satellite number and disposition were determined. — Nm Bom ел М. Су ОО D m “~N ~ --oo0o0-o0oo0 q1uxu$ej]p| * © 12101. = тосола EDERT EE D: -m 22430 LRReecclecce = edd < = кк = a 92 о © = © ©+зс о о mo я я = yZ № -"-cooonuocococo a 5 [5 `6 E E as o ooooocooococ & | E >= a t 5 vs © ooocococcooooo a н 2d a Á[uo А 3024321 "^ oocooocoooo < N Sz Áquo pU DEG щ 5 рю]4пэчу < e a8 = [201L n No \0 \о © о К. со са н ^ =“ : n S 3940 [М ооо ооо Z < m к г [| = gz © oocooo-oooo p 5 z `6 t Б УЕ о ^"ooooooo-oc — 5 > = о [2 vs © ооооноооо н е о Áquo Шат че Se OOS © с т Áquo pio[dnouy o он м о о омо № ЮМ. 900 Oy © -— сч c6 Ym сч we сє сч N C4 СЧ СЧ CN. mM t r ee ttt — ہس سم ہے س ہے‎ жч =ч е сч; A WA м м м м м м м м 'LHOH VNNOQGVTTI8 X JWüINIHdTAIQ—45Vv5 [2561 [Vor. 40 154 ANNALS OF THE MISSOURI BOTANICAL GARDEN As to the number of satellites present, the chief difference noted among the various populations was in the expression of the A-SAT, which is potentially present on two of the 6 A chromosomes in all lines, since the triploid and the Belladonna parent each should have contributed one. However, while plants in the three populations descended from the 51-213 individuals usually revealed one or two satellited A's, this satellite was not strongly manifested by the other popu- lations. Apparently, this difference had its origin in the triploid parents, since, as was mentioned above, the 50-15 individuals were characterized by possessing a more "active" A than the other lines. Aneuploidy is common in all populations, as it is within Belladonna; however, the various lines have characteristic differences (text-fig. 7, Tables IX and X). Thus, while Belladonna populations contain many hypoploids with as few as 41 chromosomes and only occasional hyperploids, this pattern occurred only in 51-219 and 51-220 (51-202-3 X self, and Х Smith) among the derived hexaploids. А somewhat intermediate situation was found in 51-222 and 51-223 (51-209-1 X self, and X Smith) and in 51-221 (51-202-7 X 50-20-1) in which plants with 48 chromosomes were frequent and both hypoploids and hyperploids occurred. A third situation was found in lines 51-216, 51-217, and 51-218, derived from ‘the fragment-bearing individuals 51-213-1 and 51-213-2. In these populations, hyperploid individuals were most numerous, though an occasional plant with as few as 45 or 46 chromosomes was found. In addition, the centric fragment was quite common, being found in 17 out of 24 individuals of 51-216, and 23 of 26 individuals of 51-218. In 51-217, which was a backcross to Smith's Belladonna, the fragment occurred in 4 of 8 individuals examined. The exact number of fragments present in each individual is not given, since the fragment does not always segregate regularly and may occasionally vary from root-tip to root-tip IRREGULAR DISTRIBUTION OF CHROMOSOMES IN THE HEXAPLOIDS Number chromosomes Number ch es deficient in escent : Number Line of plants : А В Other A B Other 51-216 24 0 с З {з : i» 51-217 9 0 ó А e A 51-218 26 0 i n : : He 51—219 12 0 1 18 0 d 0 51—220 7 0 $ 12 4 : i 51-221 24 0 { ; о : 51-222 45 0 1 18 4 i 6 51—223 20 0 М i : : : 51-224 15 0 x y А 0 : 52-311 25 0 1 49 б 1 0 51-232 25 1 F hes : : : Total 232 1 6 173 19 5 51 1953] GAGE—DELPHINIUM Х BELLADONNA HORT. 155 of a given plant. It is not known to what extent it may be possible to accumulate fragments in future generations, but in samples of two Ез populations, each of whose progenitors bore two fragments, no more than three fragments were ob- served in a total of fifteen individuals; on the other hand, all but one of the remainder possessed one or two fragments. Since transmission of the fragment through both the pollen and the ovule has been demonstrated (in the pollen, on the diploid and hexaploid levels, and in the ovule, on the triploid and hexaploid levels), it may well be that the failure to find many individuals with an augmented fragment number reflects some lethal effect of numerous fragments. It may be noted that the formation of isochromosomes by misdivision of the fragment, such as was noted by Darlington in Fritillaria (1940) and by Rhoades in Zea (1940) has never been observed here, although such a chromosome should be easily recog- nized by its distinct morphology. Fertility of tbe Derived Hexaploids.— А systematic study of all the plants has not been made, since they number now well over 1000. Аз has been noted above, the initial triploid hexaploid hybrids were of good fertility. In succeeding generations (to the F4 and various back- crosses to Belladonna) plants exhibiting all degrees of fertility have been found, as evidenced particularly by seed set. Detailed meiotic studies have been made in only a few of these plants ( Table VIII), but the degree of pairing has generally been found to agree well with that characteristic of D. Belladonna. ‘Thus, 51-221-11 forms no more than 6 uni- valents and usually only 2 at MI; no multivalents were noted; and both anaphases are fairly regular. A similar situation, with slightly more numerous irregularities in the anaphases, was found in 51-222-38. In aneuploid individuals, irregularities were, to be sure, more frequent. Thus, 51-216-10 and 51-216-16, both of which Possessed an A chromosome in excess as well as one other excess chromosome, often formed 24 bivalents, but were also characterized by low multivalent formation, chiefly involving the extra A chromosome. А hypoploid (51-219-4; 46 chromo- somes) averaged 21 bivalents, but the anaphases were not especially irregular. As in Belladonna, the number of laggards at anaphase was always much less than that of univalents at metaphase. While infertile plants have been found both among the aneuploids and among apparently normal plants with 48 chromosomes, a study of seed germinability among those plants which have set some seed shows that, on the whole, the highest number of viable seeds were produced by plants with 48 chromosomes. However, even in this group, germinability ranged from 0 to about 90 per cent, and, with опе exception, the production of seed with at least 40 per cent germinability was confined to plants having 47—49 chromosomes (Table XI). Segregation in the Derived Н exaploids.— On the whole, each line within this group exhibits considerable uniformity by the Ез (or equivalent) generation. However, there are few individuals in each 156 [Vor. 40 ANNALS OF THE MISSOURI BOTANICAL GARDEN TA SEED GERMINABILITY IN DELPHINIUM BELLADONNA AND DERIVED HEXAPLOIDS Chromosome umber Number % Germin- Рим umber seeds sown seedlings ability Smith X self 27 21 78 Smith X Bellamosum 48 M ? 86 25 29 Lamartine X self 24 14 57 Lamartine Х Smith 48 X 48 52 8 17 216-3 X self 8 4 0 0 51-216-4 X self 49, 2F 2 0 0 51-216-12 X self 48, 2F 75 15 20 51-216-13 X self 52 15 0 0 51-218-2 X self 49, F 8 3 37 51-218-9 X self 48 14 2 14 51-218-16 X self 48, 3F 42 23 55 51-218-18 X self 49, 2F 2 0 0 51-218-19 X self 46, 2F 29 18 62 51-218-29 X self 50, 2F 68 23 34 51-218-39 X self 48, F 8 2 25 51-221-2 X self 49 50 30 60 51-222-11 X self 48 3 0 0 51-222-13 X self 47 36 4 11 51-222-19 X self 48 11 0 0 51-222-28 X self 48 4 0 0 51-222-7 X self 48 200 165 82 51-222-8 X self 48 47 21 44 51-222-14 X self 48 23 17 73 51-222-36 X $ 48 12 0 0 1-222-37 X self 48 32 14 44 51-222-43 X self 47 22 3 13 1-222-46 X self 48 8 1 13 51-222-50 46 33 16 48 51-222-54 X self 49 20 1 5 51-222-61 X self 48 22 13 59 51-222-67 X self 48 5 0 0 51-222-77 48 9 4 45 51-222-95 elf 49 15 7 47 51-222-100 X self 48 ра 1 14 51-223-1 X self 49 3 0 0 51-223-21 X self 48 10 6 60 51-223-22 X self 48 15 5 33 51-223-38 X self 48 14 5 34 51-223-44 X self 47 7 0 0 group which deviate noticeably from the rest, especially fee SEE in leaf shape. In only one instance (an Fs family derived from 5 1-209) have more than a few individuals deviated in the direction of either D. grandiflorum or D. elatum, and the variation here is in the direction of rather more finely dissected leaves than generally occur in D. Belladonna. Taken as a whole, however, the general aspect of the assemblage is that of a Belladonna population in which a few individuals are more or 169 branched, shorter or taller, or with leaves more or less finely dissected than in D. Belladonna Hort. (pl. 10). 1953] GAGE—DELPHINIUM X BELLADONNA HORT. 157 Color inheritance in these forms is such as to raise some interesting questions, but detailed treatment of this subject will be reserved for future consideration. Discussion and Summary.— Although the analysis of meiosis in the first-generation hybrids was not com- plete, it is still clear from the above that while somewhat more irregular than in Belladonna, the meiotic process does not appear to be greatly disturbed, and in only one individual (51-213-1) was the regular production of multivalents noted. Chromosome numbers among the progeny of these plants suggest for the various crosses what homologies and disharmonies may have been operative in their parents. Crosses 51-219-, 51-220-, and 51-223- are most similar to Belladonna in that chiefly euploids and hypoploids were produced. In the first two of these lines, which represent the self and backcross to Smith's Belladonna, respectively, of the Р» of the cross 48-6-2 X Smith's Belladonna, there has been no intentional building up of the Belladonna complement by backcrossing. Therefore, unless the introduced chromosomes of the triploid have been appreciably eliminated, no strong barriers of chromosomal differentiation appear to exist between the parental triploid and Belladonna. The aneuploidy thus seems to be of the same nature as in D. Belladonna Hort. On the other hand, 51-223 is a first backcross to Smith's Belladonna of the original hybrid between 48-27-5 and Smith's Belladonna. Since the same hybrid When selfed gave rise to line 51-222 (in which both hyperploids and hypoploids were frequent, and in which excessive numbers of supernumerary А chromosomes were found), it seems in this case that occasional formation of multivalents in- volving the large chromosomes must occur, and perhaps that more pronounced structural differences may exist between the parental genomes than in the previous instance. The first conclusion is based on the occurrence of 4 А chromosomes out of a total of 10 chromosomes extra in the hyperploid individuals of line 51-222, more than three times the expectancy if this chromosome were segregated irregu- larly only as often as the small chromosomes, and even more unlikely as а random event if the higher chiasma frequency of this chromosome is considered. In the absence of any indication of multivalent formation by the smaller chromosomes, it is necessary to assume that some other factor is operative in Producing individuals possessing these chromosomes in excess. This may be ac- counted for if gametes hyperploid for these chromosomes are not at the selective disadvantage which appears to occur in D. Belladonna. Such an hypothesis can t be explained if certain euploid combinations in these individuals are actually deficient for necessary genetic material which may be supplied by the extra chromosomes. The extreme frequency of hyperploidy in the 51-213 derivatives, where hyper- Ploids represent the bulk of the populations, seems to indicate both high multi- valent formation by the A chromosomes, occasional multivalent formation by the B chromosomes, and to a greater extent than in the preceding case, the operation of some factors which favor the functioning of gametes hyperploid for the smaller [Vor. 40 158 ANNALS OF THE MISSOURI BOTANICAL GARDEN chromosomes. This is particularly true of 51-216 and 51-217, the Ез and first backcross respectively of 51-213-1, the plant in which multivalents were noted above. That transmission of the excess large chromosomes may be through the pollen grain as well as through the egg is suggested by the fact that a comparable excess of large chromosomes was present in 51-217, the backcross line, as compared with 51-216. As to the fertility of the derived hexaploids, it has already been noted that euploid chromosome numbers and fertility are not necessarily related, but that, on the whole, more plants which yield seeds with good germination have about 48 chromosomes than have chromosomes either much in excess or in deficiency of this number. The tendency to preserve the hexaploid number appears to be almost as strong as in the Belladonna lines themselves. So far as the homologies of the genomes derived from triploid and hexaploid sources are concerned, it seems that they must be very nearly similar, since there as been no strong morphological or cytological disharmony in the derived hexa- ploids as compared with D. Belladonna. In fact, the differences are very slight indeed if one contrasts the behavior of these plants with that of the intervarietal crosses between Belladonna and Bellamosum. Although the latter variety was derived from the former, yet reduced seed germinability and more frequent aneuploid production have been found to occur in their hybrids than characterize either variety when it is selfed, GENERAL Discussion It is concluded that D. Belladonna Hort. is an allopolyploid hybrid between the tetraploid species D. elatum L. and the diploid species complex of D. grandi- orum L. In general, D. Belladonna may best be classed as a complex segmental allopolyploid, but it should be emphasized that classification is especially difficult since strong chromosomal barriers to hybridization do not seem to exist between even distantly related species of Delphinium (e.g., D. elatum and D. nudicaule ot D. cardinale). Unlike many other segmental allopolyploids, however, D. Bella- donna is easily distinguished from both parents. It has already been suggested that regularity in pairing, enforced by differential affinity of the chromosomes, as well as by the not insignificant effect of low chiasma frequency, is probably re- sponsible in large part for this behavior. Moreover, the quantitative mode first occurrence of fertility was reported; second, it has been subject to the care and selection of gardeners and plant breeders over the years, so that the difficulties (for instance, multivalent formation by the large chromosomes) such a hybrid would encounter during the first few generations in nature were mostly obviated. 1953] GAGE—DELPHINIUM X BELLADONNA HORT. 159 Several explanations may be offered to account for the elatum tendencies noted in the experimentally produced triploids. First, while D. grandiflorum has a num- ber of varieties, those varieties grown in gardens are, on the whole, similar with respect to the features which they contribute to the phenotype of Belladonna— the large self-colored bee, the wide open flower, the few-flowered inflorescence, the finely dissected leaves, and much-branched habit. D. elatum, on the other hand, exists in many natural and horticultural variants; probably a D. elatum bred for a showy "bee," with a relatively open inflorescence and rather narrow leaf seg- ments, would yield offspring of the type characteristic of Belladonna when such a variety was crossed with grandiflorum. However, the elatum variety Summer Skies, seed parent of triploids 50-15, has a small bee, broad leaf segments, and very dense strong main spike with numerous flowers; the triploids plotted in text-fig. 5 are indeed intermediate with respect to their known parents. All one can say of the presumptive elatum parent of Belladonna is that it was almost certainly a smaller, less extreme variety than those grown today, judging from historical plates of Delphiniums cultivated in the middle of the last century. The question of the nomenclature of the triploid and hexaploid garden Del- phiniums may be viewed in three very different aspects: first, as it concerns the taxonomist, second, the geneticist, and third, the gardener or nurseryman. While the first two views may be reconciled to a certain extent, the unfortunate truth is that the people most concerned with these Delphiniums (which are, after all, cultivated plants) will probably perpetuate the past errors and contribute new cases of mistaken identity in the future. For the geneticist, it is desirable that the exact parentage, insofar as it is pos- sible, be given for every hybrid where it is known. Convenience in the herbarium, on the other hand, requires that some category be devised which will adequately contain the assemblage. Thus, according to the International Code of Botanical Nomenclature (Lanjouw, ed., 1952), "hybrids or putative hybrids between two species of the same genus are designated by a formula, and whenever it seems useful ог necessary, by a name." In the latter case, the specific name is to be preceded Y "X." Further, where different hybrids of the same parentage exist, they may be united in a collective taxon in which the forms are recognized as nothomorphs, "designated by an epithet preceded by the binary name of the hybrid group and the term "nothomorph." While the general practice in systematics is to give Priority to the earliest validly published name, for cultivated plants provision is made that the most generally employed name may be used. Of the early names applied to hybrids of D. grandiflorum and D. elatum, only D. formosum and D. Belladonna are in wide usage. Since the name Delphinium formosum was given in 1856 to a wild species (Boissier and Huet), it seems advisable to maintain D. X Belladonna Hort. ex Bergm. (see footnote, p. 119) as the epithet of the hexa- Ploid Delphiniums and their triploid allies. Varieties such as Lamartine, Capri, Moerbeimi, and the triploids and derived hexaploids of this study may then be Considered as nothomorphs of this species. [Vor. 40 160 ANNALS OF THE MISSOURI BOTANICAL GARDEN SUMMARY 1. Previous workers have postulated that D. Belladonna Hort. originated either as a variety of D. cbeilantbum Fisch. ex DC., or as а hybrid of D. cheil- antbum with some unknown other species, or by hybridization between such diploid and tetraploid species as D. grandiflorum L. and D. elatum L. It is pointed out that solution of the problem must be based upon a combination of methods, including historical research, morphological analysis, and cytogenetic data. 2. A study of the taxonomic and horticultural literature indicates that D. cheilanthum is a coherent diploid species, rare in gardens, native to Siberia, Mon- golia, and Kumaon (China), with relatives in adjacent areas; D. Belladonna, on the other hand, is a hexaploid, known only in gardens, where it has arisen upon different occasions. Hybrids between various lines of D. elatum and D. grandiflorum var. cbinensis have been produced and have been found to be morphologically com- parable to D. X Belladonna Hort. ex Bergm. and, moreover, within the limits im- posed by their triploid condition, interfertile with the "species." 4. Character analysis of the various species and hybrids indicates that both Belladonna and the experimentally produced triploids are intermediate with regards to D. elatum and D. grandiflorum, but that D. cheilanthum is morphologically and cytologically distinct from D. Belladonna and could not have been the diploid parent of the latter if D. elatum was the tetraploid parent. 5. Methods are given for the study of somatic and meiotic chromosomes and of pollen mounts. 6. It is pointed out that for the species and hybrids studied, size of pollen grain may be used as an indicator of polyploid level in both living plants herbarium material. 7. Study of the morphology of the somatic chromosomes has shown that the genomes of the two diploid types which have been considered possible parents of D. Belladonna may de distinguished on the basis of their satellite distribution. n D. grandiflorum, satellites occur on chromosomes C and Н, while in D. < antbum the A and the C chromosomes are satellited. Further, the length of the satellited arm of chromosome A in D. cheilanthum is longer than the correspon’ ing arm of the same chromosome in D. grandiflorum, suggesting that there bis been a translocation of the satellite-bearing segment in the evolution of the genom* 8. Differential activity of the satellites was found to characterize even diploid species: in D. grandiflorum the C satellite is about twice as “active” ^ the H satellite; in the closely related D. tatsienense, both are equally active. >: D. cheilanthum the A satellite is about twice as active as the C satellite. 9. The tetraploid species D. elatum was found to have a gametic complement corresponding to the summation of the grandiflorum and the cheilanthum genomes: Amphiplasty of the H satellite occurs. 10. In all of the experimentally produced triploids the potential satellite number was found to be 1 A, 3 C’s, and 2 H's, but in triploids of lines 48-27 and 1953] GAGE—DELPHINIUM X BELLADONNA HORT. 161 48-6, the A satellite was rarely evident on account of amphiplasty. The mor- phology of the chromosomes of Belladonna supports the hypothesis of similar parentage for these types, since the total Belladonna complement has double the potential satellite number of the triploids. In both the triploids and in Belladonna, amphiplasty of the A satellite was found in some individuals. 11. Meiosis in D. grandiflorum and D. elatum is largely regular, the chromo- somes forming bivalents (rarely some univalents in the latter species), and assorting on the whole regularly. 12. In the triploid hybrids, prophase and metaphase pairing is chiefly in univalents and bivalents, trivalents being formed only rarely and only by the A or B chromosomes. Chiasma frequency is somewhat reduced and both inter- genomal and intra-genomal pairing of the chromosomes occurs. Bridges and frag- ments are characteristic of both anaphase divisions, although the number of chromosomes lagging at anaphase is considerably less than that of univalents at metaphase. Micronuclei are frequent, and rare viable unreduced pollen grains are produced. 13. Meiosis in Belladonna is more or less regular, but the production of uni- valents commonly occurs, and aneuploid gametes are produced and function. Chromosome numbers as low as 41 have been found, but ordinarily only those plants with about the hexaploid number are fully fertile. It is suggested that D. Belladonna is a segmental allopolyploid which preserves a uniform appearance because low chiasma frequencies of the chromosomes enforce bivalent formation and because of the quantitative mode of inheritance of many characters. 14. Lines of derived hexaploids have been produced by pollinating various triploids with pollen of Smith's Belladonna, and further selfing or backcrossing of the plants produced. On the whole, these lines are only slightly more variable than is D. Belladonna, and fertility is only slightly impaired as compared with that "species." A small amount of multivalent formation characterizes some of es. _ 15. It is concluded that D. X Belladonna Hort. ex Bergm. embraces the en- ure assemblage of present-day varieties derived from crossing D. elatum with D. grandiflorum, as well as а number of hybrids of the last century which are no longer in cultivation. 16. A telocentric fragment has been transmitted from a D. grandiflorum Pollen parent through four generations. No isochromosomes have been observed. BrBLIOGRAPHY ее, (1907). Roy. Hort. Soc. Jour. 32:lxxviii. m duds » (1911). Delphinium Moorheimii. Gard. Chron. Ш, 50:352-353. ее Oe Miscellaneous annuals etc. at Wisley. Delphinium. Roy. Hort. Soc. Jour. ды с! (1949). А tentative checklist of Delphinium names. Publ. Roy. Hort. Soc., London. Don, E. (1948). Introgressive Hybridization. John Wiley & Son, М. Y. ; : reer (1951). Concordant versus discordant variation in relation to introgression. Evolution 1133-141. [Vor. 40 162 ANNALS OF THE MISSOURI BOTANICAL GARDEN Anderson, E. (1952). Plants, Man, and Life. Little, Brown & Co., А and A. Gage (1952). Introgressive hybridization in hg ‘bifida, Amer. Jour. Bot. 39:399—404. Wed ae m TR. QM. j ENET for recording and analyzing variations of inter- node . Mo. . Gard. 31:239— Ashby, Е. (зө). didis i in a ê Ati БА leaves, I. An essay on leaf shape. New Phytol. 7:153—17 Bailey, L = (uno ). The ovis of Larkspurs. € iens Co., New York. = . 7. Bailey (1930). Hortus. я Millan ‚ New York. Beckman, I. ane oo. eg peng FE aga уды tes ae orientale, Hereditas 11:107-128. Bergmann xs үгө, Vaste Planten Ap DEDE Haarlem Bro (1944—5). Studies of developmen in larkiper. E тона sequence in the first ten кыы Bot. Gaz. 106:103—1 Brühl, id (1896). Some new к DM Minus from India and adjacent regions. Roy. Bot. d. Calcutta, x 5:7 ы В. C. T авары (1928). пуч, of Digitalis ambigua апі Digitalis purpurea, their рч and cytology. Jour. Genet. 19:2 Clausen, J., D. Keck, and . Hiesey (1945). ока tal studies on the nature of species. I. Plant nr evolution Sisi, мары dy and autoploidy, with examples from the Madiinae. Carn. Inst Publ. Clausen, R. E., and D. R ; Cameron (1944). Inheritance in Nicotiana tabacum. XVIII. Monosomic ore. “Generics 29: rd =, Я ( се. oe in Nicotiana tabacum. IV. The trisomic Ви кы ” Ibid. 9:181— — — ——, (1925). Interspecific h e А in Nicotiana. П. А PET se n experimental verification of Winge’s n 10:27 ь (1937), Advances їп Cytolog ., London › (1939). Misdivision and the genetics of the scienti Jour. Genet. 37:341-364. › (1940). The origin of isochromosomes, Ibid. 39:351-361. La Cour (1947). The Handling of Chromosomes. 2nd ed., London. ili 340- 4 De Candolle, A. Р. (1818). Regni vegetabilis systema naturale 1:340—364. ——, . Prodromus systematis naturalis regni € per Dietz, R. (1952). Variation in the perfoliate Uvularias. Ann. Mo. Bot. od. 39:219-247. Elliot, E. (1950). A new phase of amphiplasty in Leon bag New Phytol. 49:345—349. Emsweller, $. L., and pas n (1938). Crossing-over, fragmentation, aid formation of new hromosomes in an species hybrid. Bot. Gaz. 99:729—772. Engler, A., and K. d d elphinium, in Die satüclichén Pflanzenfamilien 32:59-6 Epling, C. «i : = Н. Lewis (1952). Increase of the adaptive range of the genus Delphinium. Evd- поп 6:253— Ewan, J. Sade DR A synopsis of the North American species of Delpbinium. Univ. Colo. Studies, мее К. (1929). Die Neue sire Berlin к — (1933). Foerster's strain of Del phiniums. Delphinium. L. Н. Leonian, ed. рр. Mo Gajewski, У. (1946). Cytogenetic reden on Anemone L. I. Anemone janczewskii, а amphidiploid species of hybrid о Soc. Bot. = Acta 17:129—194. Gates, R. R. (1942). Nucleoli and а) nucleolar structures. Bot. Rev. 8:337 Glehn, P. v. а T hniss der im Wit isd chins unde < На, pu Hortus Petropol., Acta 4:1—96. Gmelin, J. G. (e Flora Sibirica sive historia plantarum Sibiriae 4: nn. m Ta 76. Goods - H., and M. V. Bradley (1 и). ся a Rev Goualt, A (1854). Delphinium Henderson Gray, A e Delpbinium, an ipit to ie vate Er ме American species. Bot. Gaz bec ‚ W. x deni riis umen and cytological studies in the Ranunculaceae. Am. Phil. 50 rans., Hall, M. T. Р pur Variation and hybridization i in aoe Ann. Mo. Bot. ане 39:1-64. Harrison, I à ad Delphinium Hendersoni. Flori . Cab, ge Flor. Mag. 21 - Questions, answers, and rem ite UR 21:85. licher ке E d " i Die Ursache der gesetzmassigen Zahl, Lage, Form und Grösse pf kleolen. фера 12:775—84. 9 de e 24), Observations sur le nombre des chromosomes chez quelques Renonculacée end. 8 Huth, E. ( 1895). аена а Gattung Delphinium. Engler’s Bot. Jahrb. 20:322—499. 1953] GAGE—DELPHINIUM X BELLADONNA HORT. 163 Kagawa, F., and Nakajima, G. (1933). ивы and cytological studies on species hybrids in amoclit. Jap. Jour. Bot. 6:315—32 Kostoff, D. (1939). Studies on uvis Vie. XXI. Cytogenetic behavior of the allopolyploid ybrids ida aes Grah. X N. Langsdorffii Weinm. and their evolutionary significance. Jour. Gene penes У. E. oo 455 Жош. ce of monosomics in Nicotiana rustica. Genetics 17:689—696. Langlet, O. F. К (1927). Mes mae zur Zytologie der ت‎ анне Svensk Bot. Tidskr. 21:1-15. —‚ (1932). Uber Chromosomenverhiltnise und Systematik der Ranunculaceae. Ibid. Lanjo не, У (1952). regir see code of botanical Mappe Utrecht. Lawrence, hg tG 6). the origin of new forms in Delphinium. PM 18: pre Ledebour, C. F. "eg Flora Rossica sive enumeratio plantarum in totius Rossici. 1:60—65; Lemaire, s um. Delphinium cheilanthum var. Chauvieri (var. iE Sort. dus. "rd. egt Lemoin 1914). Catalogue ме 186. Мапсу, Е TN ۳ ur » Cytological studies in the Allium paniculatum group. Hereditas d 317-370. H. (1 947). Leaf variation in Delp inium variegatum 1. Torr. Bot. сес Ж ; C. Epling, С. А. L. Mehlquist, and С. С. ria d (1951). Chr mbers of Californian Delphiniums and their geographical occurrence. Ann. Mo. Bot. Gard. 3%: 101—118. Lewitsky, С. A. (1931). The karyotype in systematics. Bull. A Appl. ae 27:187—240. Lindley, J. (1820). Delphinium cbeilantbum. Edwards’ k Reg. 6:1. 473. 1837). Delphinium Barlowii. Ibid. 2 234. 1944. Linnaeus, с. i d in Species Plantarum 1:530—5 Love, M., a Pa i bea see е и NR hybrids and eir i ie arti n . Вог. 32:451—456. McClintock, B. (1934). The eal relationship of a BM ун chromosomal element to the development he nucleoli in Zea Mays. Zeitschr. f. Zellfor vd did — à (1941). The нета of эн ends of ch Mays. Genetics 26:234—282. Mehlquist, G. A. L., C. O. Blodgett, and L. Bruscia ; (1943). peter induced tetraploidy in el phinium В Jou E Hed. 34:187—192. Meurman, O., and E. Ta (1939). Stia on the chromosome morphology and structural ridity > А, genus Clematis. Cytologia 10:1-14 Meyer, к es 945). Prefixing with paradichloedbenzene to facilitate chromosome study. Stain Pini. Moore, T (аз). Delphinium cheilanthum var. Hendersonii А. Henfr. Gard. Mag. Bot. 2 üntzing, А. (1932). Cytogenetic investigations on synthetic Galeopsis Prosar Hereditas té: diis: 154 — ——5 (1937). The effects of chromosomal variation in Dactylis. “ат, d HT Myers, У. M. ed ). Crilogr and genetics of forage grasses. Bot. —421. 7 ——, and H. D. Hill (1940). Studies of chromosomal association ==” fedet н“ трлр of пасагу іп autotetraploid grass species, nbus grass, tall oat grass, and crested whea Brass. he 102:236—255. Navashin, м" (19 34). Chromosome alterations caused by hybridization and their bearing upon certain ке genetic problems. Cytologia 5:169—203. AN 2 С. F., and С. Pellew (1929). Primula kewensis and its derivatives. Jour. Genet. 20: Paxton V. ODE Delphinium Barlowii. Paxt. Mag. B , . Mag. Bot. рь С. A. (1949). Delphiniums, Their History and Culti tie Eyre & Spottiswoode, London. 939 i h сораса, ). Was ist eigentlich unsere Rittersporne? Gartenschónh. 20:283—284. ш, Кр 40). СР Chromosomenzahlen von Delphinien und ihre Auswertung für die ung der Gartenformen. Die Gartenbauwissenschaft 14:642—650. Regd, p nmi utiles ung дег. Baikalien, Dahurien, und am Amur ..... Gesammelten Rh . Abt. Soc. Imp. Nat. Mace Bull. 34: 1-211. : Ў "d › M. M. (1940). — a a арша. seem in maize with reference to t t Mak, Bc ri В. (1933)? "The first red Delphinium," in Delpbinium, L. Leonian, ed. sum 35737. ischkin, В. К. (1937). “Del hinium," in V. L. Komarov, Flora "URSS 7:101 pie E. s дч, ploi — of wheat. II. Additional chromo- berrations in Tri vulgare. Genetics 29:232—24 Schrank, P. von (1819). Vorrei, cheilanthes.” Plantae asain Horti Academici Monacencis. Sim, б ©. Y. (1942). Bivalent pairing in an induced tetraploid of Tradescantia. Genetics 27:635- [Vor. 40 164 ANNALS OF THE MISSOURI BOTANICAL GARDEN Stebbins, G. L. (19 m - be жи their classification and significance, in М. Demerec, Advances in Genetics 1:4 ‚ (1950). Variation pene. dum on in Plants. Columbia Univ. Pres, New York. —— ———, (1952). The — - ыы plants = weeds. Evolution 6:445—448. Tjebbes, K. (1927). The chrom es of three Delphinium species. a E e MP. Tournefort, J. P. (1700). Delph Вы um. Р ФА Aloüetie, in in Inst. Rei Herb. Trautvetter, E. R. (1847). Florula taimyrensis phanerogama, in A. T. аа, Sibirische Reise 1:17—64 » (1877a). Plantas Sibiriae borealis ab A. Czekanowski et F. Mueller. Hortus Petropol. Acta 5:1-1 —— ——, (18 77b). Flora riparia Kolymensis. lbid. 5:497-574. Turczaninow, N. (1842). Flora Baicalensi-dahurica . . . Part L Soc. Imp. Nat. Moscou, Bull. 15 Li Vanderbilt, N. Е. (1933). American Pioneers. I. The Vanderbilts. In Delphinium. L. Leonian ed." Van r ht 1857). Delphinium formosum, Hortul. Fl. des — en - ^ і 1783; Wilde, E. к ). Studies of се genus -— een Exp Wod APPENDIX CHROMOSOME NUMBERS OCCURRING IN THE SPECIES OF DELPHINIUM О Ru NaN ых Species 2n number Author and date Ajacis L. 16 reges 1927; Tjebbes, 1927; Gregory, 1941 Andersoni Gray 16 Lew: al, azureum Michx.* 32? pao 1931; Gregory, 1 Belladonna Hort. 48 Langlet, 1927; Lawrence НН Propach, 1939, 1940 lochmanae Greene 16 LH а " ay brachycentrum Ledeb. 16 Langlet, Brunonianum Royle 16 IE UY Propach, 1940 bulleyanum Forrest 32 Gregory, 1941 californicum T. & G. 16 Lewis е? al, 19 cardinale Hook. 16 Bonk 1940; Mehlquist, 1943; Lewis e£ = Ре cardiopetalum D.C. 16 Tjebbes, Pe Lewitsky, 1931; Gregory, 1 arolinianum Walt.* 32? Gregory, casbimirianum Royle Propach, 1940 саисаѕі Меу. 16 Gregory, 1941 cheilanthum Fisch. ex DC 16 Propach, 193 Consolida L. 16 Hocquette, ЧИ т) ebbes, 1927; MP 1927; 1; Gregory, 194 cuyamacae Abrams 16 Lewis e£ al, ч decorum Fisch. and Mey, ssp. Tracyi Ewan 16 Lewis e£ al, 1951 Delaveyi Franch. 16 Lewitsky, 1931; Propach, 1940 Dubmbergi Huth 32 Lewitsky, 1931 16 Propach, or 32 Lewitsky, 1931; Propach, n O 1941 e ак. & Ки. 32 сон, DAS Propach, 1 flexuosum Raf. 32 Gregory, 1941 formosum Boiss. & Huet 32 Эзел 1941 gayanum Willmott 16 Gregory, 1941 glaucu $. 16 Lewis e£ al, 1951 gracilentum Greene 16 Lewis ef al, 1951 grandiflorum L. 16 Propach, 1940 dp Qu Mr >. — ener iier igo: has determined these American species as diploid. His studies of plants tive habitat, whereas the other determinations were probably made from material ae out ү т European botanical gardens, which often con fuse D. elatum with D. carolinianum and D. azureum. 1953] GAGE—DELPHINIUM Х BELLADONNA HORT. 165 Species 2n number Author and date үе Ewan 16,32 | Lewis e£ al, 1951 p pem ene 16,32 | Lewis e£ al, 1951 besperium 16 Lewis e£ al, 1951 besperium н Зог Ewan 16 Lewis e£ al, 1951 iliense Huth 16 Propach, 1940 inopinum (= D. parishii var. inopinum Jeps.) 16 Lewis e£ al, 1951 Lamartine Hort. 48 Lawrence, 1936 Moerbeimii Hort. 24 Lawrence, 1936 nudicaule Torr. & Gray 16 Tjebbes, 1927; viri 1940; Gregory, 1941; Lewis et al, 1 Nuttallianum Pritz. 16 Lewis e£ al, 195 Т ientale (?) orientale Losc.; ‚ orientale S. Gay) 16 Beckma = 1928; n 1941 Parisbii Gray 16 Lewis, 951 Parryi Gray 16 Lewis, т al. 1951 Parryi ssp. seditiosum Jeps.) Ewan 16 Lewis, т 1951 batens Benth. 16 Lewis, 1951 Penardi Huth 16 Ск, ч ЪЪ 16 Gregory, 1941 bictum Willd. 16 Gregory, 1941 bolycladon Eastw. 16 Lewis e£ al, 1951 Purpus dg 16 Lewis e£ al, 195 recurvatum Greene 16 Lewis e£ al, 1951 Ruysii Ho 32 Lawrence, 1936 scopulorum Gray 16 G 1941 Specios ?) Ф speciosum Janka — m L.; ; speciosum "Bieb. 16 Langlet, 1927 Stap bis sagria L. ! 16 унво A и Langlet, 1927; Lewitsky, 1931; ; Gregory, ips sulph cn Boiss. & Hausskn 16 Gage, this p fatsienen е Franch. 16 Lewitsky, 1931 Propach, 1940 tric corne: 16 Gregory, 1 trolliifolium Gay 16 Lewis e£ x ibi truncatum Lang. 16 Langlet, 1927 Ашын, Curran 16 Lewis e£ al, 1951 umbraculorum 16 is e£ al, 1951 айн ots Torr. & Gray 16,32 | Lewis ef al, 1951 Yunnanense Franch 16 regory, 194 Zalil Aitch. & Hemsl. 16 Gage, this paper ee ee ee ence CHROMOSOME NUMBERS OCCURRING IN THE HYBRIDS DETERMINED BY H. PROPACH, 1940 с Variety Breeder or Source Chromosome Number п. Eisberg E. Benary, Erfurt 16 Blauglut K. Foerster, Bornim 24 Blautanne » 24 Al e x = Berghimmel Ж » 1а 4 » Blickfang x 166 [Vor. 40, 1953] ANNALS OF THE MISSOURI BOTANICAL GARDEN in blaues Wunder De Enzianherold Enzianturm Ernst von Borsig Fón ;lasfenster Gletscherwasser Gnom Grossenwahn Gute Nacht ichflam Quem Clothilde empelgong aei ad У 1 jov musoz g ~ پا‎ p Fig. 2. Delphinium Hendersoni, from Revue Horticole. 1854. Fig. 1. Delpbinium cheilanthum Rariores Horti Academici i Monacensis,’ pl. 52. K. Foerster, Bornim ээ EXPLANATION OF PLATE PLATE 7 Fisch., PRS from von Schrank, ‘Plantae 1819 PLATE 7 ANN. Mo. Bor. Garp., Vor. 40, 1953 رر ررر As‏ a PS PS EE a Ki GAGE- DELPHINIUM BELLADONNA HOR T.