eR se RE Pk $B cs 5 | ees CT tes DEPARTMENT OF AGRICULTURE. BUREAU OF PLANT INDUSTRY—BULLETIN NO. 210. B. T. GALLOWAY, Chief of Bureau. HINDI COTTON IN EGYPT. . BY 0. F. COOK, Bionomist in Charge of Crop Acclimatization and Adaptation Investigations. ; Issurp May 11, 1911. - WASHINGTON: GOVERNMENT PRINTING OFFICE. bp Fa Ea rey at hy, ey ra barn | ceo EPA oie N Tl OF AGRICULTURE. BUREAU OF PLANT INDUSTRY—BULLETIN NO. 210. B. T. GALLOWAY, Chief of Bureau. PET COTTON IN EGY PT BY OnE COOK Bionomist in Charge of Crop Acclimatization and Adaptation Investigations. IssuppD May 11, 1911. A NW WASHINGTON: GOVERNMENT PRINTING OFFICE, LOT BUREAU OF PLANT INDUSTRY. Chief of Bureau, BEVERLY T. GALLOWAY. Assistant Chief of Bureau, WILLIAM A, TAYLOR. Editor, J. E. ROCKWELL. Chief Clerk, JAMES I. JONES. Crop ACCLIMATIZATION AND ADAPTATION INVESTIGATIONS. SCIENTIFIC STAFF. O. F. Cook, Bionomist in Charge. G. N. Collins, Botanist. L. Lewton, Assistant Botanist. II. Pittier, Special Field Agent. A. T. Anders, J. H. Kinsler, Argyle McLachlan, and D. A. Saunders, Agents. C. B. Doyle and R. M. Meade, Assistants. 210 9) = el LETTER OF TRANSMITTAL. U. S. DeparTMENT OF AGRICULTURE, Bureau or Puant Inpvustry, OFFICE OF THE CHIEF, Washington, D. C., January 15, 1911. Sir: I have the honor to transmit herewith a paper entitled * Hindi Cotton in Egypt,” by Mr. O. F. Cook, of this Bureau, and to recom- mend its publication as Bulletin No. 210 of the Bureau series. This paper reports the results of a visit to the cotton-growing districts of Egypt in June and July, 1910. It shows that the ad- mixture of inferior Hindi cotton is a serious burden upon the Egyp- tian industry and that our more intelligent farmers can secure an important advantage through the improved system of selection that has been developed by experiments in Arizona. A careful compari- son of the results of the Arizona experiments with the conditions actually’ existing in Egypt became necessary in order to determine whether a satisfactory degree of uniformity has been attained in our acclimatized strains of Egyptian cotton. A previous study of the problem of diversity of the Egyptian cotton had been made in Arizona, as reported in Bulletins Nos. 147 and 156 of this series. Respectfully, Won. A. Tay or, Acting Chief of Bureau. Hon. James WILSon, Secretary of Agriculture. 210 MPCE ECR Coa Sener ae een eee es eee Rs eo. See Sto Neo ne ae ca ees Sa Sh es0 importance oi uniformity in Eeyptian cottom-.-:.-..-........22.-...<-.:.-5-- mand seed) charactersiol Hindi cotton. <5 -..cicci- jee 2 -e~ see ses ence cs Distinctive characters’ok Handi plants: 2.2 osi.ce2 0-5. oss. 23d osc eee se ncn - ok He hier Oram ia Ol Maing COUloms <0) noes 2 Jeo. sce Soe ees Soca nes Heedieb rac tens Ol Mim dl: COOMe <2 ee acie =. 52 Lk 5 a Hae cet wid wie urse ee oe HGtitychana ckerdoreran GilncOubOn as sasee ee erat oe ye epee se ele = Enevalenee of Hind cottondn Moyptss..22tinc S. Sec cog. onsen coe eae sien’: Chraractenstote rum clislaay nid Seas teeta ey ery Pye ee verge ae a eee Estimate of damage from the Hindi contamination......................-.--- Other causes of deterioration of the Egyptian crop..........-.....--+-----+--- Prospects of Egyptian cotton in the United States ............-...-.-.......-- Creel SLO 116 = eae ee tera) oe ee ee ee iin lee a Peak Cece (alee Puate I. Tt. ie TV. Wie VI. LEEUS TER ADT ONS: Fig. 1.—A field of Egyptian cotton intermixed with Hindi. Fig. 2.— Fig. 1. in cotton field at Benha, Egypt. Fig. 2.—Large cotton field at Benha, Egypt, with natives irrigating. . : Bracts and calyxes of Hindi cotton from Mcsopetumis eal con Hayum, Hey pte ies tcc accede es cme > oe chs eens oe ee oe Bracts and calyxes of Egyptian cotton and of a Hindi hybrid.....-. Bracts and calyxes of Hindi-like Upland cotton from Cochin China - and of a relative of the Egyptian cotton from central Africa. ----- Bolls:of Hey ptian and of Himdiicottomss:s52e-25-5--65 92.2 = ee 210 6 54 54 B. P. I.—643. PN I COrTON IN EGYPT. INTRODUCTION. Inspection of many cotton fields in different parts of Egypt shows that the so-called Hindi cotton is a general contamination of the Egyptian stock, responsible for a large amount of diversity and de- generation. Expression of inferior Hindi characters renders many of the plants not only worthless from the standpoint of production, but dangerous to future crops. The establishment of a profitable culture of Egyptian cotton in Arizona and southern California de- pends largely on the exclusion of the Hindi contamination." The Hindi cotton complicates the problem of acclimatizing and adapting the Egyptian cotton to the cultural conditions found in the United States. In this case a problem of heredity had to be studied. Instead of the physical factors alone, it has been necessary to analyze the characters of the plants in order to determine the causes of im- purity and find means of elimination. a“ Ffindi is the name applied in Egypt to an undesirable type of cotton with a short, weak fiber, that injures the high-grade Egyptian varieties by infesting them with hybrids. The skill and cheapness of the native Egyptian labor enable the exporters to have the cotton sorted by hand in their baling establishments, so that a high reputation for uniformity has been secured in spite of the Hindi admixture. - “The introduction of the Egyptian cotton into the United States brings also the problem of the Hindi cotton, but without the resource of cheap labor which enables the difficulty to be surmounted in Egypt. The practicability of estab- lishing a commercial culture of the Egyptian cotton in the United States de- pends largely upon the elimination ®f the Hindi contamination and other forms of diversity, so that the fiber may be produced in a satisfactory condition of uniformity. The Hindi cotton problem might be compared to that of the red rice that mixes with the white and depreciates the value of the crop. In the case of the cotton, there is a better prospect that adequate knowledge of the vegetative characters may enable the undesirable plants to be removed from the fields without too seriously increasing the cost of production.” (See Circu- lar 42, Bureau of Plant Industry, U. S. Dept. of Agriculture, entitled “ Origin of the Hindi Cotton,” 1909, p. 3. This circular contains the results of a previ- ous study of the Hindi cotton made in connection with experiments in Arizona. It will be sent free on application to the Secretary of Agriculture. ) 210 8 HINDI COTTON IN EGYPT. During the first years of its cultivation in Arizona the Egyptian cotton produced only small yields and rather inferior fiber. After the yield and quality began to improve, an undesirable amount of diver- sity appeared. A study of this diversity showed that it was due in part to hybridization with the common American Upland cotton, and that this danger was unusually serious in Arizona when the two kinds of cotton are grown in the same locality, owing to an unusual abun- dance of wild bees. The Hindi cotton is an additional factor of diversity inherent in the imported Egyptian stock, more difficult to understand because not previously known in the United States. Experiments show that both of these sources of diversity can be eliminated by a more careful system of field selection, applied early in the season before the inferior plants have begun to flower, and hence before they have cross-fertilized the neighboring plants. The value of the Arizona Egyptian cotton and the prospects of cultivating this crop on a commercial scale in the United States depend largely on the degree of uniformity that can be attained in the fiber, in com- parison with that of the Egyptian product. Hence, the necessity for an inspection of the cotton fields of Egypt in order to determine the extent of diversity in the crop as raised in that country. The high cost of labor in the Southwestern States forbids any direct imitation of Egyptian methods, either in raising the crop or in preparing it for market. Other solutions of the problems of production have to be sought. The requirement of uniformity has been met in Egypt by a system of careful grading of the cotton after picking that would be very difficult to establish in the United States, and too expensive to leave any assurance of profit for the farmer even if it were established. The Egyptian cotton trade is organized on an entirely different basis from the American. Instead of merely ginning and baling the farmer’s cotton as he brings it from the fields, it is the regular practice of the Egyptian ginning establishments to buy the seed cotton from the farmer and prepare it for the market by sorting, grading, and blending. Instead of depending entirely on samples, as with Ameri- can cotton, Egyptian cotton is sold largely by the marks or brands that are placed on the bales by the ginning establishments. Cotton of the same mark is supposed to represent a definite uniform quality. This is much more practicable in Egypt than it would be in most parts of the United States because of the much greater uniformity of climate and soil in Egypt. In comparison with the wide range of soils, climates, and seasonal vicissitudes in the cotton-producing districts of the United States, the Egyptian cotton industry gives at first an impression of com- plete uniformity. Although people in Egypt supposed that cotton would be more advanced in Upper Egypt than about Cairo, this did 210 IMPORTANCE OF UNIFORMITY IN EGYPTIAN COTTON. 9 not prove to be the case. It is quite possible that the crop of Upper Egypt comes to maturity earlier in the fall, owing to hotter weather in the summer, but there was very little difference at the middle of June. The effect that would naturally be expected from higher day temperatures in Upper Egypt may be neutralized in the early part of ‘the season by cooler nights, due to the greater radiation allowed by the drier air. In any event the cotton was found at nearly the same stage of development about Beni-Suef as about Cairo and Tanta. (See Pl. I, fig. 1.) Even at the middle of July much of the cotton in Upper Egypt, between Beni-Suef and Minieh, was still quite small, having searcely reached the flowering stage. In some fields the plants were only 6 or 8 inches high. The same was true of many fields in Lower Egypt in the region of Mansurah. (See Pl. I, figs. 1 and 2.) To what extent the later planting was respon- sible for the more backward state of the cotton in these districts was not learned, nor the reasons that may exist for later planting. The most important local differences perceptible in Egypt were not those of the external conditions or of the methods of cultivation. The superiority of the cotton raised in the Delta region may be due in part to superior conditions, as generally assumed, but better knowl- edge of the Hindi cotton among the native cultivators is another factor of great importance, since it determines whether the inferior Hindi cotton shall be rogued out or left to mature in the fields. Many native cultivators at Beni-Suef pay no attention to the Hindi cotton, while about Mansurah it seems to be known to everybody. But even about Mansurah the human factor is by no means uni- form, as shown by widely varying proportions of Hindi cotton in the different fields. IMPORTANCE OF UNIFORMITY IN EGYPTIAN COTTON. The requirement of uniformity increases with the presence of other good qualities of cotton. A long, strong cotton commands higher prices, because it can be spun into stronger or finer thread and used to make stronger or finer fabrics. An admixture of short, weak fibers not only reduces the strength of the threads and impairs the quality or durability of the fabric; it interferes also with the work of the spinning and weaving machinery by the more frequent breaking of the threads. The superiority of the Sea Island cotton does not consist alone in its length and strength, but in its extreme uniformity. This is main- tained by a highly developed system of selection, well recognized among the Sea Island planters but not vet applied to any other commercial type of cotton. The seed for each season’s crop is raised by itself, apart from the general planting, and traces its ancestry 210 10 HINDI COTTON IN EGYPT. back to a single superior individual of two or three generations before.¢ In the Egyptian system of cotton culture no attempt seems to be made to imitate the methods of the Sea Island planters. Even less consideration is given to selection than in the Upland-cotton industry of the United States. While very few planters of Upland cotton have been accustomed to select their own seed, it has at least been possible for them to buy seed of selected stocks of many of the Upland varieties, whereas planters in EKevpt do not appear to have any recognized source of supply from which to secure uniform stocks of seed of the Egyptian varieties free from the Hindi contamination. Differences between the seeds of the Hindi and the Egyptian cotton enable a selection to be made, even after ginning, but it seems evident from the condition of the fields in Egypt that a considerable quantity of Hindi seed must be planted and that many Hindi plants are allowed to grow to maturity and so to maintain the contamination.? The advantage that the individual planter might gain by a careful and persistent selection of his own seed is difficult to realize under the Egyptian system of selling the seed cotton to the ginner. There is also a custom of exchanging seed between different villages on the theory that better yields can be obtained in this way. Thus growers of Mit Afifi cotton near Mansurah obtain their seed from Kefir Zeyat, between Tanta and Alexandria, a place that is commonly supposed to produce seed of a superior quality. Such exchanges of seed are @Webber, H. J. Improvement of Cotton by Seed Selection, Yearbook of the Department of Agriculture for 1902, p. 374. b“'The seed reserved for sowing is passed through special riddles, which re- move small and dead seed; purity can not be obtained by this means, but merely a better looking sample; that is to say, as far as general appearance is con- cerned, the sample may be excellent, but closer examination reveals the pres- ence of seed not true to variety. Small cultivators do not, as a rule, trouble even to secure the best seed which is procurable, but content themselves with the employment of that resulting from the ginning of common qualities of all pickings, regardless of origin and purity. Were this seed purchased at a low price it would provide no excuse for such a short-sighted policy, but even this is not the case, the price paid to the village merchant being, as a rule, considerably higher than that for which the better qualities could be obtained. “In order to overcome this difficulty, the Khedivial Agricultural Society, in conjunction with the Agricultural Bank, distributes annually to small culti- yators the best seed obtainable at cost price. the value of which is collected at the end of the following cotton season. “Tt must be remarked, however, that the seed so distributed is merely the best that can be procured. “That it is vastly superior to that which in the absence of such a system of distributing would be employed is without doubt. At the same time this system does nothing to actually improve the seed.” (See Foaden, G. P., “ The Selection of Seed Cotton,” Yearbook of the Khedivial Agricultural Society, 1905S ps2) 210 LINT AND SEED CHARACTERS OF HINDI COTTON. ali well calculated to preserve and distribute the Hindi contamination. Even the introduction of new, carefully selected varieties could be expected to give only temporary improvement unless the whole sys- tem were changed. The process of deterioration would be resumed at once as a result of the crossing between adjacent fields of different varieties and the exchange of seed between different localities. After selection is relaxed the rapidity of deterioration of a vari- ety of cotton depends on two cooperating factors, variation and cross- ing. Both of these factors must vary in different places, for they are influenced by external conditions. When cotton is grown under new or unfavorable conditions, more numerous variations appear. Abundance of bees or other cross-fertilizing insects causes a more rapid spreading of variations through the stock. Relatively uni- form conditions and apparent scarcity of insects may give longer life to varieties in Egypt than in the United States, but the general tendencies and results of deterioration seem to be quite the same." The history of cotton culture in Egypt shows that a succession of new varieties has replaced the old at intervals of a few decades. The modern Egyptian cotton industry began with the variety dis- covered and popularized by Jumel, a French engineer, about 1820. The Jumel cotton was replaced by the Ashmuni after 1860, the Ash- muni by the Mit Afifi about 1890, and more recent varieties, such as the Jannovitch and Nubari, are now replacing the Mit Afifi. Other varieties, such as the Bamieh, Gallini, Zafiri, Abbasi, Sultani, etc., have either failed to gain any general popularity or have aroused only temporary interest. LINT AND SEED CHARACTERS OF HINDI COTTON. The character that renders the Hindi cotton so unwelcome as an element of admixture in the Egyptian stock is the much shorter and coarser fiber. The Hindi fiber is also pure white in color, whereas in the more popular Egyptian varieties the lint is a somewhat creamy white, tinged with buff or brown. White-linted varieties of Egyptian cotton have been cultivated to a small extent, but have never become popular in Egypt. The difference in the color of the lint is of much assistance in the work of sorting out the Hindi admixture after the fiber has been picked and brought to the ginning establishment. Any thorough separation of the inferior Hindi fiber from a white variety must be “Though very few insects were noticed in the Egyptian fields in June and July, they may be more abundant later in the season. Balls reports between 5 and 10 per cent of crossing, and even 25 per cent in one of his experiments. (See Balls, W. L., “* Cross-Fertilization in Cotton,’ Cairo Scientific Journal, vol. 2, 1908, p. 405.) 210 12 HINDI COTTON IN EGYPT. much more difficult, if not entirely impracticable. From this point of view it is easy to understand why the culture of Sea Island cotton or of the superior white varieties of Egyptian, such as Abbasi, has not become more extensive. The superiority claimed for the lint of the white varieties, such as Abbasi, is in accordance with other indications of a general correla- tion between the color and the length of the lint. Study of the lint characters of many variations and hybrids seems to indicate a general tendency in brown fibers to be shorter and coarser than white fibers. Thus the Jannovitch variety has lnt longer and whiter than the Mit Afifi, though still with a very slight tinge of brown. The Abbasi lint is still longer, but is pure white in color. If the need of sorting the fiber were removed by more effective methods of eliminating the Hindi variations, the way would be open to a larger use of white-linted varieties. Though brownish lint is preferred for a few purposes, the color seems to be valuable chiefly for the aid it gives in sorting out the inferior fiber that results from the Hindi contamination. If American growers are sufficiently care- ful to keep out the Hindi contamination, they may be able to grow white varieties that have longer and stronger fibers than the brown- linted varieties now popular in Egypt. In addition to the long fibers that compose the lint, the seeds of typical Egyptian plants are always provided with short fibers, or “ fuzz,” that continue to adhere to the seed after the lint has been removed by ginning. The fuzz may be confined to small tufts at the ends of the seed or may extend down one side, or may be more widely spread over the surface. The seeds of the typical Hindi cotton, on the other hand, are entirely without fuzz. The black surface is left entirely naked after the lint has been removed. The absence of fuzz makes the small, sharp-pointed, black stalk or funiculus at the base of the seed much more conspicuous in the Hindi cotton, though it is present in other varieties. The seeds of the Hindi cotton are more angular in shape than those of the Egyptian cotton. Though not adhering like the seeds of kid- ney cotton, they seem to be more closely crowded together in the boll than the seeds of the Egyptian cotton, and this mutual pressure tends “The production of Abbasi cotton is said to be irregular because the price fluctuates with Sea Island cotton. When Sea Island cotton is cheap there is small demand for the Abbasi. Another variety that gave very promising results in an experiment in Arizona in 1909, the Nubari, is said to be not very highly appreciated in Egypt because of a tendency to produce small bolls. While many smail-bolled plants were found in the Nubari field in Arizona, there was less diversity in this and other respects than in any other lot of plants grown from imported seed. 210 LINT AND SEED CHARACTERS OF HINDI COTTON. LS to make the Hindi seeds longer and more angular. Fully developed Egyptian seeds are usually plump, with all the sides distinctly con- vex and with a larger diameter than the Hindi seeds. The smooth surface and narrower shape of Hindi seeds make it possible to separate most of them by sifting, as the Egyptian ginning establishments are said to do. Nevertheless, it is not to be expected that any complete elimination of the Hindi cotton can be accom- plished in this way, for Hindi plants are occasionally found with fuzzy seeds much like the seeds of American Upland cotton. The seeds of Hindi hybrids are also somewhat fuzzy, often in the same way as the Egyptian seeds. Hand selection of seed intended for planting is said to be done in Egypt, though it does not seem to be a regular practice. Experi- ments carried on by Mr. Argyle McLachlan in Arizona indicate that Hindi variations and other aberrant tendencies can usually be de- tected if the seeds are studied with sufficient care and discrimination. The sorting out of the Hindi cotton is also assisted by the fact that the Hindi lint is very lightly attached, allowing the black surfaces of the seeds to be very readily seen. Even before the cotton is picked from the plants this difference is often very apparent. In addition to being short and coarse, the fibers of the Hindi cot- ton are relativeiy straight and have very little tendency to cling together, like the longer and more abundant. fibers of Egyptian and Upland varieties. After the Hindi bolls are open the seeds soon begin to separate and fall out, especially if they have a little assistance from wind or rain. In other words, the Hindi cotton is conspicu- ously lacking in storm-proof qualities. The naked surfaces of the Hindi seeds may be responsible for the fact that young plants of the Hindi cotton often appear to make more rapid growth than adjacent Egyptian plants. Experiments have shown that the germination of fuzzy-seeded varieties may be seriously delayed in dry weather, while seeds without fuzz may germinate promptly in the same soil. Obviously, too, a Hindi seedling that had germinated promptly and had sent out roots to absorb water would retard the germination of other seeds in the same hill. The cotton is planted in Egypt in relatively dry soil, the young plants being easily destroyed by any excess of water. Under such condi- tions there is usually a very unequal development of the young plants. Two or three plants in each hill, or perhaps only a single one, may develop several leaves and attain a height of 8 or 10 inches, while the other seedlings of the same hill remain with only the cotyledons expanded. 210 14 HINDI COTTON IN EGYPT. DISTINCTIVE CHARACTERS OF HINDI PLANTS. HABITS OF GROWTH OF HINDI COTTON. If the Hindi cotton could be recognized only by the characters of the lint and seeds, it might be impossible to effect a complete elimina- tion of the Hindi characters by selection. As long as Hindi plants are allowed to flower in the fields with the Egyptian plants and cross- fertilize them the undesirable Hindi characteristics may be expected to reappear. Even if no seeds of the Hindi form are planted, some of the apparently normal Egyptian seeds are likely to contain Hindi hybrid embryos, and these in turn can grow to maturity and produce pollen for continuing the Hindi infection to further generations. It is fortunate, therefore, that the Hindi cotton has several very definite differences in the vegetative parts, so that all Hindi plants can be recognized and rogued out of a field or a seed plat before the age of blooming and cross-fertilization is reached. The general form or habit of growth of the Hindi plants is different from that of the Egyptian cotton, though this is not so apparent in the Egyptian fields, where the plants are crowded closely together, as in experimental plantings, where more space is allowed the individual plants. The tendency of the Hindi cotton is to produce a broader and more bushy plant, more like the Upland than the Egyptian cotton. (See Pl. I, fig. 2.) | There is a general impression that the Hindi cotton is larger and more luxuriant than the Egyptian, but this may relate to the Hindi hybrids rather than to the genuine Hindi individuals. The Hindi plants may appear larger early in the season, perhaps as a result of more prompt germination, but they are usually outgrown by the neighboring Egyptian plants by the time the fruiting stage is reached. The Egyptian cotton, as well as the Hindi, shows different habits of growth under different conditions. In the cooler climate of Lower Egypt there is no such luxuriance of vegetative growth as in Arizona, but the branches are more spreading and the foliage more open. The habit of the Egyptian cotton in Egypt is more like that of Upland cotton in our Southern States. The similarity was especially strong in the Fayum Oasis, where some of the cotton is planted on rather poor land. It flowers and fruits when only 8 or 10 inches high, ma- turing small, bushy plants, ike Upland cotton on poor soil in the South. Something of the exuberant tendency was shown in an ex- perimental planting of Egyptian cotton at Siut (Assiut), in Upper Egypt. The habits of branching of the Hindi cotton are also different from those of the Egyptian. The fertile branches are less definitely spe- cialized than in the Egyptian cotton and have a stronger tendency 210 DISTINCTIVE CHARACTERS OF HINDI PLANTS. 15 to grow in upright or oblique positions and to assume the functions of vegetative branches, the flower buds being often aborted. LEAF CHARACTERS OF HINDI COTTON. The leaves of the Hindi cotton are characterized by thinner texture and lighter color, a fresh, bright green that forms quite a definite contrast with the duller grayish or bluish green of the Egyptian leaves. The surfaces of the leaves of the Egyptian cotton are some- what duller and more hairy in Egypt than in Arizona, though not so grayish as when the Egyptian cotton is grown in the cool climate of the Pacific coast, near Los Angeles. The color is usually darker before the fruiting stage of the Egyptian cotton is reached, when the foliage usually takes on a lighter and more yellowish tone. The dark foliage of the vegetative phase may be retained under condi- tions of abnormal luxuriance, or the change to the yellower shade of green may occur prematurely if the plants are affected by some unfavorable condition, such as too much water or too little. The veins of the leaves of the Hindi cotton are usually reddish, and the red color becomes very pronounced at the pulvinus or cushion-like thickening at the bases of the veins, where they pass into the petiole or stem of the leaf. The two large veins on each side of the midrib are particularly likely to be grown together at the base, giving the pulvinus of the Hindi cotton an oblong shape. The leaves of the Egyptian cotton do not have an enlarged pulvinus, the veins passing more directly into the petiole without becoming much swol- len or united at the base. The surface of the pulvinus of the Hindi cotton is naked, or with only a few scattering hairs, while the cor- responding part of the Egyptian cotton is usually quite hairy. The lack of specialization of the bases of the veins in the Egyptian cotton seems to render the leaves less capable of movement. They do not appear to change their positions to face the sun in the morning and afternoon as much as the leaves of the Hindi cotton. The turn- ing of the leaves to the sun renders the Hindi plants more conspicu- ous in the morning and afternoon than in the middle of the day, when the leaves have a horizontal position. Advantage was taken of this fact in making inspections of fields from moving trains, as will be explained later. Even in the first leaves or cotyledons of the young seedlings the reddening of the veins and the basal spot enables the Hindi cotton to be recognized and separated from the Egyptian. The difference of coloration is not so obvious in the first few leaves that appear after the cotyledons, for even in the Egyptian cotton these are likely “YP imorphic Branches in Tropical Crop Plants, Bulletin 198, Bureau of Plant Industry, U. S. Dept. of Agriculture, 1910. 77267° —Bul. 210—11 2 16 HINDI COTTON IN EGYPT. to have a somewhat reddish spot at the base, especially if the condi- tions are not favorable for rapid growth. ‘The differences become more obvious as the plants grow, until the flowering stage is reached, but they may lessen or disappear at maturity. In adult Egyptian plants the veins of the leaves often become reddish, while those of adult Hindi plants may become pale.“ After the color contrasts have disappeared, the recognition of the Hindi plants requires notice of other less obvious details of the leaves, flowers, and bolls. Thus the leaves of the Hindi cotton have the lobes broader, more abruptly narrowed toward the apex, and usually produced into longer terminal points. In Hindi hybrids there are often 5 to 7 lobes which are often somewhat folded or plieate, as in the Egyptian cotton, the true Hindi plants having the leaves nearly flat. The rounded basal lobes of the leaf are broader in the Hindi cotton, so that the leaf as a whole is more nearly square or oblong in shape. The corresponding margins of the Egyptian leaves are likely to converge or slope backward toward the stem.? The sinus or notch at the base of the leaf, where the petiole is Inserted, is usually much broader in the Hindi cotton, exposing the upper surface of the end of the petiole. In the Egyptian leaves the sinus 1s generally very narrow or completely closed by the contact or overlapping of the margins of the lobes. The wider separation of the lobes of the Hindi cotton may be considered as a consequence of the thickening of the veins and the enlargement of the end of the petiole FLORAL CHARACTERS OF HINDI COTTON. The involucre that incloses the bud of the cotton plant is composed of three bracts, small leaf-like organs, each margined with a fringe of narrow teeth. The bracts of the Hindi cotton are more broadly rounded at the base and have longer and more numerous teeth than those of the Egyptian cotton. Comparison of the Hindi bracts shown in Plate III with the Egyptian bracts at the top of Plate IV will enable these differences to be understood. Another diagnostic feature of the Hindi bracts is that the teeth run down nearer to the base, a tendency that is shared by the Hindi hybrids. Three hybrid bracts are shown at the bottom of Plate TV. The bracts of the Egyptian cotton seemed to be somewhat more cordate in Egypt than in the United States, but the narrowly triangular form, straight sides, and small teeth, remote from the base, generally render them *Mutative Reversions in Cotton, Circular No. 53, Bureau of Plant Industry. U. S. Dept. of Agriculture, March, 1910, pp. 10-11. ®’¥For natural-size illustrations of leaves of Egyptian and Hindi cotton, see Circular No. 42, Bureau of Plant Industry, December, 1909, pp. 4 and 5. 210 DISTINCTIVE CHARACTERS OF HINDI PLANTS. IS? quite different from the Hindi bracts, in spite of endless variations in the minor details. The calyx of the Hindi cotton has five distinctly prominent triangular lobes, one or two of which are often produced into a narrow needle-like point. In the Egyptian cotton the lobes of the calyx are very short and broadly rounded, never produced into long points. Three examples of the toothed calyx of the Hindi cotton are shown in Plate IIT; an Egyptian calyx and the calyx of a hybrid in Plate IV. The fresh, newly opened flowers of the Hindi cotton have pale creamy-white petals lke those of Upland cotton instead of lemon- yellow petals ike Egyptian cotton. In the afternoon the flowers of both sorts change to a reddish pink, but the Hindi flowers attain a much deeper shade than the Egyptian. The petals of the Hindi cotton are shorter than those of the Egyptian and open more widely. The Hindi flower may be de- scribed as cup-shaped, the Egyptian as tubular. The purple spot found at the base of each petal in Egyptian flowers is lacking or only faintly indicated in typical Hindi flowers, though often quite pronounced in Hindi hybrids. The pollen of the Hindi cotton is of a much paler yellow and the individual pollen grains are much smaller than those of the Egyptian cotton. FRUIT CHARACTERS OF HINDI COTTON. The bolls of the Hindi cotton have a rounded conic shape, broadest near the base, and taper abruptly to a short point. Egyptian bolls are more fusiform, narrower at the base than near the middle, and taper less abruptly to a rather blunt apex. The shape differs appre- ciably with the conditions, the less luxuriant plants in Egypt having a broader and more conic form than is usual in Arizona, more like the bolls produced by the Egyptian cotton in the vicinity of Los Angeles. (See Pl. VI.) The surface of the Hindi bolls has a rather dull pale pea-green color, with only slight indications of the deeply buried oil glands. Egyptian bolls, on the contrary, have a fresher, darker color, with the surface smooth and shining, but rather deeply pitted around the numerous superficial oil glands, each of which appears as a distinct black dot. These differences appear somewhat less pronounced in Egypt than in Arizona. Pale-green bolls were found on many plants that seemed in all other respects to represent true Egyptian cotton. The darker color of the bolls in Arizona may be connected with the greater luxuriance of the plants. The number of carpels, or “ locks,” varies in the Hindi cotton from 3 to 5, the majority of bolls having 4 locks. In the Egyptian cotton 210 18 HINDI COTTON IN EGYPT. the locks range from 2 to 4, with 3 as the prevailing number. Very few 4-locked bolls could be found in the Egyptian fields, but they are somewhat more numerous on the larger and more luxuriant plants grown in Arizona. PREVALENCE OF HINDI COTTON IN EGYPT. Familiarity with the vegetative characters of the Hindi cotton made it possible to secure definite information regarding the prevalence of this type of cotton in Egypt and thus obtain a basis of judgment regarding the value of the methods of selection that are being ap- plied to the Egyptian cotton in Arizona. In attempting to judge of the practicability of establishing the culture of Egyptian cotton in the Southwest, it is obviously important to understand how far ithe commercial reputation of the Egyptian cotton for uniformity depends on the special methods of sorting and preparing the cotton for market. This will enable us to appreciate the advantage that may be gained by growing a more uniform fiber in the fields and avoiding the necessity of the subsequent labor in sorting and blend- ing the fiber into a uniform product after it comes to the ginhouse. Some writers have given the impression that the native cultivators rogue out all the Hindi plants during the process of thinning the young cotton early in the spring and thus avoid an admixture of the Hindi fiber. Others have referred to the Hindi cotton as a wild plant in Egypt, or even a common weed, making it seem almost impossible to avoid contamination. Neither of these impressions seems to correspond with the facts. Though many of the native cultivators will hasten to assure the inquirer that they pull out all of the Hindi plants, a goodly rem- nant of typical Hindi individuals is to be found in nearly every field. On the other hand, one does not find the Hindi cotton, any more than the Egyptian cotton, outside of regularly planted cotton fields. Seeds scattered near permanent watercourses or about towns may sometimes grow to maturity, but it is not easy to understand how the idea of wild cotton growing at large in Egypt could have gained currency. Other plants that casual observers might mistake for cotton, such as the okra or bamieh (Hibiscus esculentus), the Deccan hemp (Hibiscus cannabinus), or even the cocklebur (Xan- thium), are all strictly dependent upon cultivation and irrigation. It is difficult to believe that a plant of the habits of the cotton could exist as a native or truly wild species in the Nile Valley. And if sucht a species did exist naturally it would be dependent upon the annual flood for its water, and would be a winter-growing species. The commercial culture of cotton was not developed in Egypt under the historical system of basin irrigation direct from the annual flood 210 PREVALENCE OF HINDI COTTON IN EGYPT. 19 of the Nile. The period of high water comes during the late summer and autumn, the fruiting season of the cotton. Egypt did not gain importance as a cotton-producing country until the modern system of perennial irrigation from stored water was developed, in the nineteenth century. The Egyptian system of close planting greatly increases the diffi- culties of finding the Hindi individuals and of counting the Hindi and Egyptian plants to determine the percentages of each. Early in the season, while the plants are still small, each one can readily be seen as a separate individual, but with larger growth they fuse together, as it were, to form a solid mass of fohlage. Early inspection has the further advantage of utilizing the differences in the color of the foliage that are readily appreciable in the vegetative phase of development, but tend to disappear after the fruiting stage has been reached, as already explained. If actual countings are not made, the proportion of Hindi cotton is likely to be seriously underestimated after the plants have reached the adult or flowering stage. It has been said that the Hindi plants can be distinguished from the Egyptian by their taller growth, but this seems to be true of hybrids or of young individuals rather than of mature plants of the true Hindi type. It was noticed at Calioub and at several other points that while many of the hybrid plants ran several inches above their Egyptian neighbors, the true Hindi plants had usually been outgrown by the Egyptian. In fact, some of the Egyptian cultivators consider that the hybrids rather than the true Hindi plants ought to be pulled out. They have noticed that many of the large overgrown hybrids produce very little fruit and are willing to pull them out so they shall not crowd their more productive neighbors. Careful roguing in the early part of the season is more likely to take out all of the true Hindi plants and leave a few of the hybrids, so that careful cultivators are more likely to be familiar with mature hybrids than with mature Hindi individuals. The true Hindi plants, being less obtrusive when the stage of maturity is reached, are very easily overlooked unless special care is taken to separate and count the plants of each hill. Though two plants are usually left at thinning, regularity in this respect can not be depended upon. It often happens that only one plant sur- vives, or careless cultivators may leave occasional hills with three or four plants. It may be that the value of countings as the basis of general esti- mates of the proportion of Hindi cotton would not be seriously impaired by assuming two plants to each hill. The saving of time in this way would enable more extensive counts to be made. This plan was followed in a few of the later countings mentioned below. 210 20 HINDI COTTON IN BGYPT. at Calioub and Siut, in fields where the plants had grown very large. The hills were each noticed in turn to see whether they con- tained Hindi plants. Hills with no Hindi were assumed to have two Egyptian plants. The general effect of this plan would be to reduce somewhat the apparent proportion of Hindi plants, since it is prob- able that in most of the fields there would be more hills with a single plant than with three or four plants. Nevertheless, it might be that the figures obtained in this way would be more reliable, in view of the larger areas that might be inspected in a limited time. To serve as a general basis of judgment regarding the prevalence of the Hindi cotton in Egypt, countings of individual plants were made in several different localities. In most localities several sepa- rate counts were made, usually in fields of different proprietors, or at least of different tenant cultivators. The figures obtained do not represent the full extent of Hindi contamination of the stock, for in most cases a more or less careful roguing out of the Hindi plants had already taken place. The psychological factor of the individual cultivator enters, therefore, as an important element in the calcula- tions. One field might have only a few Hindi plants, while the next would have a considerable percentage. Thus of two adjacent fields at Tanta one showed less than 3 per cent of Hindi, the other 15 per cent. Questioning of the native cultivators showed wide differences of individual opinion. Some of them were quite alive to the need of pulling out all of the Hindi cotton and showed annoyance or offered excuses 1f reminded that many Hindi plants were still to be found in their fields. Others took a more languid interest in the matter. One cultivator might claim to have pulled out large numbers of Hindi already, while his neighbor might not think it necessary to admit any responsibility for pulling out the Hindi at all. He would not deny, perhaps, that he had heard of the need of pulling out bad cotton plants, but would insist that very few people did it. The popular impression in Egypt among people who consider themselves informed about cotton growing is that selection receives proper attention in the Delta region, where the Mit Afifi and Janno- vitch, the principal varieties of Egyptian cotton, are grown, but is very much neglected in Upper Egypt, where the Ashmuni and other inferior stocks are produced. It seems, however, that this impres- sion may relate to more careful sorting done in the ginning establish- ments of the Delta rather than to any really efficient selection in the field. Even about Tanta and Mansurah, the recognized centers of production of high-grade fiber, a conspicuous representation of the Hindi cotton was seen in a large proportion of the fields. The percentages of Hindi plants counted in fields at Tanta, in Lower Egypt, are about the same as those obtained at Beni-Suef. in 210 PREVALENCE OF HINDI COTTON IN EGYPT. Pall Upper Egypt. (See Table I.) The idea of Hindi cotton seemed to be more common about Tanta, but no indication of a serious effort to eradicate the Hindi type from the fields could be gathered from native cultivators. They are willing to pull out the Hindi plants rather than the Egyptian at the time that the hills are thinned down to the usual two plants, but have no idea of destroying any more plants after the thinning has been done. One very zealous native showed interest to the extent of pulling up some of the Hindi plants that were pointed out to him, where there was an Egyptian plant in the same hill. But when there were two Hindi plants together in a hill he would pull up only one. Nor could he be induced to sacrifice any of the Hindi individuals that stood by themselves, although he believed (as was afterward learned) that a Government inspection was being made. The Egyptian Government sends entomological inspectors through the fields to guard against outbreaks of the Egyptian bollworm. Beni-Suef is considered the chief center of cultivation of the Ashmuni cotton, this variety being now confined largely to Upper Kegypt. Inspection of fields in this locality on June 6, 1910, showed a general prevalence of Hindi and great lack of uniformity in other respects, though not as great nor as obvious as in experiments with this variety in Arizona. There is the same tendency to red spots at the base of the leaves, which is recognized as a mark of this variety to distinguish it from Mit Afifi, Jannovitch, and other more care- fully selected varieties. The more general tendency to the red spot may be a result of a more general contamination with the Hindi type of cotton. A special count was made at Beni-Suef to learn the extent of Hindi contamination as indicated by the presence of the distinct red spot at the base of the leaf. This included true Hindi plants, _obvious hybrids, and all other plants that would have been considered as having too red a callus for varieties of Egyptian cotton other than Ashmuni. Of 213 plants examined for the color of the callus 133 had the callus green or only slightly tinged with red, as usual in Egyptian cotton, while 80 plants were noted as having the callus distinctly red, as in the Hindi cotton. In the oasis of Fayum still less attention seems to be paid to the Hindi cotton than about Beni-Suef. Native cultivators knew that some of the plants produced inferior cotton, but did not claim to be able to distinguish them except by the white flowers. There was evi- dently no intention of pulling out any of the white-flowered plants. The variety planted at Fayum was not considered to be Ashmuni, but was merely called Beladi, or “ native,” cotton. Other countings of Hindi were made in the Beladi cotton at Siut. Cotton is not regularly planted about Siut, but experiments are 210 oF HINDI COTTON IN EGYPT. being made with seed brought from Fayum. The percentage of Hindi is much larger than appeared at Fayum, though the planter claimed that he had taken out numerous Hindi plants when the field was thinned. In addition to the plants counted as Hindi, much diversity was apparent, almost as much as in a field of Ashmuni cotton grown in 1909 at Somerton, Ariz. Such cases suggest the pos- sibility that transfer to new conditions may have the effect of in- ducing additional variations in these diverse stocks, but the pro- portion of Hindi in either parent stock could not be ascertained. Whatever the cause of the phenomenon, it is a significant fact that the proportion of Hindi plants and obvious hybrids may run as high as 20 per cent. The census of Jannovitch cotton at Tanta was somewhat more rigorous than that at Beni-Suef and included some plants with dis- tinctly red leaf bases; plants with distinctly red leaves and other obviously aberrant tendencies that might have been omitted in the Ashmuni fields, where the red callus is so common a feature. But many other definitely aberrant plants with light-green leaves were not included when they lacked the red callus. These hght-colored plants have the more ample and luxuriant foliage of the Hindi hybrids and may represent a second-generation splitting of the Hindi characters. Such a splitting might be expected with a color charac- ter like the basal spot that also shows seasonal reversibility. The smallest proportions of true Hindi plants were found in fields in the vicinity of the barrage (a few miles below Cairo) and at Calioub, in the same district. None of the fields that were inspected in these places showed any large percentages. About two-thirds of the plants counted as Hindi were plants of the type considered as first-generation hybrids. In one field at the barrage and in another at Calioub no true Hindi plants could be found, even after a rather careful search, though several obvious hybrids were present in each field. At Benha, on the contrary, the Hindi percentages not only ran higher but a larger proportion of the plants represented the true Hindi type. In the neighborhood where the counts were made near Mansurah the native cultivators placed much importance on the elimination of the Hindi plants, though they were known by a different name, “ Haga,” the word Hindi not being recognized. It was estimated that about 5 to 6 Hindi plants had been removed from each row of 100 to 150 plants at the time of thinning, in addition to those that remained to be counted. This would indicate a total Hindi represen- tation of between 5 to 10 per cent in this stock of seed at the time of planting. In several instances it was noticed that the Hindi plants seemed to be more numerous on the higher, drier ridges or dikes that bounded 210 PREVALENCE OF HINDI COTTON IN EGYPT. 93 the different sections into which the fields were divided for irriga- tion purposes. Separate counts were made of plants along some of the dikes, but without securing any definite evidence. It would be interesting to know whether such differences of conditions would have an influence over the expression of the Hindi characters. Other explanations were possible—that the higher ridges had been neglected at the time of thinning the plants or that the Hindi plants had an advantage in germinating in the drier soil of the higher ridges, because of the smooth seeds. The cotton often appears to be more luxuriant on the higher dikes than in other parts of the fields. Indeed, such dikes are usually planted with double rows of cotton, as though to take full advantage of the more favorable conditions. TABLE I.—Countings of Hindi cotton plants. Per- Per- ae EgZyP-| gsnaq;| 6 Plants] Egyp-| y:4;| Location. count-)| tan oe | eee Location. count-| tian coe baa ed. | type. | °° | Hindi. eat )| types! =" ~ laindk: 445 435 10 2.24 || Fayum, Upper ed 871 819 52 5.99 274 242 32 11. 67 (Beladi variety)... 676 629 47 6.95 512 457 55 10. 74 | - Beni-Suef, Upper 165 155 10 6.06 Moy R2y le pee meee 1,547 | 1,448 99 6. 41 Egypt (Ashmuni 178 161 17 9.55 | = WATIG UA) yota= etetsieese 446 435 11 2.48 609 | 494 115 18. 88 327 294 33 10.09 467 398 69 14.77 245 224 21 8.56 || Siut, Upper Egypt... 316 260 56 | 17.72 130 124 6 4.61 444 354 90 | 20.27 467 383 84 17.98 Motels 2.25.2 2,722 | 2,527 195 7.16 - } Motaleee ssc 2,303 | 1 889 414 17.97 595 569 26 4.36 = = 886 829 57 6. 43 844 829 15 Mees 464 441 23 4.96 476 472 4 . 84 464 437 27 5.82 560 555 5 | -89 Tanta, Lower Egypt |} 368| 340/ 28] 7.61 ery ower lr sai |, 528 8| 1.50 (Jannovitch variety)|)1,028 | 923} 105 | 10.21 Bet "Mit Afi va- |) 720| 758 32| 4.05 806 738 68 8. 44 rieties) 669 662 Ti 10s 134 118 16 TAO 20 ll ieererte saa’ atone seat 528 517 11 2.08 566 550 16 2.83 598 584 14 2.34 566 476 90 | 15.09 934 924 10 1.07 Motels saeco te 5,877 | 5,421 456 Tare otal. saaoeeee 5,930 | 5,824 106 1.78 1,149°| 1,124 25 2.17 || 857 810 47 5.49 424 410 14 3.31 202 200 2 .99 \} 551 543 8 1.45 || Benha, Lower Egypt 461 429 32 6.93 Barrage, near Cairo. - 483 474 9 1.86 655 633 22 3.36 567 559 8 1.41 558 542 16 2.86 334 328 6 1.79 | _ 511 486, 25 4.89 Total. = eos 2, (33) |) 2,014 119 4.36 wOtalica mses == 4,019 | 3,924 95 2. 36 1,043 954 89 8 = F 889 816 73 8. 21 417 | 412 Eaipeminia |p octeb Ealestine- ht ozeelitge7 | a) On | lad Calioub, near Cairo. . 497 493 4 . 80 982 946 36 3. 66 1,216 | 1, 202 14 1.15 Potala sete ss 4,889 | 4,663 226 4.62 2,130 | 2,107 | 23 | 1.07 | Count was made of 32,150 plants in all, of which 1,733 were re- corded as belonging to the Hindi type, a percentage of 5.39. If the percentages for the different localities are averaged, a somewhat higher general average, 5.98 per cent, is obtained. One series of countings of Hindi plants was made in an experiment with Egyptian cotton in Palestine, at a locality called Beteha, near 210 94 HINDI COTTON IN EGYPT. the north end of the Lake of Tiberias, not far froin the ancient Capernaum. The first two counts at Beteha were made in late- planted fields that had not yet been thinned or rogued for Hindi. The percentages obtained in these cases, 8.53 and 8.21, may be taken to represent the amount of Hindi contamination represented in the seed before planting. Early-planted fields at Beteha seemed to be as far advanced as any seen in Egypt, the date of the visit being June 23. In order to obtain a more general and yet a-not altogether in- definite indication of the prevalence of the Hindi cotton, the apparent presence or absence of Hindi cotton was noted for a considerable number of fields that could be seen to advantage from the railroad. Such inspection is greatly facilitated by a fact already considered, namely, that the leaves of the Hindi cotton have greater freedom of motion than those of the Egyptian cotton, and that they make pro- nounced changes of position in order to face the sun in the morning and afternoon. The Hindi plants are much more readily seen from a distance at these times than in the middle of the day, when the leaves are in a horizontal position to face the sun overhead. The presence of tall hybrids gives a general impression of uneven surfaces to the fields and thus betrays the presence of Hindi cotton, even when details of individual plants can not be made out. But when the broader, fresh-green leaves of the Hindi plants are formed into rosettes to face the sun, they become conspicuous and unmis- takable. Indeed, it is sometimes more difficult to distinguish them from the okra that is often planted in the fields than from the Egyptian cotton. The Egyptian okra (bamieh) has broad leaves of the same color as those of the Hindi cotton and also a red spot at the junction with the stem. Such observations are greatly assisted by the fact that the Egyp- tian railroads are usually elevated on embankments. By being able to look down on the fields a more accurate impression can be gained than by viewing the plants from the side, as one is obliged to do when standing on the same level. It is to be expected of course that Hindi plants would be found by more careful inspection in most of the fields where they were not ap- parent from a passing train. But at least it may be considered that fields showing no apparent Hindi have been rogued. In a large pro- portion of cases the Hindi plants and hybrids were very conspicuous. Fields that have had the Hindi plants and hybrids rogued out often appear remarkably even in height and color. Such an inspection could not be made to any advantage after the Egyptian cotton has entered the fruiting phase, when the color changes from a dark to a lighter green, thus destroying the contrast with the Hindi cotton, so marked during the earlier vegetative phase. 210 PREVALENCE OF HINDI COTTON IN EGYPT. 25 In addition to the lighter color assumed by the foliage of the Egyp- tian plants as the season advances, the proportion of yellow in the fields is increased by the abundance of bracts and flowers. At the time these changes were taking place, about the middle of July, the dark-green tone of the vegetative phase was still shown with much uniformity in some of the fields, while others had gone over to the yellower shades or were still more completely dominated by the abundance of yellowish bracts and still yellower flowers. These changes seemed to have come rather suddenly, for most of the fields seemed to represent one phase or the other quite definitely, only a few showing pronounced individual diversities of coloring among the Kgyptian plants. TABLE II.—Fields with Hindi cotton apparent from trains Fields Fields : eee Number with without Fields were noted between towns— of fields. | apparent | apparent Hindi. Hindi. Calionb andi@hebinvell Kanetenssrers cco scmecctecece sons sneccecccatesecse 48 46 2 GhebimgancdiMia che tou lssseee cee ne oes ciara cise a clon cele Soo oer aisim Seclemecle’e 53 50 3 Macnetomlean Gptsil heissertee= nas seek Ue ese mc es ee ee ee ae 81 76 5 Bi beisrant enon 7 Saapeee eas meme open Ia eee aN aes Se cee Sagi ook 82 81 1 LEN EVAL £29 6YO | YAN) OY) BE C{) 0 b Daa I Ieee ats Ae ae 88 82 6 FAH OUS Kein and wk Saki Men ee. foe coisas cue natseraeisceeee sate ci mien & 24 22 2 KatiSaknandeApomelihekouksen: 222.2 25.25 sseeesacee sees cece eas ee wen 30 24 6 AbomeliChekoukiandSimibellaoueines 2-2. see .-2 ese ec ace eee eee ences 51 | @(10) 44 7 Simbellaouein and Baklieh..........-- aoe 13 (3) 18 0 BakliehtandsMansurahe. eo 2-22.52 2aces- <5 -| 24 (4) 17 7 ManIStina hear Gas aim anOUG saa eeryae ace aaa ake oe ee hes oe Se ee ceases | 90 (16) 73 17 pamanoudandeMehalla WKebinjr. jas. c ose ace tense es ad cwasock wie aelec wie coe 29 26 3 Mehalla Kebir and Mehallet Roh....- re eee er eee rh See eee 34 (4) 27 7 Mehalletavohtandien snitass:- seme. okey be ae oe 3 tiny coed waa | 49 (3) 42 7 TPROLI eas oe Se, rate oe nn ae re te ee 696 623 iB ET EE TIA Corey a Seen ears eee anism a ee Ses as nen eee eral basesiemees 89. 51 10. 48 a In some localities fields that showed a strikingly large proportion of Hindi cotton were specially noted, and the numbers of such fields are given in parentheses in the table. It would be safe to estimate that the proportion of Hindi cotton and obvious hybrids in such fields was more than 5 per cent. Many fields between Bilbeis and Zagazig appeared to be quite as thickly sprinkled with Hindi as any in Upper Egypt where percentages of 15 and 20 were counted. Tn addition to fields noted in Table IT, many other inspections were made in the region between Cairo and Tanta. Several hundred fields were seen in Upper Egypt, in every one of which indications of Hindi contamination were found. In the district between Abou el Chekouk and Mansurah much of the cotton at the middle of July was still too small and irregular to give faverable conditions for seeing Hindi plants from the train. . Many of the fields had not begun to flower. In many the stand was irregular, or the plants of irregular sizes, perhaps as a result of alkali in the soil. Fields of rice interspersed among the cotton showed the same irregularity. The unfavorable conditions may be partly responsible for the larger proportion of fields with no apparent Hindi in this district. Fields with larger plants often showed great 210 °6 HINDI COTTON IN EGYPT. abundance of Hindi. Most of the cotton to the west of Mansurah was in better condition and afforded a more reliable indication of the prevalence of Hindi, or rather the prevalence of roguing. Though the proportion of fields apparently clean of Hindi seemed to be dis- tinctly larger than in other districts, many of the fields showed un- mistakable Hindi plants in great abundance. Unless the conditions are favorable for the detection of the Hindi plants such inspections could have very little value, but if made at the right time the presence of the Hindi contamination and the relative amount in different districts could be judged very easily in all localities accessible by railroad. The time would differ with the growth of the cotton in the different localities, probably extending through the month of July. Before June 20 the Hindi plants could seldom be seen from the trains, but during the second and third weeks of July they were easy to see in all except the more backward districts. CHARACTERS OF HINDI HYBRIDS. DISTINCTIVE FEATURES OF HYBRIDS. Except in cases that are especially noted, the plants enumerated as Hindi in the preceding tables comprise two elements, the typical Hindi plants and the pronounced Hindi hybrids, those that resemble the first generation of the crosses that have been made between the Tiindi and the Egyptian cottons. When the fields are in the earher vegetative phase, the pronounced hybrids can be distinguished from the Egyptian plants by the hght color of the leaves and the red pulvinus at the base of the veins, al- most as easily as the true Hindi. The larger size of the hybrids also attracts attention. The leaves of the hybrids become larger than those of the true Hindi plants, and most of the larger leaves have five or seven distinct lobes instead of three. The lobes of the hybrids are somewhat folded or channeled, like those of the Egyptian cotton, instead of spreading out nearly flat, as in the Hindi cotton. The larger size of the involucral bracts of hybrids is another feature usually quite obvious. (Pl. IV, B.) The teeth do not always run down toward the base of the bracts, as in the Hindi cotton, though there is a general tendency in this direction. In Arizona the Hindi hybrids have shown a marked tendency to sterility or to very late bearing, but in Egypt, early in June, some of the hybrids seemed to * be more advanced toward flowering than their Egyptian neighbors. The countings of the Hindi plants and obvious hybrids do not by any means indicate the full extent of the Hindi contamination in the Egyptian fields. There is background of diversity too multi- farious to be counted or even noted in detail without careful inspection 210 CHARACTERS OF HINDI HYBRIDS. OT of the charac¢ters of individual plants. Crossing between hybrid plants and Egyptian must produce many very dilute hybrids with little or no expression of the Hindi characters. Indeed, it may well be doubted whether any of the Egyptian stock would be found to be entirely free from the Hindi contamination if all of the ancestry could be traced. As yet we have no knowledge of the effects of slight dilutions of the Hindi blood upon the expression of characters, but experiments are being made to obtain information on this point. Two principal elements might be recognized in the study of the diversity that exists in the Egyptian fields. One element might be ascribed to the prevalence of the Hindi cotton, the other to variation inside the Egyptian type. But in the present state of our knowledge it is often quite impossible to determine at once whether a variant plant is a dilute Hindi hybrid or an unusual example of the Egyptian stock. Evidence on this question can be secured by planting the seed to see whether the progeny “come true” to the characters of the parent, as in a mutation, or show more pronounced reversions to the Hindi type. But many mutative variations are also to be considered as reversions. The practical fact is that the Hindi con- tamination is responsible for a large amount of diversity outside of the obvious hybrid forms that resemble first-generation crosses. Among the plants enumerated as Egyptian are many that are appreciably different from the Egyptian type, even in the early part of the season. Without departing seriously from the Egyptian form and habits of growth, some of the plants have broader or narrower leaves, lighter or darker than their neighbors. Though the form of the leaves may be that of the Egyptian cotton, the bases of the veins may be reddened as in the Hindi. Or plants with Egyptian foliage may have unusual habits of growth, the more frequent tendency being toward taller stalks and more strictly upright branches. The large cordate bracts that characterize the most obvious Hindi hybrids are not entirely confined to that class of plants, but may be found on other large plants with foliage of Egyptian shape and color. The pulvinus may have the Hindi size, shape, and color, though concealed by more abundant hairs. In addition to the large circular, or very deeply cordate bracts, with the teeth running well down, such plants often have the calyx distinctly toothed, though the teeth do not have the long slender points that occur so frequently in the Hindi cotton. (See Pls. III and IV.) As the season advances such differences become more apparent. When flowering and fruiting begin the hybrid nature of many indi- viduals becomes unmistakable. even in plants that might not have been suspected of hybridity from the vegetative characters alone. Roguing must not be limited to the time of thinning in the early 210 298 HINDI COTTON IN EGYPT. spring if any complete elimination of the Hindi characters is ex- pected.” The tendency to revert to small bolls is one of the most frequent and least obvious evidences of Hindi contamination. Small bolls can often be found on large-bolled plants, but many individuals pro- duce only small bolls. The shape of the bolls may not suggest Hindi, though other Hindi characters may be found, such as naked seeds, sparse white lint, or pale spots in the flowers. To make a complete enumeration of all the plants that show any of the Hindi characters it would be necessary to watch a field of cotton through the whole season, for in some plants only the lint and the seeds may betray the Hindi ancestry. Already, at the beginning of the fruiting season in Egypt, it became evident that many of the aberrant Egyptian plants were really Hindi hybrids, in addition to the type of hybrids that had been included in the countings. Even in the fields that had been quite carefully rogued, as at Man- surah, so that only very small percentages of plants with the Hindi foliage were left, many white-flowered individuals remained. The leaves of the white-flowered plants seemed to be a little broader than those of adjacent yellow-flowered Egyptian plants, but the difference was not enough to be noticed if attention had not been attracted by the flowers. COHERENCE OF CHARACTERS IN HYBRIDS. It is not yet certain that all of the more Hindi-like hybrid plants are really first-generation hybrids, the direct result of cross-fertiliza- tion between Hindi and Egyptian plants. All that is known at pres- ent is that the crossing of Egyptian with Hindi does produce plants of the Hindi-like hybrid type. The experiment has been made in Egypt by Mr. Balls and in Arizona by Messrs. McLachlan and Meade. It is possible, however, that some of the Hindi-like hybrid forms may represent the progeny of hybrid parents. According to the Mendelian theory of heredity a part of each generation of hy- brids should resemble the first generation, while the remainder should show other combinations of the parental characters. In typi- cal Mendelian hybrids the contrasted parental characters are sup- posed to have entire freedom of chance combination in the second and later generations. In reality there does not seem to be such complete freedom of com- bination of the two sets of characters that represent the two parental types. Plants that have the Hindi foliage, or that of the Hindi-like hybrid type, invariably have the white petals of the Hindi cotton. “Cotton Selection on the Farm by the Characters of the Stalks, Leaves, and Bolls, Circular No. 66, Bureau of Plant Industry, U. S. Dept. of Agriculture, 1910. 210 CHARACTERS OF HINDI HYBRIDS. 29 White flowers always have the more open, cuplike form of the Hindi cotton instead of the longer and more tubular form of the Egyptian cotton. It very rarely if ever happens that any single Hindi character is brought into definite expression by itself—that 1s, without being accompanied by the more or less definite expression of other Hindi characters. It is hardly to be supposed that any of the Hindi plants, any more than the Egyptian, are pure bred in the sense of having had no Egyptian ancestors, and yet the Hindi type is nearly as uniform as the Egyptian, in spite of all the selection that has been directed against it. Neither is it reasonable to assume that all of the pronounced hybrid plants have the same proportions of Hindi and Egyptian blood, though they form nearly as definite a group as the parent types. Hindi-like lint and seeds sometimes occur on plants that give little or no external evidence of Hindi contamination, but plants that have previously shown Hindi leaves or flowers very seldom, if ever, have typical Egyptian bolls or lint of good Egyptian quality. In a field of Jannovitch cotton raised in Arizona in 1909 from imported Egyptian seed numerous individuals were found that seemed, early in the season, to depart from the normal Egyptian type only in the lighter and more pinkish tinge of the purple spot at the base of the petals. But when these plants were examined again in the fall it was found that the bolls and lint also departed from the type of the variety. All the pale-spotted individuals had small bolls, and some of them showed naked seeds and short Hindi-hke lint. That the depth of color of the petal spot can be, in itself, a matter of any direct significance in the economy of the plant is hardly to be believed, but it seems to have an indirect significance as indicating a tendency for the Hindi or other abnormal characters to come into expression. White petals may be considered in the same way as ev1- dence of a still stronger tendency to express the Hindi characters in the parts to be subsequently formed. Very pale yellow flowers were noticed on a few Egyptian-hke plants at Mansurah, but in nearly all cases a departure from the normal Egyptian color involved a complete change to the creamy white of the Hindi flowers. Although white Hindi-like flowers are rarely to be found on plants that have produced Egyptian foliage, such sudden changes in the expression of the characters do not appear to be normal phenomena of heredity, at least in cotton hybrids, for plants with these incon- eruous combinations of characters are generally infertile and some- times completely sterile. *Mutative Reversions in Cotton, Circular No. 58, Bureau of Plant Industry, U. S. Dept. of Agriculture, 1910, p. 6. 210 30 ' HINDI COTTON IN EGYPT. Coherence of characters is not confined to Hindi hybrids, but apparently has to be reckoned with in any attempt to combine the characters of different types of cotton... The phenomenon was first recognized and described in the study of Egyptian-Upland hybrids in Texas and Arizona. It differs from correlation in affecting whole groups of characters instead of only two or three. Thus a general correlation may be said to run through many different types of cot- ton—between the shape of the boll and the length of the lint or between the color of the lint and its strength. Correlation refers pri- marily to the fact that certain characters tend to vary together, one increasing or diminishing in relation with another. The fact that the weight of ears of corn increases with their length is reckoned as a correlation. Coherence refers to the expression of characters in hybrids. It denotes a condition in which characters derived from the same parent remain together in expression instead of being expressed in chance combinations as in Mendelian hybrids. Correlations often appear entirely arbitrary, unless they are merely mathematical expressions, as in the case of the corn ears. From the mathematical standpoint it seems impossible to understand why long fibers should not be packed into round bolls as well as into pointed bolls or why brown fibers should not grow as long as white fibers. But after the tendency to coherence of much larger groups of char- acters has been recognized as a fact correlations appear somewhat less mysterious. The general association of longer lint with more pointed bolls in any particular type of cotton may be connected with the other general fact that the long-linted types of cotton have more gradually tapering bolls than short-linted types of cotton. Coher- ence implies that the expression or nonexpression of one character may determine whether other characters shall be patent or latent. A striking example of coherence of characters was observed in Egypt in a block of hybrids made by Mr. F. Fletcher, director of the School of Agriculture at Gizeh, between an American Upland variety called Jackson’s Limbless and an Egyptian variety called Voltos, somewhat similar to Nubari, Voltos being the male parent. In addition to many other courtesies of hospitality Mr. Fletcher most generously insisted upon a full use of his interesting series of experimental plantings of cotton at Gizeh, which yielded many in- teresting facts with special relation to problems of diversity. Instead of the usual tendency of some of the Egyptian traits to predominate in the first generation, this lot of hybrids showed an unusually definite expression of the Upland characters. Very few of the plants would have been taken for Egyptian cotton, even on casual examination, and none of them showed any close approxima- tion to the Egyptian type. On the other hand, a considerable pro- portion of the plants adhered very closely to the characters of the 210 CHARACTERS OF HINDI HYBRIDS. 31 Upland type. Several of these were distinctly clustered and some were quite limbless, like the Upland parent, though the majority did not have the shortened internodes. Coherence of characters was shown very conspicuously in the fact that all of the definitely clustered or limbless plants had the Upland type of foliage, all were quite hairy, and all had white petals, as in Upland cotton. The only definite mark of hybridization on several of these plants was the purple spot at the base of the petals. When the purple spot was lacking there was no definite evidence of hybridi- zation, but some plants that would have been taken for pure Upland in all other respects had very faint spots, showing that they were hybrids. There was no complete dominance of the yellow flower color as reported in some Egyptian-Upland hybrids. None of the yellow flowers were as yellow as those of Egyptian cotton. All of the yellow flowers had pale-purple spots at the base of the petals. Some of the white flowers had spots as dark as any of the yellow flowers. In this respect the hybrids may be said to afford an example of the Mendelian law of free combination, but these variations occurred in the first generation, where Mendelian crosses are expected to give more uniform results. Another lot of hybrids produced by Mr. Fletcher by fertilizing an Upland cotton from Cochin China with pollen of the Voltos variety of Egyptian cotton showed quite a contrast in comparison with the preceding series. Nearly all of these plants looked like ordinary first-generation Upland-Egyptian hybrids, except one that showed only Upland features. But the white petals had small purple spots as an evidence that the plant represented a true hybrid, not merely a result of accident in manipulation. The plant was very hairy and the leaves and bolls showed no departure from Upland characters. All other plants of the cross had pale-yellow flowers, and all the flowers had the spots pale, sometimes entirely wanting. The spot character would have to be reckoned as nearly recessive, but not quite completely so. Two plants were found in the same lot that might have been taken for ordinary Egyptian individuals, unless it were for too much hair, but one plant was more hirsute than the other, especially on the under side of the leaves, where the stellate hairs developed into noticeable tufts. This also must be taken as a sign of hybridity. The other plant was somewhat abnormal, in that it produced several sterile involucres composed of only a single bract. In a third lot of hybrids between the Voltos variety of Egyptian cotton as the female parent and the Cochin China Upland as the male there were several more plants of a complete Upland type. Three of these plants had been grown from fuzzy seeds that appeared in the Voltos cotton, an indication that the variety was not pure. 77267 °—Bul. 210—11 GQ. v 82 HINDI COTTON IN EGYPT. The habit of these plants was much like the Cochin China parent and also closely similar to that of the Rabinal and Pachon varieties of Upland cotton from Central America. The plants were very hairy and the bracts were unusually well closed, as well or better than in the Rabinal cotton, and being also larger they remained closed to a more advanced stage. This character of the closed bracts was also shown among the hybrids. It was fully expressed, or even intensi- fied, in some of the plants that had yellow flowers and other unmis- takable evidences of hybridity. Well-closed hairy bracts have value as a weevil-resistant character, since they exclude the insects from the young buds.* The phenomenon of coherence of characters is not only of interest from the standpoint of the scientific study of heredity, but is of distinct practical importance in relation to the problem of develop- ing and maintaining uniformity in cultivated varieties. It repre- sents on the one hand a limitation of the power of the breeder to make free combinations of the characters of different species, as in ordinary Mendelian hybrids, but on the other hand it assists in main- taining the uniformity of established strains and guarding them against contamination. If there were no coherence in the expres- sion of the characters any Hindi character could come into expres- sion independent of any other. The work of selection would involve a detailed inspection of each plant by all of its characters and would require an amount of time that would make it entirely impracticable as a farm operation, even though the farmer should acquire the neces- sary skill. In short, it is the fact of coherence of characters that lends value to selection, that makes it possible by roguing to improve or maintain the quality of the crop. The success of the Egyptian method of securing commercial uni- formity by matching the color of the fiber rests also on the fact that variations in the color of the lint are not independent of other char- acters. The inferior lint of the Hindi plants and hybrids does not have the same color as the lint of Egyptian plants. If there were no coherence of the Hindi characters the brown color would be found in combination with the naked seeds and short lint of the Hindi type, but this seems never to occur. Recognition of the principle of coherence calls attention to the practical fact that plants seldom make serious changes in the expres- sion of one character without showing changes of expression on other characters. The plants that produce the inferior lint in the fall are those that have departed from the regular courses of development earlier in the season. Indeed, these departures from normal heredity 4 Weevil-Resisting Adaptations of the Cotton Plant, Bulletin No. 88, Bureau of Plant Industry, U. 8S. Dept. of Agriculture, 1906. 210 CHARACTERS OF HINDI HYBRIDS. Bf) can usually be recognized much more readily by inspecting the vege- tative characters of the plants in the earlier stages of development than after the crop is ripe and the damage of cross-fertilization has been done. It takes only an instant to see that the foliage or the habits of growth of a plant are different from those of its neighbors, much less time than is required to judge plants by their lint and seed characters at maturity, after the external differences of leaves, flowers, and bolls are no longer to be appreciated. The breeder in search of new varieties may find it desirable to pre- serve all the sports or freak plants that he can find to see whether in some rare cases they may not prove superior to normal plants of the variety, but the farmer who follows this course will lead his variety to degeneration. He must rely on the fact that the vast majority of the plants that diverge from the characters of the variety represent degenerations. His policy is to pull all the aberrant plants as soon as they can be detected. If allowed to remain, they will destroy the uniformity of the stock.¢ INTENSIFICATION OF CHARACTERS IN HYBRIDS. Another deviation from the Mendelian expression of characters in cotton hybrids is found in cases where characters are suppressed or intensified beyond the range of variation of the parental types. The crossing of the Egyptian cotton with short-staple Upland varieties” 44 writer in the Liverpool Daily Post and Mercury (Saturday, March 12, 1910) maintains that periods of prosperity for the Egyptian cotton industry haye followed the introduction of new varieties and that periods of depression ensued as the varieties degenerated: “It is to be remarked that each time a new variety of seed was sown for the first time of cultivating an increase was immediately obtained of 1 to 14 eantars weight per feddan, and as high as 12 to 14 per cent in the ginning yield. This increase diminished with the passing years and by slow degrees the seed degenerated. The excellent results of the beginning did not bear out their early promise, and after a lapse of time of more or less duration the seed cultivated had to be abandoned to give place to a new yariety. * a “And it is the same story. As in 1862, when the Jumel, old and degenerated, had to be abandoned, as in 1892 the Ashmuni had to be replaced by Mitaffifi, so to-day the Mitafiifi seems coming to the end of its career, and no one can deny the degeneration of quality. “While in 1891, 1892, and 1893 it yielded 7 to 8 cantars per feddan on the best lands and 5 to 6 on the others, at the ae day it never gives either 7 or 8 cantars, and in Lower Egypt its production has certainly diminished hy 1 to 13 cantars per feddan on an average. This cotton, which during the first years of its cultivation yielded 110 to 114 in ginning, no longer gives to-day more than 101 to 103, and that with difficulty. ‘ “Seventeen years, therefore, had sufficed for the degeneration of Jumel, and it is exactly after the same lapse of time that we are forced to notice the degeneration of Mitaflifi.” ) 210 ae HINDI COTTON IN EGYPT. usually results in a favorable intensification of the lint characters in the first generation. Notwithstanding the inferiority of the lint of the Upland parent, the lint of the hybrid is usually longer and stronger than that of a pure Egyptian progeny grown under the same conditions.? A form of intensification occasionally shown in Egyptian-Upland hybrids is an unusual development of the nectaries. An excellent example of this was found in an aberrant plant at Calioub, July 12, 1910. It was probably a Hindi hybrid, though showing no pro- nounced Hindi characters. It was much taller than its neighbors and had unusually long basal internodes on the fruiting branches, while the other internodes were short and imperfect. Many buds had aborted and no bolls had been set. Each of the involucres that remained on the plant, 15 in number, had a large nectary on each of the three bracts. In order to give a more definite indication of the extent of intensi- fication shown by the nectaries of this plant, notes were made of the occurrence of nectaries on the involucral bracts of six adjacent plants, one of which happened to be Hindi. The lower buds of the Egyptian plants were generally without nectaries, unlike the Hindi plant which had nectaries on the early as well as on the later involucres, though with no such regularity as in the aberrant plant, to say nothing of the much larger and more regular size of the nectaries of the aber- rant plant. Table III shows the distribution of nectaries on all the involucral bracts of the Egyptian and Hindi plants. Bracts with large nectaries are indicated as “N,” those with small nectaries as “n,” those with no nectaries as “o.” No nectaries as large as those of the aberrant plant were found on any of the neighboring Egyptian and Hindi individuals. Several other plants were examined in addi- tion to those that were definitely counted. One of the Egyptian plants had an involucre with only two bracts, a not uncommon occurrence. “Suppressed and Intensified Characters in Cotton Hybrids, Bulletin 147, Bureau of Plant Industry, U. S. Dept. of Agriculture, 1909. 210 CHARACTERS OF HINDI HYBRIDS. 85 TABLE III.—Census of nectaries of Egyptian and Hindi cotton plants. Adjacent plants. Adjacent plants. | a _| Aberrant | plant. | | Egyptian. Egyptian. Egyptian. | Hindi. Egyptian. Egyptian. | ono | onn 000 NNN/| ooo 000 000 000 000 000 NNN 000 000 000 000 nnn 000 NNN onn ono ono COVCay (ay |) oa value 000 NNN nno 000 oon | 00 NNN nno noo 000 NNN 000 onn oon NNN ono 000 nn N N N oon 000 nno NNN non onn oon NNN ono NNN NNN NNN NNN NNN | Another example of a notable departure from parental characters was shown in a block of hybrids produced by Mr. Fletcher by cross- ing two Egyptian varieties. The whole block showed a remarkable susceptibility to a disease of the roots similar to the wilt of the United States. The whole block of plants was notably different in behavior from either of the three other blocks of hybrids that in- closed it on three sides; the other side bordered on a roadway. All of the plants were small, with a very open habit of growth, and their foliage was tinged with red. Many of the roots were dead or dying and had changed to a grayish-brown color. The contrast between this block and its neighbors was very distinct out to the square corners, with the larger and more healthy plants on either side. . Microscopical examination by Mr. Fletcher found the fibro-vascu- lar bundles of the roots stuffed with fungous mycelium. There seemed to be no escape from Mr. Fletcher’s view that this particular stock of hybrids was unusually susceptible to the disease in compari- son with the surrounding stocks. The peculiarity may have come, of course, from one of the individual plants that happened to be used as parent of the cross, but this does not diminish the value of the evidence that some members of the Egyptian type may have marked susceptibility to the disease. Mr. Fletcher has noted other indications of such suseeptibility and is inclined to believe that the 210 36 HINDI COTTON IN EGYPT. disease may be an unrecognized cause of much damage to the crop. It appears that the symptoms are generally more pronounced on land that had cotton the year before, but the aseeyidiorins have not extended far enough to establish this point. RELATIONSHIPS OF HINDI AND EGYPTIAN COTTONS. The Egyptian cotton in the United States is exposed to the addi- tional danger of crossing with the American Upland type of cotton. It is quite as important to guard against this danger as to exclude the Hindi contamination that has caused so many difficulties and losses in Kgypt. Experiments indicate that the result of allowing the Egyptian cotton to be crossed with Upland pollen will be much the same as with the Hindi, and this is also to be expected from the fact that the Hindi cotton shares many of the characters of Upland cotton, and especially those of some of the types of U pland cotton that have been discovered recently in southern Mexico and Central America.® Though differing in minor details, there is a general agreement between the American Upland types of cotton and the Hindi in the habits of growth, the form, color, and textures of the leaves, in- volucres, and flowers. The external characters of the bolls are also much the same. The principal difference hes in the character of the seeds. In the American Upland cottons the seeds are generally covered with a dense coat of short fuzz, though some of our varie- ties show frequent variations in the direction of naked seeds, like those of the Hindi cotton. Indeed, there are occasional variations where the lint and the fuzz are both lacking, showing that the seed characters of the Hindi cotton he within the range of variation of the Upland type. Thus if the parentage of a hybrid plant is not known it may be impossible to determine whether it represents the Hindi contamination or an Upland cross. In general it may be assumed that plants with hairy stems and leaves represent Upland hybrids rather than Hindi, for the typical Hindi cotton is not hairy. Yet a few hairy Hindi-like plants have been found in Egypt as well as in plantings of imported seed in Arizona. From the standpoint. of the study of heredity it would be very desirable to determine when the Hindi contamination of the Egyp- tian cotton took place. The Hindi variations may represent a recent admixture or the crossmg may have taken place so far back as to represent a general constitutional tendency to reversion pervading the whole Egyptian type. The idea that the Hindi cotton grew as a wild weed in Egypt would allow us to suppose that the process of. “Origin of the Hindi Cotton, Circular No. 42, Bureau of Plant Industry, 18iGaSh Dept. of Agriculture. 210 f RELATIONSHIPS OF HINDI AND EGYPTIAN COTTONS. 37 contamination had been continuous, with some new crosses every year to replace those that were removed by selection. But the idea. of wild cotton in Egypt and also the theory founded upon it seem altogether improbable. The sources of the Hindi contamination must apparently be sought farther back. Another possibility is that the Hindi cotton was formerly culti- vated in Egypt before the present so-called Egyptian type was intro- duced and that the mixing occurred while the Egyptian cotton was replacing the Hindi. A difficulty with this idea is that the lint of the Hindi cotton is so sparse and short as to make its cultivation seem improbable. But it is possible that Hindi plants now appearing as reversions among the Egyptian cotton do not fully represent the possibilities of the Hindi type in the direction of lint production. While there is a general tendency to sparse lint among naked-seeded types of cotton, this is not universal. A strain of Caravonica cotton erown in Hawaii has very abundant lint, in spite of the fact that the seeds are entirely devoid of fuzz, as shown by samples recently deposited with the Department of Agriculture by Dr. E. V. Wilcox, Director of the Hawaii Agricultural Experiment Station. Mr. Fletcher has recent information indicating that Hindi cotton is still planted as a crop in Mesopotamia under the same name as in Egypt. Plants grown at Gizeh by Mr. Fletcher from seed received from Mesopotamia were carefully examined and seemed to show all the essential characters of the Hindi cotton. (See Pl, Eo) tast pos- sible. therefore, that the Hindi admixture may be traced by way of Mesopotamia. The idea that the Mediterranean countries were limited to Old World types of cotton (Gossypium herbaceum, and its relatives indicum, arboreum, etc.) even in ancient times may prove to be erroneous. In southern Italy an Upland-hke cotton is cultivated under an ancient name “bombage,” evidently cognate with the Greek “bombax.” The plants are quite small and somewhat hairy, lke American Upland cotton, but the bracts are very strongly toothed after the Hindi fashion. In this connection it may be well to mention the fact that a sample of seed of brown, rough-fibered cotton has recently been received from northern Arabia by the United States Department of Agricul- ture. While these seeds and lint do not closely resemble those of any recognized variety, they show more of an approach to the Egyptian qualities than any samples previously seen from the Old World. Another small sample of seeds and lint, received about the same time from Honduras, has a much closer resemblance to the Egyptian cotton and is stated to represent a native tree cotton. These seeds have the size and shape of Egyptian seeds with tufts of brownish 210 38 HINDI COTTON IN EGYPT. fuzz at the ends, and the lint is similar to that of the Egyptian cotton, whereas the seeds from Arabia are covered with a brown fuzz. While at Gizeh there was also opportunity, through the kindness of Mr. W. Lawrence Balls, botanist of the Khedivial Society, to see living plants of a kidney cotton raised from seed brought from the Niam-Niam country in the upper valley of the White Nile, a type considered by Mr. Balls as representing one of the parents of the Egyptian cotton. It has to be admitted that these plants show a notable agreement with the Egyptian cotton in many respects and are quite unlike any of the varieties of kidney-seeded cotton that have been seen in Mexico and Central America or received from those countries. The Niam-Niam cotton has three external nectaries present with great regularity, reniform-cordate in shape, and usually distinctly emarginate on the upper side. The nectaries are always of a red color, at least on these well-exposed plants. Inner nectaries are also present with much regularity, are broadly V shaped, and often colored red. The surfaces of the nectaries are rather coarsely granular-papillate and without hairs. Cases of supposed intensi- fication of nectaries in Egyptian hybrids might be considered as reversions to such an ancestor as this. The leaves vary from entire toe 5 lobed, the latter usually on the ‘ank growth of new shoots. Occasionally there are 6 or 7 lobes, but the additional lobes are usually small. The leaves are of the Egyp- tian form and color, somewhat more hairy than usual in Egyptian cotton, but the hairs are short, as in some variations of the Egyptian type. The pulvinus and veins are green or tinged with dull reddish, as in Egyptian cotton. The pulvinus is very hairy and not enlarged, but the outer pairs of veins show an occasional tendency to unite at the base. There are 1 to 3 leaf nectaries, those of the midribs being sagittate. The stipules of the main stalk and vegetative branches are long and slender as in rank-growing Egyptian cotton, while those of the fruiting branches are unequal, one narrow and the other broad, the latter often with two teeth. The bracts are usually connate at their base for one-eighth to one- fourth inch, as often occurs in Egyptian cotton. The calyx has very distinct, broadly rounded lobes (Pl. V, C), more prominent than is usual in the Egyptian cotton but nearly equaled under some condi- tions, as in the Egyptian cotton grown near Los Angeles in the season of 1909. The plants at Gizeh were quite woody and about 10 feet high, and had no tendency to produce elongated fruiting branches. Only one flower was borne on each fruiting branch. The pedicels of the flowers 210 ee RELATIONSHIPS OF HINDI AND EGYPTIAN COTTONS. 39 were very short and subtended by a small leaf, usually with one stipule very much enlarged and often toothed, somewhat like an in- volucral bract. One of the most striking peculiarities in which the Niam-Niam cotton agrees with the Egyptian is the tendency to enlargement of one of the stipules of the leaves of the fruiting branches. It has been noticed in Arizona that abnormally large strong-growing plants of Egyptian cotton often have this tendency very pronounced, a fact suggestive of the possibility that such plants may represent reversions toward an ancestral form similar to the Niam-Niam cotton. The unequal development of the stipules has been considered in relation to Hindi hybrids, but such a tendency does not seem to be as pro- nounced in the Hindi hybrids as in the Egyptian cotton and in this African relative. Enlarged stipules are especially likely to be found in Egyptian cotton on leaves of short branches produced from the fruiting branches and may be connected with the tendency of such branches to produce organs intermediate between the ordinary leaves and the involucral bracts. While the Niam-Niam cotton must certainly be considered in the study of the relationships of the Egyptian cotton, it seems more likely to prove a collateral relative than a direct ancestor. It is very difficult to believe that the Egyptian cotton descended from a kidney-seeded ancestor or from one that had the fruiting branches so shortened and specialized as the Niam-Niam cotton. The most significant thing regarding these cottons from Mesopo- tamia and central Africa is that they may add something to the evi- dence of the existence of genuine Old World varieties of the Upland type of cottons. The Upland variety from Cochin China recently brought forward by Mr. Fletcher as an ancestor for our American Upland cottons is also very interesting from this standpoint." As seen growing at Gizeh the Cochin China cotton shows a remark- able resemblance to some of the Central American varieties and especially to two types from the Central Plateau and the Pacific slope of Guatemala, those that have been described as Pachon and Rabinal. The Guatemalan Upland cottons and other related types from south- ern Mexico show very close agreements with the Hindi cotton in so many of the characters that a rather close relationship must be supposed to exist. This renders the close resemblance of the Cochin China cotton to the Central American varieties all the more interesting. The Cochin China cotton shows in Egypt the same bushy habit of growth with many upright vegetative branches as the Central Ameri- “Pletcher, F. The Origin of Egyptian Cotton, Cairo Scientific Journal, vol. 2, no. 26, November, 1908. 210 40 HINDI COTTON IN EGYPT. can Upland cottons when first brought to the United States, though not carried to quite the same extent under the less extreme Egyptian conditions. The stems, leaves, and involucres are densely hairy as in the Central American cottons. The bracts also have the margins hairy and very firmly appressed in the same way as in the Central American cottons and perhaps to an even greater extent. The lobes of the calyx have the same tendency to grow into long teeth (Pl. V, 4), and the bolls have the same conic-oval, abruptly apiculate form which several of the Central American varieties share with the Hindi cotton. In short, the resemblance seems so complete that if the Cochin China cotton had been found in Central America it would have been considered as only one more of the relatively shght local variations shown by the general type represented by the Rabinal and Pachon varieties. The most notable difference was an apparent absence of bractlets, but this condition could probably be found on second-year wood in the Central American varieties. While the Cochin China cotton, like the Central American varieties, appears to be a relative of our American Upland cottons, there are native Mexican varieties that seem to be still more closely related to some of our United States Upland varieties. Yet it is not impossible that Mr. Fletcher’s idea of tracing the Cochin China cotton to the United States through an early introduction of so-called “ Siam cotton ” may turn out to be true of our long-staple Upland type still grown in Louisiana. If the Cochin China cotton were more nearly identical with our United States Upland cottons it might be looked upon as an introduc- tion from the United States, but it is much less hkely that a local Central American variety has been carried to Cochin China. The information of Mr. Fletcher’s correspondent, that this cotton was really indigenous in Cochin China, may therefore be credited. While the existence of these additional relatives of the Egyptian and Hindi types of cotton in the Old World does not affect the evi- dences of relationship that have been pointed out between these types of cotton and others that appear to be natives of America, it does have a bearing upon the question of how these members of American types of cotton reached the Old World. If many sorts like the Hindi. Egyptian, Niam-Niam, and Cochin China cottons are found in differ- ent parts of the Old World it will not be reasonable to believe that they represent recent importations from America, since the time of Columbus. Jt will be necessary to consider the possibility that American types of cotton, like the-coconut palm, sweet potato, and @¥Fletcher, F. The Botany and Origin of American Upland Cotton. Cairo Scientific Journal, vol. 3, no. 38, November, 1907, p. 263. 210 SUPPOSED INCREASE OF HINDI COTTON. 41 other economic plants of American origin, were carried across the Pacific Ocean in prehistoric times.” If our long-staple varieties of Upland cotton originated in the East Indies it is reasonable to expect that other superior types of Upland cotton may be found in that part of the world. Indeed, Mr. Fletcher’s Cochin China cotton seems to be a promising type, worthy of attention from the standpoint of acclimatization. The bolls are larger than in our long-staple Upland varieties and the lint is of good length. The very large and well-closed hairy involucral bracts would have value from the standpoint of weevil resistance, like the similar bracts of the Central American varieties which exclude the boll weevils from the young buds, as already noted in describing the hybrids of the Cochin China cotton.? SUPPOSED INCREASE OF HINDI COTTON. The popular belief in Egypt is that the proportion of Hindi cot- ton is increasing, though there seems to be no way to obtain definite information on this point. Intelligent natives declare that they “Wood Plants of Ancient America, Smithsonian Report 1903, pp. 481—497. Agricultural History and Utility of the Cultivated Aroids, Bulletin 164, pt. 2, Bureau of Plant Industry, U. S. Dept. of Agriculture, 1910. History of the Coconut Palm in America, Contributions from the United States National Herbarium, vol. 14, pt. 2, 1910. 5’ The successful cultivation of a so-called ‘“‘ Cambodia” cotton in British India has been noticed in a recent Consular Report, issued while this bulletin was in preparation. The facts are of special interest in view of the many unsuccessful experiments that have been made in India with Upland varieties from the United States. The statement is as follows: “In Tinnevelly district, Madras Presidency, at the extreme southern end of the peninsula, there had been planted up to October about 17,000 acres in what is known as Cambodia cotton. This is a variety of acclimatized American cotton, introduced into the country about four years ago, which is being quite successfully grown and which yields far more fiber per acre than any of the old varieties. “Last year a total of 15,000 bales of Cambodia was produced on 15,000 acres of the black soil of Tinnevelly, and this season, in addition to the larger area already reported as planted in that district, the agricultural department is experimenting with it in several other parts of the Presidency with a view to its general adoption by growers. It is said to thrive on irrigated lands, and should it prove even partially as successful in other districts as in Tin- nevelly, there is little doubt that within a very few years it will be grown throughout the whole of south India, if not elsewhere in the country. “As the fiber of the Cambodia compares favorably with that of American Upland cottons, it is not too much to say that India may within a few years become a serious competitor of the United States in meeting the world’s demand for the commodity, instead of furnishing only the inferior grades as at present.” (Report of Nathaniel B. Stewart, consul at Madras, India, in Daily Consular and Trade Reports, December 17, 1910.) 210 49 é HINDI COTTON IN EGYPT. remember when the Mit Afifi or the Jannovitch varieties produced fields of uniform plants, all of the same height, with none of the irregularities now shown by the tall hybrid plants of the Hindi- infested fields. But in the absence of any actual countings in former years it is not possible to determine what change has taken place. From the standpoint of the Mendelian theory of heredity an in- creased representation of the Hindi characters would not be expected to occur unless additional contamination took place from outside sources, which appear to be lacking in Egypt. Mathematicians have shown that characters expressed according to the Mendelian theory would not tend to increase, but would remain at the same general proportion in a mixed population.* Nevertheless, an increasing dominance or stronger tendency of expression of the Hindi characters should not be dismissed as impos- sible, for it has been noticed in experiments with Egyptian-Upland hybrids that the Upland characters seem to attain a more and more predominant expression in the later generations, even when selec- tions are made with a view to preserve the Egyptian or intermediate characters among the hybrids. Though no direct statistical evi- dence regarding the supposed increase is likely to be obtained, it may be possible to throw ght on the question indirectly by the study of the tendencies of expression shown in artificial hybrids be- tween the Egyptian and Hindi types. Experiments of this kind were begun by the making of such hybrids in Arizona in the season of 1909. The popular impression of a gradual increase in the proportion of Hindi cotton is supported by the general opinion of the commercial world that the quality of the Egyptian cotton is declining. This may mean that poorer qualities are being sent out under the same marks or that the ginning establishments are finding it more difficult to keep their product up to recognized standards. Either of these results, or both, might naturally be caused if the Hindi cotton continues to multiply in the face of the selection that is now being applied.’ Considered on a percentage basis, a considerable amount of selec- tion has undoubtedly been directed against the Hindi cotton. In “Hardy, G. H. Mendelian Proportions in a Mixed Population. Science, n. s., vol. 28, p. 48, July 10, 1908. ’'The idea of a progressive deterioration of the Egyptian product is confirmed by a recent authoritative statement published while the present report was in preparation: ‘“ There is no gainsaying the unanimous evidence that the genera] character of Brown Egyptian cotton [by which Lancashire means Afifi] has gone down most markedly from the standard of 15 years ago. AIl the spinners of fine counts, to whom strength is everything, speak with regret of the Afifi of those days. Without exception they say that during recent years they have continually been compelled, in order to maintain their standards of strength, 210 DAMAGE FROM THE HINDI CONTAMINATION. 43 the Delta region a large proportion, probably 50 per cent or more, of the Hindi plants that germinate in the fields are rogued out. The sorting of the fiber in the ginhouses must take out a still larger per- centage of the Hindi cotton that is harvested. Some of the ginners are also said to sift out the smooth seeds, or even to resort to hand picking to keep the smooth Hindi seeds from being planted. While it is to be expected that the various ginning establishments would be found to differ greatly in the thoroughness with which these precautions are observed, the general effect must be to exclude a large proportion of the Hindi seed every year. Under any Mendelian rule or other customary idea regarding the effects of selection it might be expected that the expression of the Hindi characters would have declined long since to a neghgible quantity, but the facts certainly do not correspond to this expectation. The result demonstrates instead that the system of selection now in opera- tion is entirely inadequate to eliminate the Hindi variations. As already noted in connection with the seed characters of the Hindi cotton, the tendency to an increased representation of this type is not limited to the factor of prepotency, but may prove to be due partly or wholly to more prompt germination of the seeds, owing to the absence of fuzz that allows more effective contact with the soil. Experiments with other types of cotton have shown that varieties having less fuzz germinate more promptly, but comparisons will also be made between Egyptian and Hindi. ESTIMATE OF DAMAGE FROM THE HINDI CONTAMINATION. As the percentages of Hindi cotton in the Egyptian fields do not represent the full amount of Hindi contamination, so they do not indicate the full extent of damage to the crop. In addition to the true Hindi plants and the obviously Hindi-like hybrids, supposed to represent the first generation, more careful inspection always shows a considerable number of obscure or dilute hybrids as well as many individual variations that may reasonably be ascribed to the same general fact of Hindi contamination. These aberrant plants include those that show the white flowers, the flowers with pale spots, and other peculiarities that can often be detected only by to raise the mark or grade of cotton they use, and to add increasing propor- tions of superior varieties, such as Nubari and Jannovitch, merely to obtain the same results as they formerly secured with Afifi alone. Strength is abso- lutely essential in the manufacture of ‘twist’ yarns for warping, and in spite of improved spinning processes, greater loss in waste through taking out a larger proportion of short staple, and more careful and costly methods gen- erally, the spinners have had the greatest difficulty in maintaining the quality of their yarns.” (See Todd, John A., “The Market for Egyptian Cotton in 1909-1910,” L’ Egypte Contemporaine, no. 5, January, 1911, p. 5.) 210 44 HINDI COTTON IN EGYPT. careful comparison of all the parts, including the seeds and lint. A complete census of the aberrant plants of a field requires too much time to make it generally feasible. Moreover, the cotton in Egypt was not yet far enough advanced in July, 1910, to allow such a study to be completed. The visit was made at that season because the vegetative characters of Hindi plants were known to be more readily visible at that time. Counts made in a field of Ashmuni cotton raised in Arizona in 1909 from imported seed gave over 40 per cent of the plants showing distinct departures from the normal characteristics of Egyptian cotton, mostly in the direction of the Hindi.