BANANA WILT 
 
 BY 
 
 E. W. BRANDES 
 
 A DJ SSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE 
 
 REQUIREMENTS FOR THE DEGREE OF DOCTOR OF 
 
 PHILOSOPHY, IN THE UNIVERSITY 
 
 OF MICHIGAN 
 
 Reprinted from Phytopathology, Vol. IX, No. 9, 
 September, 1919 
 
BANANA WILT 
 
 BY 
 
 E. W. BRANDES 
 
 A Dl SSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE 
 
 REQUIREMENTS FOR THE DEGREE OF DOCTOR OF 
 
 PHILOSOPHY, IN THE UNIVERSITY 
 
 OF MICHIGAN 
 
 Reprinted from Phytopathology, Vol. IX, No. 9, 
 September, 1919 
 
*. «f 1. 
 
 JAN 17 1920 
 
 <# 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXI 
 
 Brandes: Banana Wilt 
 
 A banana plant of the Gros Michel variety in Costa Rica attacked by the wilt 
 organism. 
 
Reprinted from Phytopathology, Vol. IX, No. 9, September, 1919 
 
 BANANA WILT 1 
 
 E. W. Brandes 
 
 With Plates XXI to XXXIV and Five Figures in the Text 
 
 contents 
 I. The Host: 
 
 1. Extent of the industry 340 
 
 2. Food value 341 
 
 3. Methods of cultivation 342 
 
 4. Varietal susceptibility 343 
 
 II. The Disease: 
 
 1 . Names 344 
 
 2. History and previous literature 345 
 
 3. Geographical occurrence 347 
 
 4. Economic importance 348 
 
 a. Extent of the losses 348 
 
 b. Nature of the losses 349 
 
 5. Symptoms 350 
 
 a. External signs of disease 350 
 
 b. Internal signs of disease 353 
 
 6. Etiology 356 
 
 a. Morphology of causal organism 356 
 
 1. Sporodochia and conidia 356 
 
 2. Development of sporodochia 357 
 
 3. Mycelium 357 
 
 4. Cultural characteristics 358 
 
 1 "The work on this problem was begun while the writer was connected with the 
 Porto Rico Agricultural Experiment Station at Mayaguez, Porto Rico; it was con- 
 tinued for one and one-half years at the New York State College of Agriculture, 
 Cornell University, and was completed in a final year at the University of Michi- 
 gan. The writer desires to acknowledge his great indebtedness to Professor H. H. 
 Whetzel, of Cornell, for facilities and advice in conducting the work, and to 
 thank Professor F. C. Newcombe, of the University of Michigan, for the facilities 
 of the laboratory and advice in preparing the dissertation." 
 
340 Phytopathology [Vol. 9 
 
 b. Physiology 360 
 
 1. Isolation 360 
 
 2. Relation to heat 361 
 
 3. Relation to moisture 362 
 
 4. Relation to light 363 
 
 5. Relation to oxygen supply 363 
 
 6. Excretions 364 
 
 c. Pathogenicity 367 
 
 d. Names and synonymy 373 
 
 e. Life history 375 
 
 1. Source of inoculum 375 
 
 2. Dissemination 375 
 
 3. Infection courts 377 
 
 4. Penetration 379 
 
 5. Growth in host 379 
 
 7. Ecology 380 
 
 S. Control 382 
 
 1. Exclusion 383 
 
 2. Protection 384 
 
 3. Eradication 385 
 
 4. Immunization 386 
 
 9. Literature cited 368 
 
 I. The Host 
 
 1. EXTENT OF THE INDUSTRY 
 
 The extensive cultivation of bananas {Musa sapientum) in the Ameri- 
 can tropics for exportation to North America and Europe is compara- 
 tively recent. A few bunches were imported into this country early in 
 the past century but were regarded only as a horticultural curiosity. No 
 systematic effort had been made towards production in quantities suffi- 
 cient to create a steady demand or a fixed market price until 1885, when 
 a small fruit company, capitalized at $20,000, was formed in Boston to 
 promote the importation of bananas. The venture proved a success. 
 The numerous obstacles in the way of raising and transporting this very 
 perishable fruit have been gradually overcome and to-day the gigantic 
 industry, supplying approximately 40,000,000 bunches to the United 
 States annually, is on a secure basis and is still growing. The present 
 war has somewhat checked the distribution of bananas to European coun- 
 tries, but importations in 1914 were sufficiently large to encourage the 
 belief that consumption of bananas in those countries would soon rival 
 that in the United States. In general, bananas are retailed by weight in 
 this country, the present mainland market price being ten cents per 
 pound. Sixty pounds of edible fruit would be a low estimate for the 
 average bunch, so it is seen that the American consumers pay at least 
 $200,000,000 annually for bananas. 
 
1919] Brandes: Banana Wilt 341 
 
 The countries which furnish most of the bananas to the American and 
 European markets are, in the order of their importance, Jamaica, Costa 
 Rica, Colombia, Panama and the Canary Islands. Honduras, Guate- 
 mala, Nicaragua, and the Hawaiian Islands also export bananas in con- 
 siderable quantities to the United States. Production of bananas in 
 Cuba for export has declined in recent years owing to the extensive 
 opening up of more suitable banana land in Central America. 
 
 The home consumption of bananas in some tropical American coun- 
 tries is of more importance than the export trade, since in many places it 
 forms the principal sustenance of the natives. Porto Rico does not ex- 
 port bananas, but in the interior they are used largely as a substitute for 
 bread. Practically every peon has a few dozen plants of bananas and 
 plantains and in some places many acres are planted with some attempt 
 at careful cultivation. The city of San Juan alone consumes about 
 3,000,000 dozens of bananas annually. In the country districts where 
 the masses of people live, it takes the place of wheat, corn, and potatoes. 
 This staple food is usually cooked in various ways by the hundreds of 
 thousands of Porto Ricans for whom it is one of the main articles of diet. 
 
 2. FOOD VALUE 
 
 The banana is not merely an agreeable tasting fruit, but is also a food 
 of high value. The composition of ripe fruit (3) is as follows: Water, 
 75.3 per cent; protein, 1.3 per cent; fat, 0.6 per cent; carbohydrates, 22.0 
 per cent; ash, 0.8 per cent; fuel value, 460 calories. 
 
 It is seen that the protein and fat are not high enough in proportion 
 to the carbohydrates to make a perfectly balanced ration, but if supple- 
 mented with a small amount of beans, milk or meat, a diet of ripe ba- 
 nanas would suffice. The ash is composed of the phosphates, sulphates 
 and chlorides of potash, soda, magnesia and lime, all of which serve useful 
 purposes. The following table (25) shows the composition of the ash: 
 
 Corn-position of the ash of bananas 
 
 per cent 
 
 Silica 2 . 19 
 
 Lime 1 .82 
 
 Iron oxide . 18 
 
 Phosphoric acid 7 . 68 
 
 Magnesia 6.45 
 
 Soda 15 . 11 
 
 Potash 43.55 
 
 Sulphur trioxide 3.26 
 
 Chlorine 7.23 
 
342 Phytopathology [Vol. 9 
 
 Ripe bananas are easily digested, cheap, and wholesome. Owing to 
 their many superior qualities, they have won a well deserved popularity 
 in American and European markets. Anj r factor which threatens to 
 curtail the production of this staple crop merits careful investigation, 
 with a view towards eliminating the cause. 
 
 3. METHODS OF CULTIVATION 
 
 A few words in regard to the different methods of banana cultivation 
 in various countries will serve to make clear the necessity for adopting 
 procedures in the investigation of banana diseases, which vary some- 
 what from the usual practice among plant pathologists. The banana is a 
 great perennial herb, requiring from twelve to twenty months from time 
 of planting to attain maturity. The cultivated bananas, Musa sapientum 
 and Musa cavendishii, have small, degenerate seeds, which are function- 
 less and useless for propagating. New plants are started by means of 
 suckers that arise from the tuberous underground stem or rhizome. 
 These suckers are cut away from the mother plant when they are several 
 months old. The top and roots are pruned off and the young "bulb" or 
 "bit" is planted in an excavation made to receive it. The bulbs may be 
 set out immediately or they may be allowed to dry for several days. In 
 Hawaii, some planters place them in heaps and cover them lightly with 
 trash, allowing them to remain so for a month before planting. The sev- 
 ering of the sucker from the mother plant, usually done with a.sharp spade, 
 results in a large wound, from six to ten inches in diameter, depending on 
 the size of the rhizome. In Cuba many fields may be seen in which the 
 tops of the suckers are not pruned off close, but rather at a height of 4 
 to 6 feet above the surface of the ground. Such bulbs are supposed to 
 produce a mature plant in less time, but the root system is apt to be 
 weakly developed and the plants may be blown down in moderately high 
 winds. In Cuba, Porto Rico, Jamaica and the Hawaiian Islands, con- 
 siderable preparation of the land is made. It is cleared, plowed, and in 
 general, the best agricultural practice followed. The plants are placed 
 uniformly in rows, the distance varying from 6 to 10 or more feet apart, 
 both ways, depending on the variety. In Central America, however, 
 when new banana land is opened up, the timber is cut and allowed to rot 
 where it falls. Bulbs are merely thrown into holes dug among the fallen 
 logs and nature is allowed to do the rest. The weeds and brush are cut 
 with machetes two or three times a year, but further than that no care 
 whatsoever is given the plants. 
 
 A banana plant produces but one bunch of fruit, which is harvested by 
 cutting down the whole "stem," a sucker from the rootstock being allowed 
 
1919] Braxdes: Banana Wilt 343 
 
 to take its place. The sucker may have arisen a foot or more from the 
 old plant, so it is evident that after some years the stool may have mi- 
 grated to some distance from the original point where it was set out. 
 This results in irregularity of the rows. In Central America, where no 
 tillage is employed, this is a matter of indifference, but where the land is 
 tilled with horse drawn implements, it is necessary to replant the fields 
 to restore regularity. Lands must be suitably drained, and in arid regions, 
 such as the south coast of Jamaica and the banana districts of Colombia, 
 irrigation is employed. 
 
 4. VARIETAL SUSCEPTIBILITY 
 
 Banana wilt, the disease with which this paper is concerned, exhibits a 
 decided variation in its ability to attack the different cultivated varieties 
 of bananas. It is a curious fact, that in any particular region, this dis- 
 ease attacks most severely, and sometimes exclusively, the variety which 
 is there most esteemed and therefore most widely planted. Other varie- 
 ties may be planted in the immediate vicinity and optimum conditions 
 for infection appear to exist, but they are not attacked. It is also in- 
 teresting to note that varieties which are strongly attacked in one country 
 appear to be resistant in another, although the disease is present in great 
 abundance on some other variety. It will of course suggest itself imme- 
 diately that we have here strains of the pathogene which exhibit bio- 
 logic specialization. 
 
 Owing to the great number of varieties of the banana (estimated by 
 some at more than a thousand) and the fact that in different countries the 
 same varieties show variations, especially in regard to size and habit of 
 growth, the problem is manifestly complicated. In addition, the bo- 
 tanical nomenclature' is confused, and different local names are used for 
 the varieties, even in different parts of the same country. 
 
 Typical cases of banana wilt have been observed by the writer affecting 
 only four varieties, namely the Chamaluco in Porto Rico, the Manzana in 
 Cuba, the Gros Michel in Cuba, Panama, Costa Rica, Guatemala, 2 Hon- 
 duras, 2 and Jamaica, and the Red banana in Panama. The organism 
 was easily isolated from all of these. In addition to the four varieties of 
 Musa sapientum listed above as having proved susceptible, it has been 
 suggested by one or two investigators and a number of planters, that other 
 varieties and even other species of Musa are attacked. Among them are 
 included the Dwarf or Chinese banana, Musa cavendishii, and the culti- 
 vated plantain, Musa paradisiaca. Dr. J. R. Johnston, Plant Patholo- 
 
 2 Mr. L. W. Waters of the United Fruit Company gave the writer cultures of the 
 pathogene from plants in Guatemala and Honduras. 
 
344 Phytopathology [Vol. 9 
 
 gist of the Cuban Experiment Station at Santiago de las Vegas, reports 
 having seen affected dwarf bananas in Panama in 1912, and to have iso- 
 lated a species of Fusarium from affected tissues. I was unable to verify 
 this in 1917, although scores of suspected plants were examined in the 
 same region. However, there is no reason for doubting that this species 
 may be affected under suitable conditions. The plantain is more vigorous 
 than the banana, but occasionally plants are seen which are unthrifty 
 for some reason or another, and planters are apt to ascribe the condition 
 to "Panama Disease" or banana wilt. A number of such plants have been 
 carefully examined, and in one of them a Fusarium was found occupying 
 the tracheae in the pseudostem. Upon cultivation, however, it proved 
 to be a member of the section Discolor of the genus Fusarium. These 
 have also frequently been found associated with the true pathogene in the 
 last stages of the disease and are to be regarded as secondary infections. 
 The presence of Fusaria in the vascular tissue must therefore not be 
 taken to mean that they are the causal agents. 
 
 II. The Disease 
 
 1. NAMES 
 
 Various names have been applied to this disease, since it was first defi- 
 nitely recognized as such, among them being "the disease," "banana 
 disease," "Panama disease," "banana blight," "banana wilt," "droop," 
 "tired bananas," etc. in English-speaking countries, and "la enfermadad 
 del platano," "la enfermadad" and "enfermadad Panama" in Spanish 
 speaking countries. The name "Panama disease" first gained wide usage. 
 It was probably originally used in Surinam on account of the resemblance 
 of the banana disease which broke out there in 1906, .to the disease which 
 had played havoc in the plantations of Panama for some years prior to that 
 time. It is improbable that the name "Panama disease" would be used 
 first in Panama itself, since it would not there be particularly descriptive. 
 The disease is by no means confined to Panama, but is widespread in the 
 American tropics, being present in Central America, the West Indies, 
 South America, the southern extremity of North America, also in the 
 Hawaiian Islands and probably in the Old World. 
 
 A more satisfactory and descriptive name for the disease is "banana 
 wilt," used in 1915 to describe the disease in Jamaica (Jamaica Department 
 of Agriculture publication "The Law and the Orders Issued in Accordance 
 Therewith with Regard to Diseases of Plants" p. 12, 1915.) Another less 
 serious disease of bananas had previously been called "banana wilt" in 
 Jamaica, but it was subsequently proved to be a bulb rot and not a true 
 wilt. Furthermore, it is now widely known as the "Bonnygate disease," 
 
1919] Brandes: Banana Wilt 345 
 
 so there is very little risk of confusion if the name "banana wilt" is 
 applied to our disease. It is very similar both in symptoms and in iden- 
 tity of the causal organism to the well-known "cotton wilt," "okra wilt," 
 "tomato wilt," "cowpea wilt," "watermelon wilt," "potato wilt," etc., 
 so it is believed that planters will accept the name which is in accordance 
 with the best usage among plant pathologists. 
 
 2. HISTORY AND PREVIOUS LITERATURE 
 
 The first mention of- this disease in the literature seems to have been 
 made by Higgins (17) at Honolulu. It is a very meagre account, but 
 identification of the associated fungus was made by a good mycologist, and 
 since the disease has subsequently been proved to exist there (8, 31), it 
 is probable that what Higgins observed was the true banana wilt. 
 
 It was in Panama and Costa Rica, however, that the disease first at- 
 tracted wide attention on account of its destructive nature. McKenney 
 (23) reports that the disease had attained alarming proportions in Panama 
 and Costa Rica in 1904. He states, probably on the authority of planters, 
 "As early as 1890 a few isolated spots were known to be affected, and 
 from these the spread of the disease can be traced." The same statement 
 was made to the writer in 1917 by a planter of long experience in Panama. 
 McKenney was not able to isolate the causal organism, which is cer- 
 tainly surprising in view of the ease with which it may be cultivated, 
 but his careful description of the symptoms leaves no doubt as to the 
 identity of the disease. 
 
 Dr. Erwin F. Smith (28) at the same time described a disease of ba- 
 nanas from Cuba. He isolated a species of Fusarium from the discolored 
 vessels of diseased material. Pure cultures of this fungus were inocu- 
 lated into the midrib, leaf stalk and pseudo-stem of healthy banana 
 plants at Washington. It was found that the organism would invade the 
 vascular tissue of the inoculated plants to a distance of from 5 to 8 feet 
 from the point of inoculation. The experiment was broken off before any 
 secondary signs of the disease appeared and was not resumed. Doctor 
 Smith named the organism which he isolated Fusarium cubense, but did 
 not publish any technical description of it. He gives no description of the 
 symptoms of the disease,, since he had not at the time seen affected plants 
 in Cuba. He states, however, that the disease is similar to a disease of 
 bananas described by Earle in Jamaica in 1903 (11). This is certainly 
 due to his not having seen banana blight in the field, since Earle's de- 
 scription shows conclusively that his disease has no connection with true 
 banana wilt. 
 
 In 1911, Essed (13) describes what was undoubtedly banana wilt in 
 Surinam, but his investigation bears all the earmarks of hurried work. 
 
346 Phytopathology [Vol. 9 
 
 In 1912 Drost (10) also working in Surinam published a quite extensive 
 account of his researches on what was unquestionably banana wilt. His 
 conclusion that it is caused by an ascomycete, Leptospora musae, is 
 based on insufficient evidence. 
 
 Research of a careful nature is indicated by the papers of Ashby (1, 2). 
 He states that the banana wilt was first noticed in Jamaica in 1911. A 
 good description of the symptoms are given. He obtained cultures of a 
 Fusarium from banana tissues in a not far advanced stage of the disease, 
 by the poured plate method. A very good description of the growth and 
 appearance of this organism in pure cultures, the first to appear in print, 
 is found in this paper. No inoculation experiments were attempted and 
 no further contribution to our knowledge of the life history of the parasite 
 is made. Recommendations for control of the disease, chiefly prophylactic, 
 and rather drastic are set forth. It is suggested that the destruction of all 
 plants within a radius of 66 feet of a case of banana wilt be accomplished 
 and the area fenced and quarantined for an indefinite period. 
 
 Johnston (18) reports that banana wilt had become very destructive in 
 the western end of Cuba in 1915. The Manzana variety was most sub- 
 ject to the disease and the Johnson (Gros Michel) was also strongly 
 attacked. 
 
 Fawcett (14) gives the first published report of banana wilt in Porto 
 Rico in 1910. In his brief account, he states merely that the disease is 
 occasionally troublesome, and that two years previously a species of 
 Fusarium was isolated from diseased materials. In a later paper (15) he 
 records inoculation and control experiments, both of which gave negative 
 results. 
 
 The present writer (5 and 6) records having isolated the organism from 
 cases of banana wilt and also from the soil in banana plantations and 
 adjacent fields in 1915. A detailed account of proof of causation, which 
 will be repeated in the present paper, is also given. 
 
 Basu (4) notes the presence of a banana wilt in India, the symptoms of 
 which agree very closely with the disease in the New World. The asso- 
 ciated fungus is a vascular parasite of the genus Fusarium. It is de- 
 scribed as having exterminated the most profitable variety of bananas at 
 Chinsurah, while other varieties are not at all attacked. 
 
 Tryon (29) states that the disease has long been prevalent in Australia, 
 but judging from his account it does not cause serious damage. His 
 account is very incomplete and it is by no means certain that the disease 
 he mentions is identical with ours. He speaks of a vascular parasite, 
 causing discoloration of the bundles, and concludes from a consideration of 
 its symptoms that it is identical with the " Panama disease." 
 
1919] Brandes: Banana Wilt 347 
 
 Rijks (26) has recently described a banana disease in the Saleyer 
 Islands, Dutch East Indies, which resembles banana wilt in many ways. 
 The symptoms as described are not exactly typical of the disease in Cen- 
 tral America and the West Indies but the differences may be due to the 
 fact that the varieties of bananas grown in the two regions are not the 
 same. No organism is mentioned in Rijks' paper. 
 
 Butler (7) and Shaw (32) report diseases of the banana in India in 
 which a Fusarium is involved, but they are unmistakably rots and not 
 true wilts. Many other papers on banana wilt and similar maladies have 
 appeared but for the most part they are compilations, abstracts or reviews. 
 
 3. GEOGRAPHICAL OCCURRENCE 
 
 From the foregoing and other papers and from personal observations, 3 
 it appears that the disease has a very general distribution throughout the 
 tropical regions of the world. With a few notable exceptions, probably 
 due to climatic conditions, its range may be said to coincide with that of 
 the cultivated banana. In the West Indies, it has been reported from 
 Jamaica, Cuba, Porto Rico, Trinidad, and Barbados (9). The writer has 
 made personal observations and comparisons of diseased plants and of the 
 associated organism in the first three of these countries, and there is no 
 reasonable uncertainty as to the identity of the disease. No doubt, this 
 disease is present in other islands of the West Indies but on account of the 
 limited extent of the banana industry and lack of agricultural investi- 
 gations, it has not yet been reported from them. 
 
 In Central America, the disease is present in Costa Rica, Nicaragua, 
 Guatemala, Honduras, and British Honduras. The writer found the 
 disease in a most virulent form in Costa Rica (1917) and received cul- 
 tures of the organism from Guatemala and Honduras from Mr. Waters 
 of the United Fruit Company. 
 
 In North America, banana wilt has been reported from Mexico (24). 
 
 In South America, the disease has long been known. It first attracted 
 wide attention in Panama early in the present century. No evidence 
 that the disease has ever existed in Colombia, one of the greatest banana 
 producing countries, can be found. In May and June, 1917, the writer 
 made a diligent search for cases of the disease there but without success. 
 This matter will be discussed later. The destructive nature and wide- 
 spread occurrence of the disease in Surinam has already been noted. 
 
 3 At the invitation of Dr. Erwin F. Smith, the writer made a trip to Cuba, Panama, 
 Costa Rica, Colombia, and Jamaica in 1917 for the purpose of investigating this dis- 
 ease, on funds furnished by the Laboratory of Plant Pathology. 
 
348 Phytopathology [Vol. 9 
 
 Evidence that it exists in India is fairly conclusive, and probably it is 
 present in Australia and the Dutch East Indies as well. 
 
 On account of the great antiquity of the cultivated banana, and the 
 scant records or lack of records concerned with its diseases, it would be 
 difficult to trace the progress of the disease. That it was introduced into 
 the Hawaiian Islands from Costa Rica seems certain, since the outbreak 
 of disease there quickly followed the first importation of Gros Michel 
 bananas from the latter into the former. The promiscuous shipping of 
 stock from a common source into the various countries of Central and 
 South America and the West Indies by large, commercial interests having 
 holdings in all of those countries, has been responsible for its wide distribu- 
 tion in tropical America. The intensive cultivation of bananas on the 
 same land year in and year out has resulted in the disease becoming 
 steadily worse. Where bananas are grown in a desultory way for home 
 consumption the disease may be present but is never serious. 
 
 4. ECONOMIC IMPORTANCE 
 
 a. Extent of the losses 
 
 As early as 1910, banana wilt was regarded as one of the most im- 
 portant, if not the most important, fungus disease of cultivated plants 
 in the American tropics. The extent of the industry which it threatened, 
 its wide distribution, and apparently infectious nature alarmed banana 
 planters almost to the same extent as did the dreaded Phylloxera the vine- 
 yardists of France some years ago. Although endemic, the outbreak in 
 Panama in 1904 assumed the proportions of an epiphytotic. McKenney 
 states (23) in 1910 that in Panama at least 15,000 to 20,000 acres of 
 banana plantations had been abandoned and many thousands more were 
 seriously affected, while in Costa Rica the damage had been even greater. 
 Since that time the disease has continued to spread rapidly. In 1917 the 
 writer visited the banana districts of Panama and found the older sections 
 to be full of abandoned farms. In the Changuinola division, there is 
 one tract of about 15,000 acres where the Gros Michel banana cannot be 
 grown at all, which according to an official of the company had been their 
 finest plantation five years previously. Another area of the same size 
 was found to be practically abandoned, and the newer Sixaola division of 
 45,000 acres was already badly diseased. It has become the policy of the 
 company to replant the abandoned areas with cacao and coconuts. Some- 
 times the Red banana, which is a little more resistant than the Gros 
 Michel, is planted, and several harvests made before it in turn falls prey 
 to the ravages of the parasite. In Costa Rica in 1917 conditions were 
 little better. The newly opened farms show only an occasional dis- 
 
1919] Brandes: Banana Wilt 349 
 
 eased plant, but it is only a matter of time when they too will become 
 badly diseased. The seriousness of the situation is fully appreciated 
 by the banana planters, and efforts have been made to ascertain the cause 
 of the trouble and if possible to find a remedy. The United Fruit Com- 
 pany has established a small experiment station at Zent, Costa Rica 
 where research is carried on in connection with this disease. It would be 
 difficult to estimate the money loss in these two countries during the past 
 decade, but it certainly amounts to many millions of dollars. 
 
 The history of the banana industry in the Dutch colony of Surinam is 
 well-known to the student of tropical agriculture. In 1906, by' agreement 
 with the government the United Fruit Company supplied bulbs of the 
 Gros Michel variety to the estates, and bound themselves to carry the 
 product to the markets. Some cases of banana wilt were observed the 
 next year, 1907, and by 1908 it had already done great damage. Still, 
 in 1909, 648,636 bunches were exported from the colony, showing that the 
 planters were determined to establish the industry. By 1910 every field 
 was affected, and the attempt to grow the Gros Michel variety was aban- 
 doned entirely. Another variety, the Congo, was substituted but it did 
 not meet the requirements of the market and it too was abandoned. A 
 large majority of the planters, who had received advances from the 
 government on the prospect of their crops, were ruined. 
 
 Last year, in the western end of Cuba, there was scarcely a plantation 
 of Manzana bananas which did not show some signs of the disease, and 
 many fields were seen in which every plant was affected. This variety 
 is the most popular in Havana markets, but is not exported. 
 
 In recent years, the Chamaluco variety in Porto Rico has suffered 
 severely from banana wilt. The plantation on the Experiment Station 
 Grounds at Mayaguez is a veritable hotbed of infection. The suckers 
 which come up from old stools appear healthy at first but they rarely 
 bear a bunch of fruit. This variety, which is always cooked, has been in the 
 past one of the main articles in the diet of the average Porto Rican. 
 
 Jamaica has not suffered any great direct loss on account of this dis- 
 ease, due in part to the vigilance of the government, which has estab- 
 lished a quarantine of all affected areas, but also undoubtedly to 
 its climate and soil which are quite different from those in Central 
 America where the disease is rampant. 
 
 b. Nature of the losses 
 
 The nature of the losses due to banana wilt may be divided into three 
 classes (1) injury to the fruit, (2) destruction of immature plants, and (3) 
 depreciation of the value of the land. The first of these is the least serious 
 
350 Phytopathology [Vol. 9 
 
 since affected plants do not usually bear a bunch. If, however, a bunch 
 is "shot," it is apt to be small, the individual fruits are "bottle-necked," 
 the flesh has a yellowish tinge and is pithy. Such a bunch would ripen 
 unevenly and too rapidly and is always rejected by the inspectors on that 
 account. It is a total loss. Destruction of the immature plants is the 
 most serious phase of the disease. The plants wilt and die in enormous 
 numbers, before any fruit is produced. The plant falls prostrate in a 
 short time, but the rootstock is not immediately killed, and may send up 
 fresh suckers. These will invariably become diseased also, and only 
 rarely produce a marketable bunch. 
 
 The third type of loss, viz. depreciation in value of the land, is due 
 to the fact that the causal organism may remain for long periods in the 
 soil, and finally increase to such an extent that bananas can no longer be 
 grown. Since in banana growing countries, the greatest profit is derived 
 from these plants, and sometimes there is absolutely no sale for other 
 products, it is evident that the land is greatly reduced in value or made 
 utterly worthless. 
 
 5. SYMPTOMS 
 
 a. External signs of the disease 
 
 In a field where only occasional and isolated cases of the disease are 
 found, indicating that the pathogene has not yet become thoroughly 
 and uniformly distributed in the soil, the affected plants are not apt to 
 exhibit any external signs of the disease until they are approaching ma- 
 turity. Sometimes the disease is not made manifest until after the bunch 
 has started to form. It is not uncommon to see a banana tree with a 
 half formed bunch of fruit (plate XXIV, fig. 1), whose development has 
 been arrested at that stage by the death of the plant. In these large 
 plants, the symptoms are not unlike those brought about by severe 
 drought. A typically diseased plant first shows a yellowing of the lower 
 or outer leaf blades and petioles (plate XXI). The transition from the 
 normal dark green color of the leaf to a vivid yellow is usually sudden 
 and startling, and proceeds from the margin inward. Such leaves stand 
 out conspicuously in a planting of bananas, the contrast with the healthy 
 leaves making it easy to detect them even at a considerable distance. 
 Their appearance, to one familiar with the disease is umnistakable, and 
 they are not apt to be confused with any other trouble, if, as is usually the 
 case, they appear when drought symptoms are not to be expected. A 
 field was seen in Jamaica in 1917, in which practically all of the plants 
 exhibited this symptom, but on examination no further evidence of disease 
 was found. It was raining every clay at the time, so drought was out of 
 
1919] Brandes: Banana Wilt 351 
 
 the question. Mr. Cousins, director of agriculture, of the Jamaica De- 
 partment of Agriculture said that the particular field had the same appear- 
 ance every year and suggested that it was owing to the character of the 
 soil. Such a condition must be exceedingly rare, since it is the only time 
 it has come under observation in three years study of the disease. The 
 yellowing of the lower, or outer leaves then is a practically unmistakable 
 symptom of the disease. There is no risk of confusing it with the normal 
 sloughing off of the lower leaves. These lower or older leaves gradually 
 die off in a healthy plant, due to abandonment of the roots which supply 
 the lower leaves. This natural phenomenon is characterized by the 
 regularity with which the leaves die in sequence from below upward, and 
 by the fact that at no stage is there a brilliant yellow color of the leaf, 
 but only a gradual transition from dark green to brown. 
 
 Cases of banana wilt have been seen in which a younger leaf, possibly 
 second or third from the bottom, was yellow, while those below it remained 
 green, but usually the oldest (lowest) leaf is the first to go, then, in order, 
 the next oldest and so on up to the topmost (youngest) leaf, which 
 invariably remains green the longest. 
 
 The first attacked leaves begin to wilt almost immediately. Within a 
 day or two the fleshy leaf stalk buckles at a point three or four inches 
 from the pseudostem, and the leaf hangs pendant from this point. Some- 
 times buckling of the leaf-stalk or the large midrib takes place at any point 
 out to the middle of the leaf blade or beyond, but the first mentioned 
 condition is typical. 
 
 The leaf now rapidly withers and becomes brown (plate XXI). The 
 process is rapidly repeated in the other leaves until the topmost or inner- 
 most (youngest) leaf is reached. If the fruit bud has not yet emerged, 
 this leaf may be only partially or not at all unfurled. It is in a state of 
 rapid growth and seems to resist a trifle longer than the other leaves. At 
 any rate, diseased plants are very numerous, in which the youngest leaf 
 stands erect and turgid, like a green lightning rod, for some time after the 
 other leaves have succumbed. Finally, this leaf droops and withers, and 
 the plant stands for a few days or weeks with the dry, brown leaves dang- 
 ling and rattling in the wind (plate XXIV, fig. 2). A puff of wind, 
 stronger than the rest, eventually sends the stately plant crashing to the 
 earth, where it lies prostrate and quickly rots, due to secondary invasion 
 by putrefactive organisms. 
 
 Another symptom of the malady, which is especially prominent in very 
 heavily infected young plants, such as may be seen in abandoned fields 
 where the disease is of long standing, is a decided dwarfing, or stunting 
 of the entire plant. This dwarfing may range from a slight retardation 
 in development, to the extreme condition, observed in the western end of 
 
352 Phytopathology [Vol. 9 
 
 Cuba, where the plants do not grow to be more than one or two feet tall. 
 The pseudostems of these diminutive plants have a constricted, or "hide- 
 bound" appearance and the margin of the leaves may be undulating or 
 the whole leaf curved or distorted in various ways. The leaves do not 
 wilt so rapidly as those of larger plants, nor is the yellow color so con- 
 spicuous. Exactly the same signs were seen in plants growing in soil 
 artificially inoculated heavily with the causal organism at Mayaguez, 
 P. R. 
 
 As stated above, this dwarfing may be less decided, and this is more 
 often the case. The plants may attain a third or a half of their normal 
 size in the time required for complete development. This condition means 
 that the plants arose from diseased bulbs or were strongly infected imme- 
 diately after planting. The leaves become yellow and wilt just as do the 
 leaves of larger plants. 
 
 Another symptom that frequently accompanies stunting, is a longitu- 
 dinal splitting of the outer leaf bases (plate XXV, figs. 1 and 2, and plate 
 XXVI, fig. 1) which form the pseudo-stems. The split may extend all the 
 way from the rhizome, at the base or surface of the ground, to the collar 
 or place where the leaves diverge, or it may be more limited in extent. 
 Only the outer leaf bases may be involved, or the opening may extend to the 
 center of the pseudo-stem, in which case it may happen that the young 
 leaf becomes diverted from its course up through the center, and pro- 
 jects out through the split. This condition also has been produced in 
 artificially inoculated plants (plate XXVI, fig. 2). 
 
 All of the foregoing symptoms have been observed in the varieties 
 Chamaluco, Manzana, Gros Michel and Red, in Porto Rico, Cuba, Costa 
 Rica and Panama respectively and in the Gros Michel in Jamaica. From 
 the literature (10) it is learned that they occur in Surinam also. 
 
 If a bunch of fruit has formed on an attacked plant, it also shows signs 
 of the disease. Since, in the main, the differences between the various 
 varieties are most marked in the characters of the fruit, the bunches of the 
 several susceptible varieties exhibit symptoms that vary somewhat. In 
 general, however, it may be said that the bunch will be found to be small. 
 Development may be completely arrested after a few hands have been 
 formed, or if a normal number of hands are produced, the individual 
 fingers are small and "bottle-necked" • i.e. the ovaiy is constricted or 
 " pinched-in" at the calyx end. The fingers do not ripen uniformly. 
 Occasional fingers, scattered throughout the bunch become yellow rapidly. 
 The flesh is inclined to be pithy, acrid and yellowish and judging from the 
 taste, the starch is not converted into the soluble sugars. Since a bunch 
 is rarely produced, this symptom is of little value in diagnosis. 
 
1919] Brandes: Banana Wilt 353 
 
 b. Internal signs of the disease 
 
 Healthy banana tissue, both in the rhizome and pseudostem is almost 
 dead white when first cut open (plate XXII, figs. 1 and 2). After a few 
 minutes, especially if it has been cut with a steel knife, a purplish dis- 
 coloration will appear uniformly distributed over the cut surface, due to 
 the presence of oxidizing enzymes. If a plant in the incipient stages of 
 banana wilt be removed from the ground and a transverse cut made in 
 the lower part of the rhizome, it presents a quite different appearance, 
 (plate XXIII, fig. 2.) Within the stele, small dots and irregular threads 
 of a yellowish or very light brown color are seen, either distributed evenly 
 over the whole cut surface, or more frequently arranged in a band just 
 inside of the endodermis. Sometimes they are localized in one or more 
 patches just within the endodermis. Such a patch at a more advanced 
 stage is shown in plate XXIX, figure 2. Upon examination these dots 
 and lines are made out to be the discolored vascular bundles, which in 
 this situation run in every conceivable direction, so that there may be 
 transverse, longitudinal, and oblique sections of bundles in the same 
 plane. Upon cutting successive sections of the rhizome towards the apex, 
 it is found that the discoloration gradually becomes less pronounced in an 
 upward direction and finally disappears altogether. If a plant in a some- 
 what later stage of the disease is selected and the operation repeated, it 
 will be found that the bundles in the lower part of the rhizome are more 
 deeply stained, and more numerous, the color ranging from reddish to 
 reddish brown. A few scattered discolored bundles may be seen in the 
 cortex. Successive sections cut at intervals towards the apex reveals 
 discolored bundles in the stele extending upward towards the base of 
 the pseudo-stem. At various points near the apex of the stele, some of 
 these discolored bundles are seen to pass through the endodermis, and 
 traverse the cortex in an upward and outward direction, passing thence 
 directly upward into the leaf-bases. During their passage through the 
 cortex where they form the leaf-traces, these discolored bundles are very 
 conspicuous in the otherwise healthy and white cortical tissue. The 
 previously mentioned scanty diseased bundles in the cortex represent these 
 leaf traces. The most successful method of following their course is to 
 cut a diseased plant in half longitudinally (plate XXVII, fig. 1). Some of 
 them will be revealed, and by carefully excavating the surrounding tissue 
 with a sharp scalpel, one of them can be followed continuously from deep 
 in the stele, across the cortex and far up into the leaf petioles (plate 
 XXVII, fig. 2). Here the color gradually becomes lighter as they progress 
 towards the collar and eventually, in this stage of the disease, the dis- 
 coloration is no longer found. 
 
354 Phytopathology [Vol. 9 
 
 In such a plant, if the pseudo-stem is cut off transversely near the 
 base, only cross sections of discolored vascular bundles will be found 
 (plate XXIII, fig. 1). The bundles here are straight and vertical, and 
 are arranged in concentric arcs, conforming to the clasping leaf bases of 
 which the pseudo-stem is composed. 
 
 At a quite advanced stage of the disease, say when most of the leaves 
 are wilted, a cross section of the rhizome shows the vascular bundles of 
 the stele to have changed in color, from reddish brown to purple or even 
 black and to be so numerous that most of the stelar tissue, especially 
 that near the periphery, is quite dark in color (plate XXVIII, fig. 1). 
 The cortex is still white and healthy looking excepting for the few scat- 
 tered discolored bundles forming the leaf-traces. Somewhat later the 
 entire stele becomes entirely black (plate XXVIII, fig. 2). Secondary 
 rots have set in by this time and it is more difficult to isolate the causal 
 organism from such tissue. As a result of such rots, which are occa- 
 sionally putrefactive, a disagreeable odor is sometimes given off. It has 
 been noted also that a yellowish or brownish discoloration of part or all 
 of the stelar tissues takes place when putrefaction sets in but this is of 
 infrequent occurrence. A cross section of the pseudo-stem now shows 
 reddish, purplish or black vascular bundles at any point up to the 
 crown, and even in the leaf -stalk or midrib. 
 
 At this point it may be mentioned that certain more or less uniform 
 differences exist in the appearance of the diseased pseudo-stem cross 
 sections in the several attacked varieties. In all of them, the leaf bases 
 at the center of the pseudo-stem remain healthy after the bundles of the 
 outer ones have become discolored. This is to be expected when we re- 
 call that under " gross appearance, etc." it was mentioned that the prog- 
 ress of the disease is from the lower (outer) leaves to the upper (inner) 
 ones. In the Gros Michel, this phenomenon is not so conspicuous, that is, 
 there is a more even distribution of the discolored bundles throughout the 
 cut surface of the pseudo-stem (text fig. 1). 
 
 With the Manzana variety a more pronounced delay of invasion of the 
 inner leaves is seen. The condition in the Chamaluco is still more com- 
 plicated, for in typical cases the cross section shows a decided band of 
 diseased tissue, with healthy tissue both within and without (plate XXIX, 
 fig. 1). Occasional discolored bundles may be found, it is true, on both 
 sides of the band, but the tissues there are decidedly less attacked. 
 
 Passing to the roots, which of course originate at the stele, and pass 
 thence through the thick, fleshy cortex and so into the soil, symptoms of 
 the disease are exhibited in various ways. The main roots are fleshy 
 and of uniform diameter throughout. From these small thread-like roots 
 arise laterally, and branch profusely. On the thread-like roots and back 
 
1919] 
 
 Brandes: Banana Wilt 
 
 355 
 
 of the growing tip of the main roots are borne the root hairs. The pri- 
 mary meristem occupies only a very small portion behind the root cap, 
 and is very tender. The main roots do not seem to adapt themselves 
 well to irregularities in the soil, but upon meeting a stone or other ob- 
 struction, the tips by rapid cell division and growth push into it and are 
 crushed. The root dies back and a lateral thread-like root behind the 
 apex takes the lead, and becomes the main root. It is not surprising that 
 many roots are seen which are dead and decayed. Unless complications 
 have set in, this must not be taken to indicate a diseased condition. 
 
 Fig. 1. Transverse Section of Psettdostem of Gros Michel Variety in 
 
 Jamaica 
 
 Notice the rather even distribution of discolored vascular bundles 
 
 Nematodes, insects and other agencies also may injure the roots. How- 
 ever, umnistakable evidence of disease in the roots may be found in con- 
 nection with wilt. Blackened roots, close to the bulb and extending into 
 diseased portions of the stele are frequently found (plate XXVII, fig. 1) 
 and this condition has been proved to be due to the banana wilt organism 
 (see discussion under "life history"). 
 
 It may be proper to discuss here some variations in the situation of 
 primary diseased stelar tissue. An explanation for the condition will be 
 reserved for discussion later. In dissecting hundreds of diseased plants, 
 it is seen that there are two regions where the disease apparently first gains 
 
356 Phytopathology [Vol. 9 
 
 a foothold: (1) at the cut surface of the rhizome, where the sucker was cut 
 away from the mother plant resulting in general infection (plate XXVIII, 
 fig. 1) and (2) at the side of the bulb, where no wound is present (plate 
 XXIX, fig. 2). The evidence for this is deduced from the fact that the 
 diseased tissue in the early stages of the malady is confined to either or 
 both of those regions, and in more advanced stages the tissue in those 
 regions is manifestly more strongly attacked than elsewhere, as can be 
 readily seen by a glance at the illustration. Where the diseased stelar 
 tissue originates at the side of the bulb, a diseased root can invariably 
 be found leading out from it into the soil. 
 
 6. ETIOLOGY 
 
 a. Morphology of the causal organism 
 
 The causal organism is Fusarium cubense E. F. Smith amended E. W. 
 Brandes. 
 
 1. Sporodochia and conidia. Sporodochia are found at the surface of 
 leaf stalks, and blades, sometimes also of leaf bases, emerging through 
 the stomatal openings (plate XXXIV, fig. 2). They are most numerous 
 on the upper epidermis of leaf stalks at the point where they diverge from 
 the pseudo-stem. On account of their location in stomata, they natu- 
 rally stand separate, and are not joined by any type of stroma. The 
 sporodochium arises from a globose mass of pseudoparenchymatous tissue 
 26 to 30 /J. in diameter which entirely occupies and distends the sub- 
 stomatal cavity. This pseudoparenchymatous tissue is composed of quite 
 large, thin walled, isodiametric cells. The cells vary from 2.6 n to 10.4 n 
 in diameter. This fungus tissue wedges the guard cells apart at the 
 aperture to an average width of 21 /jl. At this point the cells are no longer 
 isodiametric, but are elongated in the direction of outward growth, and 
 are united in filaments. Occasional large globose, bladder-like cells, up to 
 10 [x in diameter are found intercalated in the filaments or attached termi- 
 nally, the tissue here being somewhat vesiculose. The individual conid- 
 iophores arise just where the fruiting body emerges from the stomata, 
 and diverge from one another so that the somewhat loose aerial tissue 
 flares out in all directions and is roughly obovate. Conidiophores are 
 verticillately branched, with two and occasionally three one-celled 
 branches in a whorl. The average length of conidiophores is 70 ii, and 
 of the lateral branches about 14 p. The largest diameter of both main 
 stalk and branches is 4 /x. The apical ends of the lateral branches taper 
 abruptly, but the tip of the main axis is drawn out into a slender needle- 
 like end. The main axis is about seven times septate, and the whorls 
 of branches arise at the extreme upper end of cells of the main axis. 
 
1919] Brandes: Banana Wilt 357 
 
 Usually three whorls of lateral branches are found on a conidiophore. 
 Coniclia are borne at the apical ends of both the lateral branches and the 
 main stalk. These are of two distinct types. At first 0- and 1-septate 
 microconidia are born in great abundance. They are ovate, or somewhat 
 elongated, and range from 5 to 7 ^ by 2.5 to 3 n. A few abortive, 1- and 
 2-septate sickle-shaped conidia are found interspersed among the micro- 
 conidia. These vary greatly in size, approaching the microconidia on the 
 one hand and normal macroconidia on the other. Finally, the large, 
 hyaline pedicillate, sickle-shaped macroconidia are produced (plate 
 XXXI, fig. 1). More than 95 per cent of them are 3-septate, a few 4- 
 and 5-septate individuals being found. They vary from 22 to 36 n in 
 length, and from 4 to 5 n in their largest diameter. The mature sporo- 
 dochia are dry and powdery. They appear white by reflected light. On 
 account of their minute size and the rather restricted area upon which 
 they occur normally on the host, it is not surprising that they have never 
 before been reported. Pure cultures obtained from these sporodochia by 
 the loop dilution method do not differ from those obtained by plating out 
 diseased internal tissue. 
 
 2. Development of sporodochia. Tangential, radial and transverse serial 
 sections of affected leaf stalks, stained with safranin and haematoxylin, 
 show sporodochia in all stages of development. They arise from intra- 
 cellular mycelium in the epidermal and subepidermal cells of the host 
 (plate XXXIV, fig. 1). In the vicinity of the substomatal cavity the 
 mycelium possesses occasional bladder-like cells, which are terminal or 
 intercalated. Such cells are not seen in the vegetative forms of the fun- 
 gus in banana tissue. This mycelium penetrates the substomatal cavity, 
 where it loses its filamentous character altogether. The first cells to 
 enter the cavity are large and bulbous. They divide rapidly, producing 
 cells of a like nature, until the whole cavity is distended with pseudo- 
 parenchyma. The growth is very irregular, and there is no suggestion 
 of chains of cells. At first this mass of fungus tissue conforms to the 
 original shape of the cavity, but by the pressure of its growth, exerted 
 in all directions, it soon becomes globose. Before this portion of the sporo- 
 dochium has attained its maximum size, strands of hyphse push out 
 through the stomatal aperture. Successive outgrowths result in a 
 wedging apart of the guard cells in a short interval of time. These same 
 strands of hyphse continue to grow apically, by cutting off end cells. Just 
 subsequent to passing through the aperture, they diverge, and become 
 the conidiophores. 
 
 3. Mycelium. The mycelium is intracellular. Evidence of intercellular 
 mycelium has been seen in a few sections, especially near the point of 
 penetration in roots and stele, but this type seems to be rare. In the cor- 
 
•358 Phytopathology [Vol. 9 
 
 tical cells of the roots, and also in the xylem, where they occupy the 
 lumena of vessels, the strands are hyaline and septate, the septa being 
 spaced at intervals of about 14 ju on the average. There are here no con- 
 strictions at the septa, the tubes being about 2.6 p. wide and of uniform 
 diameter throughout. The contents are densely granular. In these 
 regions the mycelium takes an irregular course, but in the vessels of the 
 leaf bases it is more commonly straight. In the parenchymatous cells of 
 the leaf-stalk and blade, just previous to sporodochium formation, the 
 fungus cells are swollen and barrel-shaped. The hyphee are more richly 
 septate, with decided constrictions at the septa. Occasionally an enor- 
 mously distended cell will be found, perhaps 8-9 /x in diameter. The 
 mycelium is in general a trifle larger here than elsewhere, averaging 3 to 
 4 ijl in diameter. The large bladder-like cells occur terminally or less 
 frequently intercalated between other cells of the hyphal strand. Oil 
 drops and other inclusions are more noticeable in mycelium of these 
 regions. 
 
 4- Cultural characteristics of the causal organism. It is not proposed 
 to go deeply into the vexed question of separating the species and varie- 
 ties of Fusarium on their morphological and physiological characters in 
 pure culture. The limitations of such attempts have already become 
 evident. We do not believe that it is possible at present to separate 
 Fusarium cubense from nearly related forms on this basis. Possibly 
 methods may be devised in the future by which it can be accomplished. 
 The following data will serve mainly to show the near relationship between 
 Fusarium cubense and other species, notably F. vasinfectum. This organ- 
 ism has been under continuous observation in pure culture for the past 
 three years. The study has served to show that the same strain, origi- 
 nally from a single-spore isolation will vary considerably with the age of 
 the culture or the kind of inoculum used (mycelium, conidia, sclerotia). 
 Even two sub-cultures will vary somewhat under apparently identical 
 conditions. The following descriptions are of freshly isolated cultures, 
 material that seems to vary the least. 
 
 The organism grows rapidly and luxuriantly on potato plugs, producing 
 dead white, aerial mycelium in great abundance. Microconidia are formed 
 in enormous numbers in about two days, being abstricted at the tips of 
 the short lateral branches of conidiophores. They are one-celled, and 
 variable in size, ranging from 5 to 7/z by 2.5 to 3 n in size. Many of them 
 accumulate at the fertile tips of hyphse, being held in globose heads by a 
 gelatinous or watery matrix. Occasional normal macroconidia may be 
 found in such a culture, but they are very rare. Transition forms between 
 the microconidia and macroconidia are numerous. They are 0- 1- and 
 2-septate and may be straight or curved. In five or six days, small masses 
 of hard, flesh-colored, plectenchymatic tissue develop at the surface of 
 
1919] Braxdes: Banana Wilt 359 
 
 the substratum. These are smaller than a pin head when first seen. 
 They develop rather slowly and become indigo blue in color when ten 
 to twelve days old. These sclerotia range from one to four mm. in 
 diameter and are irregular and nodule-like. 
 
 On potato agar slants, a more spreading growth takes place. The 
 mycelium penetrates deeply into the medium and less aerial growth takes 
 place, although it is considerable sometimes. Cultures may make a low 
 slimy growth with only a little aerial mycelium. Besides the types of 
 conidia found on potato plug cultures, so-called " sporodochia" are formed 
 on various agar media if the air is moist. These are wart-like heaps or 
 limited, more or less convex layers of macroconidia borne on closely packed 
 conidiophores. They are salmon colored and range from 2 to 5 mm. in 
 diameter. When the culture has become somewhat dry, chlamydospores 
 are produced, arising from cells of the mycelium or macrospores. They 
 occur singly, in pairs, or in chains, and may be terminal or intercalary. 
 They become thick-walled when old. Individual chlamydospores are 
 globose to short oval, and in size are 5.5 to 6 m by 6 to 7 /x. Paired or 
 catenulate chlamydospores are usually a little smaller. 
 
 On steamed rice grains, a luxuriant growth of white mycelium is made 
 which penetrates to the bottom of the test tube. Very soon the sub- 
 stratum becomes colored a delicate Hermosa pink (Ridgway). The color 
 gradually deepens, but usually not uniformly, but rather is streaked or 
 blotched and varies from Hermosa pink to spectrum red. Vinaceous 
 tinges occur later in all cultures and some of them become violet purple. 
 These color changes have been observed in cultures from all varieties of 
 bananas secured in Porto Rico, Cuba, Panama, Costa Rica, Guatemala, 
 Honduras and Jamaica. When old and dry, the white mycelium of cul- 
 tures obscures the color of the substratum. Small bits of white to flesh- 
 colored plectenchyma are seen scattered through the culture and adhering 
 firmly to the glass walls of the test tubes. These have not been observed 
 to become indigo blue, as do similar bodies in potato slab cultures. Mi- 
 croconidia are formed on rice, but no macroconidia. A quite noticeable 
 agreeable odor is generated by some strains of F. cubense when grown on 
 this medium. When requests have been made of various people to 
 identify the odor, it has been described as similar to the odor of "orange 
 peel," "ripe watermelon," "cucumbers," "fermented grapes," "lilacs," etc. 
 
 On stems of Melilotus alba, the fungus first makes an abundant growth of 
 white aerial mycelium. Within a few days, however, a nearly continuous 
 slimy layer of macroconidia. will form on a considerable area of the stem. 
 This layer may have a smooth surface, or may be lumpy or nodule-like, 
 the irregularities of topography being caused by the more abimdant pro- 
 duction of conidia at certain points. This type of fructification has been 
 referred to as a "pionnotes." The spores are salmon colored in mass. 
 
360 Phytopathology [Vol. 9 
 
 b. Physiology of the causal organism 
 
 1.. Isolation. The organism is easily isolated in various ways. The 
 chief requisite, is the use of tissue in the incipient stages of disease. For 
 rapid work, such tissue cut from the root, rhizome, leaf-base, or leaf 
 stalk under aseptic conditions, and planted in sterile media will give a 
 pure culture of the parasite. If the disease is far advanced, bacterial and 
 even fungus contaminations are likely to occur. If only advanced cases 
 are available, bacterial growth may be inhibited by adding to the media 
 2 cc. of a 10 per cent solution of lactic acid. 4 
 
 For refined work it is necessary to obtain single-spore strains of the 
 organism. Various methods of doing this successfully are available. 
 Tissue such as would be used for direct plantings may be removed with 
 the same precautions and crushed in sterile melted agar, after which it is 
 poured into petri dishes. Many colonies will arise from microconidia 
 present in the vessels. If clear media and thin-bottomed petri dishes are 
 used, the conidia may be located with the microscope through the bottom 
 of the plate and marked with India ink. A colony resulting from the 
 germination and growth of a single conidium can be lifted from the plate 
 with a sterile platinum spatula, and transferred to fresh sterile media. 
 
 The single-spore strain is thus obtained directly. Conidia from a 
 colony resulting from direct planting may be stirred into sterile melted 
 agar, after which point the above mentioned process is followed. 
 
 Conidia borne externally on sporodochia can be scraped off with a 
 sterile scalpel, stirred into melted acid agar, and plates poured. Single 
 conidia are easily located in the same manner and the resulting colonies 
 transferred. 
 
 Isolation of the organism from the soil is accomplished by taking a 
 small particle of soil from below the surface, at any depth from one to 
 thirty inches, under aseptic conditions (30) and transferring it to acid 
 agar or steamed rice tubes. By this method of direct planting, contami- 
 nations may occur, but in heavily infested soil, the pathogene usually 
 outgrows the other organisms present. Single spore strains are obtained 
 as above from these cultures. If it is desired to determine the number of 
 conidia in the soil, a measured quantity of soil is taken and dilution plates 
 poured. Media for this purpose should be acid to discourage bacteria, 
 and starchy, so that the pathogene may be quickly recognized by its 
 characteristic color production. 
 
 It is realized that the behavior of a fungus in pure culture under con- 
 trolled conditions is at best only an indication of what may be expected 
 
 4 This amount of acid in 10 cc. of potato agar will permit of a somewhat re- 
 stricted growth of the Fusarium and will prevent the development of the common 
 bacteria. 
 
1919] 
 
 Brandes: Banana Wilt 
 
 361 
 
 under natural conditions. Insofar as deductions to be drawn from the 
 relation of the organism to moisture, heat, light, etc. are concerned, if 
 made with the end in view of offering explanations for the various phe- 
 nomena exhibited by the organism during its parasitic existence in the 
 host or during saprogenesis in the soil, they are of limited applicability 
 and should be accepted with reserve. However, knowledge of this kind 
 is of some service in this connection, and to a certain extent may also be 
 of some value in delimiting the organism concerned from some of its close 
 allies. A few experiments have been conducted with Fusarium cubense to 
 ascertain its response to varied conditions of heat, moisture, light and 
 oxygen supply. 
 
 2. Relation to heat. The thermal death point of macroconidia obtained 
 from fresh pionnotes on Melilotus alba stems was determined by the method 
 in use by bacteriologists. A spore suspension was made in sterile dis- 
 tilled water, and drawn up into capillary glass tubes six inches long. 
 These were sealed by holding the ends in the flame of a Bunsen burner 
 for a few seconds. The tubes with included spores were then exposed for 
 ten minutes to constant temperatures in the water bath, ranging from 40 
 to 55°C, the series varying by intervals of one degree. The water in 
 the water bath was maintained at any desired temperature by the use of 
 Novy's thermo-regulator, and was continuously agitated to obviate the 
 inequality of temperature at various points due to convection currents. 
 After the exposure, one end of the capillary tube was broken off, and the 
 contents "shot" into a dish of sterile agar by applying a flame to the 
 closed end. Germination of the conidia was observed through the bot- 
 tom of the petri dish. The average result of three trials is as follows. 
 
 
 PER 
 
 CENT OF GERMINATION AFTER 
 
 TEMPERATURE 
 
 
 
 
 
 12 hours 
 
 24 hours 
 
 48 hours 
 
 °C. 
 
 
 
 
 42 
 
 99+ 
 
 99+ 
 
 99+ 
 
 43 
 
 99+ 
 
 99+ 
 
 99+ 
 
 44 
 
 99+ 
 
 99+ 
 
 99+ 
 
 45 
 
 90 
 
 94 
 
 • 95 
 
 46 
 
 40 
 
 50 
 
 50 
 
 47 
 
 15 
 
 40 
 
 40 
 
 48 
 
 15 
 
 ' 30 
 
 30 
 
 49 
 
 S 
 
 30 
 
 30 
 
 50 
 
 5 
 
 30 
 
 30 
 
 51 
 
 
 
 20 
 
 20 
 
 52 
 
 
 
 20 
 
 20 
 
 53 
 
 
 
 15 
 
 15 
 
 54 
 
 
 
 1 
 
 1 
 
 55 
 
 
 
 
 1 
 
 56 
 
 
 
 
 
 
 
362 
 
 Phytopathology 
 
 [Vol. 9 
 
 According to this arbitrary method, the thermal death point is 56°C, 
 which is remarkably high. A decided decrease in percentage of germina- 
 tion at temperatures of 46°C. and above is to be observed. At tem- 
 peratures above 50°C. germination is of a type that is decidedly abnormal. 
 The germ-tube is thick and makes a slow, feeble growth. 
 
 Sustained exposure to heat causes death at much lower temperatures. 
 Using the same technique it was found that twenty-four hours exposure 
 to 42°C. killed all of the spores, and only a small percentage germinated 
 after twenty-four hours at 38°C. 
 
 3. Relation to moisture. Freshly produced macroconidia and micro- 
 conidia are very tender to desiccation. An experiment was conducted to 
 determine how long such spores would withstand drying on clean cover 
 glasses at ordinary atmospheric humidity. A suspension of spores in 
 distilled water was made up, using material from pionnotes on Melilotus 
 alba stems, and drops of the suspension were placed on clean sterile cover 
 slips. The water evaporated quickly leaving a cloudy mass of air dry- 
 spores on the glass. The slips were placed away from the dust in a cov- 
 ered culture dish. At intervals of two hours, a slip was removed, and a 
 drop of sterile distilled water placed on the spores. It was then in- 
 verted over a Van Tieghem ring and set away. Most of the conidia failed 
 to germinate in distilled water after drying for a period of four to seven 
 hours. There were still a few feeble germinations at twenty hours but 
 none at twenty-two hours. The result of this experiment is outlined in the 
 following table. 
 
 PERIOD OF DESICCATION 
 
 GERMINATION 
 
 PERIOD OF DESICCATION 
 
 GERMINATION 
 
 hours 
 
 per cent 
 
 hours 
 
 per cent 
 
 
 
 80 
 
 14 
 
 4 
 
 2 
 
 80 
 
 16 
 
 2 
 
 4 
 
 80 
 
 18 
 
 1 
 
 6 
 
 40 
 
 20 
 
 1 
 
 8 
 
 15 
 
 22 
 
 
 
 10 
 
 4 
 
 24 
 
 
 
 12 
 
 1 4 
 
 
 
 Possibly a higher per cent- of germination could be obtained by sub- 
 stituting some nutrient solution for distilled water. Spores from cul- 
 tures that are apparently very dry and have been so for weeks will 
 sometimes germinate readily. 
 
 Chlamydospores and sclerotia are quite resistant to desiccation. They 
 will germinate after being air dry for five or six months, possibly even 
 longer, but no evidence that they will resume growth after six months is 
 
1919] Brandes: Banana Wilt 363 
 
 available. Since these, bodies have not been found in nature, no exact 
 experiments have been performed for the purpose of determining their 
 longevity under dry conditions. 
 
 In old dry cultures on bean pods, chlamydospores are sometimes found 
 to be absent, but the ordinary mycelium, which has become quite thick 
 walled, has been known to resume growth after being quite dry for four 
 months. 
 
 4- Relation to light. Attempts to vary the growth of the organism by 
 controlling the amount of light available for it have had slight success. 
 Light-tight cases were made of black opaque photographic paper in the 
 following way. Two open paper cylinders were made by gluing the 
 opposite edges of sheets together. Both were eleven inches long, but one 
 was 4 inches and the other 6 inches in diameter, so that the smaller would 
 fit inside of the larger. Caps for both ends were made by gluing the 
 rim of a short cylinder, 5 inches in diameter to a round disc of the same 
 material. The cup was so constructed that the attached short cylinder 
 would fit between the two 11-inch cylinders. The rims of the long cylin- 
 ders were toothed to insure circulation of air, and all seams made light- 
 tight with a mixture of lamp-black and paraffine. 
 
 Similar cases were made out of transparent tracing cloth. Newly 
 isolated cultures on potato plugs and steamed rice were placed in all of the 
 cases, and the latter were placed in front of a south window in direct 
 sunlight. All conditions of temperature, humidity and aeration were 
 thus exactly the same, so if light exercised any limiting influence, it would 
 be here determined. 
 
 After ten days the cultures were removed from the cases and exam- 
 ined. All had made an abundant vegetative growth. There was no per- 
 ceptible difference in amount between the cultures in absolute darkness 
 and those in strong diffused light. Macroconidia, microconidia and 
 chlamydospores had been produced in equal numbers in all of the tubes. 
 The sclerotia on potato plugs were identical. The only observed difference 
 was a slight variation in color production in the rice cultures. Those in 
 the light had slightly more prominent streaks of a vinaceous color in the 
 otherwise pink background. 
 
 5. Relation to oxygen supply. To determine the part played by free oxy- 
 gen in the development of the fungus, Erlenmeyer flasks of various sizes 
 ranging in capacity from 10 to 1000 cc. were filled to a uniform depth 
 with corn meal and sterilized in the autoclave. Each flask was then in- 
 oculated with spores from a newly isolated culture by a single stab at the 
 center. When the cotton plugs were replaced they were immediately 
 sealed securely with paraffin. 
 
364 
 
 Phytopathology 
 
 [Vol. 9 
 
 Readings were taken before the edges of colonies in the smallest flasks 
 reached the glass walls of the vessels to avoid the criticism that any 
 restriction of development was due to limited space for growth. The 
 edges of the colonies in the smallest flasks were still 6 to 7 mm. from the 
 walls when readings were made. Results: 
 
 Effect of oxygen on growth of organism,; 
 
 time nine days 
 
 FLASK CAPACITY 
 
 DIAMETER OF COLONY 
 
 CHARACTER OF GROWTH 
 
 CC. 
 
 m m . 
 
 
 10 
 
 15 
 
 Flat effuse 
 
 10 
 
 15 
 
 Flat effuse 
 
 10 
 
 15 
 
 Flat effuse 
 
 100 
 
 41 
 
 Raised 
 
 100 
 
 38 
 
 Raised 
 
 100 
 
 40 
 
 Raised 
 
 250 
 
 55 
 
 Aerial 
 
 250 
 
 58 
 
 Aerial 
 
 250 
 
 59 
 
 Aerial 
 
 500 
 
 75 
 
 Aerial 
 
 500 
 
 65 
 
 Aerial 
 
 500 
 
 72 
 
 Aerial 
 
 1000 
 
 90 
 
 Fluffy aerial 
 
 1000 
 
 70 
 
 Fluffy aerial 
 
 1000 
 
 80 
 
 Fluffy aerial 
 
 Color production in the media was the same in all cases, a deep vina- 
 ceous red. An equal production of conidia took place in all cases. In the 
 absence of any other logical explanation for the limited growth in the 
 small flasks it seems probable that the more rapid growth in the large 
 flasks was due to more abundant oxygen supply. Inhibition due to less 
 chance of wide diffusion of autotoxic excretions seems less reasonable. 
 
 6. Excretions of the fungus. When a culture stands for some time under 
 very humid conditions, large drops of liquid gather in the aerial mycelium. 
 Sometimes this is so extensive that the drops run together and form a 
 continuous film over the surface so that the culture resembles wet fur. 
 This phenomenon has been termed transpiration, and it is believed to be 
 similar in function to transpiration in the higher plants. 
 
 For the most part we have only indirect evidence that excretions are 
 made by mycelium below the surface of the substratum. Several times 
 it has been mentioned that starchy media such as rice, corn meal, etc., 
 are discolored where they are in contact with the mycelium of F. cubense. 
 In the host tissue also various colors are produced. The study of color 
 production, a characteristic which the organism has in common with 
 
1919] Brandes: Banana Wilt 365 
 
 various other fungi, is of general interest and constitutes a separate prob- 
 lem. Undoubtedly it is due to the excretion of substances by the fungus 
 into the substratum. Attempts to vary color production by growing the 
 organism in acid and alkaline media have failed. 
 
 It is known that some strains of F. cubense generate propionic aldehyde 
 when grown in certain synthetic liquid media (19), and the aromatic odor 
 which is detected when the organism is grown in these media is ascribed 
 to this substance. The same odor occurs when the organism is grown on 
 steamed rice. It has been suggested (Lathrop) that this substance may 
 be responsible for the pathologic effect on the host, on account of the 
 deleterious effect of aldehydes on plant growth. This seems improbable 
 however, since there are strains of F. cubense, which are parasites of the 
 same aggressive type, but no odor accompanies their growth on these 
 same media, and probably no aldehydes are produced. 
 
 Experiments have been conducted by the writer, however, that seem 
 to indicate that there is a relation between excretion of toxic substances 
 by the parasite, and the symptoms exhibited by attacked plants. It has 
 been noted (1) that the fungus tissue in the lumena of vessels is not 
 present in quantity sufficient to cause serious obstruction to the passage 
 of water. There are exceptions to this but as a rule it is true. Wilting, 
 then, must be caused by some other means than the mechanical plugging 
 of the lumena of vessels and the resultant restriction of water supply to 
 the leaves. Cultures of F. cubense from Jamaica and Porto Rico were 
 grown in 500 cc. of Richards solution 5 in 1 litre Erlenmeyer flask for two 
 weeks. The fungus made a rapid growth and occupied most of the 
 available space in the liquid at the end of that time. The cultures were 
 then filtered under aseptic conditions and the filtrate was poured into 
 250 cc. flasks, about 200 cc. being put into each. An equal amount of 
 the same solution that had merely been sterilized, but not inoculated was 
 placed in each of another series of 250 cc. flasks. Another series was 
 filled with sterile distilled water. The stems of seeding plants of buck- 
 wheat about ten inches tall were now cut off under water at the surface 
 of the ground, and the tops quickly transferred to the flasks containing 
 the various liquids. One plant was placed in each flask so that the 
 cut ends were deeply immersed. The experiment was performed in a 
 greenhouse at noon March 23, 1917, at Ithaca, New York. A bright sun 
 was shining at the time. Within five minutes indications of wilting 
 could be seen in the plants immersed in the filtrate from cultures. In 
 fifteen minutes the plants were very noticeably wilted, and in one-half 
 
 6 Composition of Richard's solution: 10 grams KN0 3 , 5 grams KHoPO-t, 2.5 grams 
 MgSO-4, 20 mgm. FeCl 3 , 50 grams cane sugar, 1000 cc. water. 
 
366 
 
 Phytopathology 
 
 [Vol. 9 
 
 hour the leaves hung perfectly limp, and even the tips of the stems 
 drooped to one side. The check plants, both in uninoculated Richards 
 solution and in water remained perfectly turgid for 24 hours after the 
 experiment was started. 
 
 The experiment was repeated, using bean seedlings in place of buck- 
 wheat, with the same results (text fig. 2), except that the time required 
 for wilting was somewhat longer, it being fully two hours before the 
 plants were decidedly wilted. 
 
 Banana leaves were now used with the result that it required fully 24 
 hours for the wilting to become perceptible. The banana leaf blade and leaf 
 stalk are composed of leathery and fleshy tissue respectively and it is 
 hardly to be expected that the leaf would wilt as quickly as the rela- 
 
 Fig. 2. Wilting of Bean Plants by Toxic Excretion op Fusarium cubense 
 
 The five plants on the left are immersed in the filtrate from a culture growing in 
 Uschinsky's solution. The next two plants are in plain sterile Uschinsky's solution, 
 and the one on the extreme right is in water. 
 
 tively soft, tender tissues of the other plants. The time required in this 
 case was of course long enough to permit the growth of other organisms 
 in the media. This is unfortunate, but sterilization of the banana leaf 
 stalk cannot be accomplished without injuring it and otherwise inter- 
 fering with the experiment. Results in the other two cases, however, 
 convinces the writer that the wilting here was due to the same causes. 
 
 It will immediately occur to any student that the wilting may have 
 been due to increased osmotic pressure in the filtrates from cultures. It is 
 true that before the cane sugar, a dissaccharide, of the culture solution is 
 available to the fungus, it must be inverted by some excreted enzyme such 
 as invertase, and the resulting monosaccharides, glucose and fructose, would 
 have approximately double the osmotic pressure of the original sugar so- 
 
1919] Brandes: Banana Wilt 367 
 
 lution. It was then decided to use another synthetic culture medium in 
 which the carbon is supplied in some other form and in small amounts. 
 Uschinsky's solution, in which the organic matter consists of small amounts 
 of ammonium lactate and sodium asparaginate was next used, but with 
 precisely the same results. Goss (16) by employing slightly different 
 methods, came to the conclusion that the wilt of potatoes caused by 
 Fusarium oxysporum, is brought about by the action of toxic substances 
 excreted by the fungus. In the case of the banana disease, wilting is not 
 due to plugging of the vessels by mycelium, but is probably the result of 
 toxic excretions by the fungus. 
 
 c. Pathogenicity 
 
 On account of its constant association with the disease, and on ac- 
 count of the evidence put forth as a result of some rather doubtful ex- 
 periments, the ability of F. cubense to cause the disease has been assumed 
 for some time. In previously recorded experiments by Drost (10), in- 
 oculated plants were grown under very unnatural conditions in which 
 normal development could net take place. What is worse, it is evident 
 that cultures used for inoculating the plants were not pure, since "pyc- 
 nidia" and "perithecia" are described as arising from them. These cul- 
 tures were obtained by direct plantings of diseased tissue and it is 
 probable that they were contaminated. 
 
 In the fall of 1915, at a time when the writer was not aware that work 
 had previously been done on this disease, experiments were started at 
 Mayaguez, Porto Rico in an attempt to prove the causal relation of the 
 associated fungus. As a result of those experiments, it was possible to 
 furnish formal proof of the pathogenicity of the parasite, and the results 
 were published (6) in the annual report of the Porto Rico Agricultural 
 Experiment Station for 1916. The principal experiment was conducted as 
 follows: Thirty cylindrical cement tiles, 3 feet in diameter and 4 feet deep 
 were constructed and sunk into the ground in prepared excavations so 
 that the rim projected 4 inches above the surface of the ground. Each 
 tile had a "collar" about 6 inches wide 4 inches below the rim which 
 rested on the surface to prevent settling. The tiles were all filled with a 
 mixture of clay loam and river sand. The soil in twenty of the tiles was 
 sterilized with live steam from a specially constructed apparatus. (Text 
 fig. 3.) This consisted of a ramification of pipes with small perforations 
 in three rows lengthwise of the pipe at intervals of \ inch in the rows. 
 The pipes were so disposed that the soil in no part of the tile was more 
 than 6 inches from a jet of steam issuing at 80 pounds pressure. The 
 pipes were connected by means of steam hose to a boiler capable of gen- 
 
368 
 
 Phytopathology 
 
 [Vol. 9 
 
 crating and discharging large quantities of steam. The whole apparatus 
 was mounted on a stone boat so that it could be moved to any desired 
 point by a pair of oxen. In sterilizing the soil, the rigid series of pipes 
 was driven into the rather light soil, and the steam turned on when it had 
 attained a pressure of 80 pounds. This was maintained for two hours 
 after which the steam was turned off and the pipes removed with sterile 
 wrenches. Twenty minutes after the steam was turned on the soil at 
 
 Fig. 3. Apparatus Used for Sterilizing Soil with Steam 
 
 points farthest removed from the pipes had attained a temperature of 
 120°C. The soil was then covered with discs of tar paper soaked in a 
 strong solution of carbolineum, and the rim was smeared with "tangle- 
 foot" to prevent contamination by insects. It was amply proved by 
 plating out soil from various depths that the soil was actually sterile. 
 When sterilization of the twenty tiles had been completed, the soil in ten 
 of them was inoculated with sub-cultures from a single spore strain of 
 the organism, by emptying the contents of a culture growing in plain 
 corn meal in 1000 cc. Erlenmeyer flasks into each, and stirring in well with 
 
1919] Brandes: Banana Wilt 369 
 
 sterile glass rods. All of the thirty tiles, including the ten untreated ones 
 were now planted with healthy bulbs obtained from Naguabo, Porto 
 Rico, a disease free region. The bulbs were first examined for evidence of 
 disease. None was found, but to guard against extraneous infection, 
 they were all treated ten minutes in 1 to 1000 HgCl 2 and then rinsed in 
 sterile water. 
 
 They were planted by cutting round holes in the middle of the tar 
 paper discs, after which fine mesh wire cages 3 feet high were placed 
 over the tiles. These cages rested on the " collars" of the tiles and fitted 
 closely against the outer part of the projecting rims. The plants were 
 watered with sterile water for several weeks, after which the rains set in, 
 and there was no further need of supplying moisture in this way. In a 
 month it was necessary to remove the cages, due to the growth of the 
 plants. They had served their main purpose, which was to aid in keep- 
 ing out insects and small animals and thus prevent possible contamina- 
 tions by other organisms long enough to insure unrestricted growth of the 
 Fusarium in the soil, and also to aid in preventing the possible introduction 
 of the pathogene into the uninoculated tiles in the same way. The tar 
 paper discs, "tanglefoot" and wire mesh cages were merely extra pre- 
 cautions taken to insure the success of the experiment at a time when 
 little was known of the life history of the organism. If without them 
 the inoculated tiles had produced diseased plants and the uninoculated 
 ones, uniformly healthy plants, they could reasonably be regarded as 
 superfluous. 
 
 The results of this experiment were convincing. Several of the young 
 plants growing in inoculated soil were severely stunted almost from the 
 start. Two of them died before they were 1 foot tall, but owing to their 
 extremely slow growth, this was a matter of about two months. The 
 balance of the plants in the inoculated row grew rather rapidly at first 
 and at the end of two months some of them appeared to be in as good con- 
 dition as the control plants growing in sterile soil and also the controls in 
 untreated soil. From that time on, however, a remarkable difference in 
 development took place. Eight months from the time of planting all of 
 the plants in inoculated soil showed unmistakable signs of the disease. 
 Practically every condition mentioned under "Symptoms" was present 
 in one or more of the plants, including, dwarfing, yellowing and wilting 
 of the leaves and splitting of the pseudo-stem (plate XXVI, fig. 2). All 
 of the check plants were healthy appearing and had made a rapid, vigor- 
 ous growth. The difference between the two rows of plants is shown con- 
 spicuously in the accompanying illustration (plate XXX). The check 
 plants, both those growing in sterilized and unsterilized soil, averaged 
 more than ten times as large (by weight) as those grown in inoculated 
 
370 Phytopathology [Vol. 9 
 
 soil. Dissection of the inoculated plants revealed the typical internal 
 signs of the disease. A fungus was readily isolated from the diseased 
 plants, which when grown in pure culture proved to be identical in cul- 
 tural characters with the one used for inoculation. The internal tissues 
 of check plants cut down for examination showed absolutely no sign of 
 disease. 
 
 Some of the check plants were allowed to grow unmolested to deter- 
 mine if growing in the confined space of the tiles would have any influence 
 in limiting their development. These plants all attained normal size and 
 bore excellent bunches of fruit. Koch's rules of proof were fully com- 
 plied with in this experiment, and there no longer remains any doubt as 
 to the cause of the disease. 
 
 During the course of these investigations it was determined to find 
 out if the organism could attack some of the genera nearly related to Musa, 
 and also some of the plants which are subject to wilt caused by organisms 
 morphologically indistinguishable from Fusarium cubense. 
 
 Accordingly, plants of the genera Ravanella (traveller's palm), Heli- 
 conia (parrots tongue), Strelitzea, and Canna were copiously inoculated 
 with pure cultures of Fusarium cubense below the surface of the ground. 
 No infection resulted from any of the inoculations. Later cotton plants 
 (Gossypium hirsutum) were inoculated in various ways, with more in- 
 teresting results. The fungus inoculum was proved to be toxic to cotton 
 but typical wilt symptoms were not produced, nor was there a normal 
 invasion of the vessels. Two methods of inoculation were used, inocula- 
 tion of the soil, and direct inoculation of the plant tissues. In the first 
 experiment fifteen 6-inch pots of light soil were sterilized in the autoclave 
 by heating three hours at 16 pounds pressure. The soil in twelve of the 
 pots was then very heavily inoculated by emptying the contents of a 500 
 cc. Erlenmeyer flask containing a two-weeks old culture of Fusarium 
 sp. growing in plain corn meal into each pot. The cultures were thor- 
 oughly stirred into the soil with a sterile glass rod. Three pots were so 
 inoculated with F. vasinfectum, and the balance were inoculated with F. 
 cubense, three each from Porto Rico, Cuba and Costa Rica. Cotton 
 seeds, that had been treated for seven hours in calcium hypochlorite (32) 
 were then planted in all of the pots. The very interesting result was 
 that in two weeks from the time of planting, all seedlings growing in the 
 inoculated soil were very much stunted and had curled and distorted leaves, 
 while the check plants showed no such symptoms and were fully three 
 times as tall (text fig. 4) . When dug and examined, the roots of plants in 
 inoculated soil were found to be discolored and injured, some of them 
 partly rotted. The vessels in the stem were discolored but hand sections 
 of fresh material did not disclose the presence of any fungus. White 
 
1919] 
 
 Brandes: Banana Wilt 
 
 371 
 
 mycelium growing on the surface of rotted roots was proved, upon cul- 
 tivation to be the Fusarium used for inoculum. 
 
 Cotton seedlings three weeks old were next inoculated with macro- 
 conidia by making longitudinal slits in the hypocotyl with a sharp scalpel 
 and inserting the inoculum with a platinum needle. The same number 
 of plants and the same strains of Fusarium were used for this experi- 
 ment as for the preceding. The results were not so uniform, still they 
 were convincing. Where an abundance of inoculum was used the plants 
 wilted and died in from one to two days (text fig. 5). The stems above 
 and to a lesser extent below the wounds became discolored and finally 
 rotted. Where a small amount of inoculum was used the plants were 
 undeniably injured, as was made evident by a retarded development, 
 but they seemed to recover and resumed growth. The control plants, 
 
 Fig. 4. Cotton Growing in Soil Inoculated with Fusarium cubense (left) 
 Cotton Growing in Sterilized Soil (right) 
 
 which had been wounded in exactly the same manner with a sterile scalpel 
 were all half again as tall as the largest of the inoculated plants one week 
 after inoculation. Only one of the plants inoculated with F. vasinfectum 
 showed typical symptoms of cotton wilt. The vascular bundles became 
 discolored, and the organism was reisolateci from a leaf petiole. Fungus 
 hyphae were determined to be present in the brown, injured petioles of 
 plants copiously inoculated with F. cubense also. 
 
 It is to be regretted that the results of this experiment were not more 
 decisive. The writer believes, however, that certain conclusions may be 
 drawn from this evidence, when it is supplemented with the results ob- 
 tained (p. 365) in the experiments on the effect of by-products of the 
 fungus on plants selected at random from the vegetable kingdom. The 
 fact that pathologic symptoms were produced when the banana organism 
 was copiously inoculated into cotton plants is not complete proof that the 
 organism is F. vasinfectum. 
 
372 
 
 Phytopathology 
 
 [Vol. 9 
 
 There are good grounds for assuming that the fungus excretes substances 
 that are poisonous to many plants. It would also appear that the roots 
 of the cotton plant do not exercise any selective absorption with refer- 
 ence to these excreted substances. Where an abundance of inoculum 
 is inserted into the plant tissue through a wound, it seems to have the 
 same toxic effect. The extent of the injury varies directly with the 
 amount of inoculum used. 
 
 Fig. 5. Cotton Plant at the Right Inoculated Copiously at the Hypocotyl 
 with macroconidia of fusarium cubense 
 
 Control plant at left. Notice healed wound in control plant 
 
 Altogether too little is known concerning the chemistry of the process 
 of penetration of uninjured plant tissues by parasitic fungi. It is as- 
 sumed from analogy that all such penetrations are brought about by the 
 ability on the part of the fungus to excrete enzymes capable of dissolving 
 cellulose, cutin or lignin. If that were the only essential property, so 
 far as we know of the constitution of cell walls, such a fungus ought to 
 be able to attack almost any plant, but we know to the contrary that a 
 
1919] Brandes: Banana Wilt 373 
 
 wonderful degree of specialization exists. Natural infection of its special 
 host by a fungus of this type may be brought about irrespective of the 
 amount of inoculum. It has been shown that the pathologic symptoms 
 in cotton induced by F. cubense depends on the mass action of a rela- 
 tively enormous amount of inoculum. It is believed that the injury to 
 buckwheat and beans was brought about by the same agency. It is pro- 
 posed, therefore, to consider the two organisms, F. cubense and F. vasin- 
 fectum as distinct, at least until more convincing evidence is brought 
 forward to the contrary. 
 
 In conclusion, it appears that this aggressive parasite is known to 
 attack only four (possibly one or two more) of the hundreds of varieties 
 of bananas. In view of this apparent high degree of specialization, it 
 would not seem reasonable to look for natural infection of a plant phylo- 
 genetically so far removed as cotton. 
 
 (I. Names and synonymy 
 
 Essed (13) in 1911 as a result of his studies on the disease calls the 
 pathogene UstilaginoideUa musaeperda. He figures macroconidia, mi- 
 croconidia and chlamydospores of the Fusarium type, together with an 
 assortment of ascomycetes and even phycomycetes, all of which he calls 
 the same organism. There is some doubt that the Fusarium he figures is 
 the pathogene, since he shows a preponderance of 4-septate macroconidia. 
 
 Drost (10) in 1912 designates the pathogene as Leptospora musce. His 
 illustrations show macroconidia and microconidia of a Fusarium, sup- 
 posedly the imperfect stage of an ascomycete which is also figured. He, 
 as well as Essed, evidently worked with mixed cultures, but since he 
 obtained infection by inoculation with the mixture, probably the pathogene 
 was present. 
 
 Dr. Erwin F. Smith (28) in 1910 obtained pure cultures of a Fusarium 
 from wilted bananas sent from Cuba. He demonstrated the ability of the 
 fungus to grow in the xylem of bananas in a greenhouse at Washington, 
 D. C. No technical description of the organism was given, and no com- 
 parison with other Fusaria was made, but on the strength of its location 
 in the banana tissue and its evident ability to cause the disease, he con- 
 sidered it a new species and named it F. cubense. 
 
 This name, then has priority over any other published name, and 
 there is assurance that it is the parasite in question. 
 
 The organism is a typical member of the genus Fusarium. It belongs 
 with the group of vascular parasites placed in the section Elegans, pro- 
 visionally erected by Wollenwebber (33) to facilitate the division of the 
 genus Fusarium into more or less natural groups. Its very close relation- 
 
374 Phytopathology [Vol. 9 
 
 ship to F. vasinfectum has been noted by the writer (6). In common 
 with F. vasinfectum, strains of F. cubense have the peculiarity of possess- 
 ing or lacking an agreeable aromatic odor, similar to that of some of the 
 aldehydes of the aliphatic series when grown on steamed rice and certain 
 liquid media. No culture from Porto Rico has been observed to have 
 this odor. It is present or absent in cultures from Cuba. All cultures 
 that the writer obtained in Panama, Costa Rica and Jamaica generated 
 the substance causing this odor. This difference is sharply demarked, 
 and seems sufficient to justify the erection of a new variety. 
 
 Fusarium cubense 
 
 Fusarium cubense, E. F. Smith, 1910, in Science n.s., v. 31, no. 802, 
 p. 754, 755. 
 
 Ustilaginoidella musaeperda, Essed, 1911, in Ann. Bot. 25, pp. 343-361. 
 
 Leptospora musae, Drost, 1912, in Bui. 26, Dept. van der Landbouw 
 Suriname. 
 
 Sporodochia on leaf stalks and blades, separate, arising from pseudo- 
 parenchymatous substratum in substomatal cavities, hyaline; conidio- 
 phores septate, projecting through stomatal apertures, hyaline, verticil- 
 lately branched, about 70 /x long and 4 /x in diameter, apical end tapering 
 gradually to a point; branches in whorls of 3, continuous, arising from 
 upper end of cells of the conidiophore, apical ends tapering abruptly; 
 microconidia hyaline, oval or elongate, continuous, 1- or 2-septate, 5 
 to 7 m by 2.5 to 3/x; macroconidia of the Elegans type, 3- to 5-septate, 
 sickle-shaped, pedicillate at the base, more than 95% of 3-septate macro- 
 conidia present, 22 to 36 /x by 4 to 5 xx, a few 4- and 5-septate conidia, born 
 at apical ends of conidiophores and lateral branches. 
 
 In pure culture on potato plugs, indigo blue sclerotia, irregular, nodule 
 like, 1 to 4 mm. in diameter, produced, in 12+ days. Pionnotes produced 
 on Melilotus alba steins, conidia salmon-colored in mass. On steamed rice 
 grains a pink or pinkish salmon color imparted to the substratum at first, 
 later becoming blotched or streaked with red, and finally tinged with 
 vinaceous purple or wholly blue. Strong aromatic odor on this medium. 
 Chlamydospores ellipsoidal to globose, terminal, intercalary or conidial, 
 simple, paired or catenulate, when unicellular 5.5 to 6 /x by 6 to 7 /x- Vas- 
 cular parasite, causing wilt of Musa sapientum. 
 
 Fusarium cubense var. inodoratum n. v. 
 
 Differs from F. cubense by the absence of odor on steamed rice, etc. 
 Vascular parasite, causing wilt of Musa sapientum. 
 
1919] Brandes: Banana Wilt 375 
 
 e. Life history 
 
 There are two, so to speak, distinct life histories of F. cubense, in one 
 of which the parasite may be constantly associated with the host for 
 many years. In the other, there is a definite alternation of pathogenesis 
 and saprogenesis. 
 
 The first of these owes its existence to the method of propagating the 
 banana, and the ability of diseased rhizomes to produce suckers which 
 are only slightly infected. These latter are sometimes able to make a 
 strong growth, and in turn produce suckers before the disease terminates 
 in death. The slightly affected suckers may remain attached to the 
 parent stool, or may be cut away and planted elsewhere. Owing to the 
 carelessness and ignorance of man, this practice has resulted in a wido 
 distribution of the disease. If there were no other method of spreading 
 the disease, it would be restricted to the progeny of originally diseased 
 plants. Unfortunately, however, nature has provided for an almost 
 unlimited dissemination of the pathogene to new localities under the proper 
 conditions. It will not be necessary to discuss the details of the growth 
 of the parasite in the host, where it is constantly associated with the 
 latter, since the process is essentially the same as in the other method. 
 It may be mentioned merely that the vascular systems of parent bulb and 
 sucker are of course continuous, so there is no obstruction to invasion of 
 the latter from the former. The second type of life history will be dis- 
 cussed in detail. 
 
 1. Source of inoculum. The inoculum consists of macroconidia and 
 microconidia (plate XXXI, fig. 1) produced by minute sporodochia 
 (plate XXXIV, fig. 2) which are born on both surfaces of the leaf blade, 
 on the leaf stalk or even on the leaf bases which form the pseudostem. 
 The sporodochia are most numerous in the "axils of the leaves," where 
 the leaf stalk diverges from the sheathing leaf base. Conidia apparently 
 are produced at any time of the year, the governing factors being the stage 
 of the disease and high atmospheric humidity, or abundant rainfall. 
 
 2. Dissemination. They are loosely attached to the conidiophores, 
 not being held by a gelatinous matrix or any other provision to prevent 
 dissemination when dry, so that they are easily dislodged by the wind, and 
 carried on slight currents of air. Sterile agar plates, exposed for one-half 
 hour beneath diseased banana plants and between the rows in the plan- 
 tation at.Mayaguez, Porto Rico, have yielded colonies of F. cubense, the 
 number of colonies per plates varying from 1 to 30. 
 
 It is not known just how far these conidia may be carried by the 
 wind in a viable condition. They are extremely small and light. Judg- 
 ing from the results of experiments (22) with uredospores of Cronartium 
 
376 Phytopathology [Vol. 9 
 
 ribicola in which it was found theoretically possible for them to be car- 
 ried many miles, it may be assumed that the limiting factor in the 
 dissemination of these conidia by the wind is really their desiccation 
 deathpoint. 
 
 In the tropical downpours which are so common in many banana pro- 
 ducing countries, it frequently happens that the water accumulates on the 
 surface of the ground so rapidly that it is not immediately absorbed, but 
 runs in sheets or streams for considerable distances. There can be no 
 doubt that this affords a ready means for the dispersal of spores for short 
 distances. 
 
 In the main the conidia so produced suffer one of two fates, they are 
 either carried to the infection courts of fresh victims, or they are de- 
 posited on other plants or on the soil. A preponderating majority of the 
 conidia produced suffered the latter fate. It has been abundantly proved 
 that the fungus may remain alive in the soil in some form or another for 
 long periods of time (6) . Samples of soil taken under aseptic conditions 
 from banana plantations and adjoining fields, and plated out by the loop 
 dilution method show that the organism is present therein in enormous 
 numbers. For these determinations, it is preferable to use a specially 
 prepared medium. It should be acid enough to inhibit the growth of 
 bacteria and should contain some form of starch so that the organism 
 may be quickly recognized by its characteristic color reaction. 
 
 On account of the number of colonies produced on the agar plates 
 (many thousands per gram of soil), it is thought that the organism is 
 present in the soil as spores of some kind. Slight success has attended 
 the direct microscopic examination of soil. Soil samples were taken in 
 the same manner as for plating out, and were placed in small vials of 
 sterile distilled water, about 1 gram of soil to 10 cc. of water. This was 
 shaken up thoroughly, and when the larger particles had settled, a loopful 
 of the suspension was placed on a clean cover slip. It was allowed to dry 
 in the air, and was then killed and fixed by passing rapidly through a 
 flame. The preparations were stained one minute in a 2 per cent solution 
 of Bismark brown in 70 per cent' alcohol, and mounted in Canada balsam. 
 The chief difficulty in this method is the extreme dilution necessary. It 
 is estimated that with the maximum number of fungus bodies present 
 according to plate count, in whatever form they might be, many such 
 preparations could be made that would not contain a single body. 
 They would, however, contain thousands of fragments of organic matter, 
 minute bits of sand, crystals, etc. Confusion due to the latter two is 
 largely eliminated by the fact that they do not take the stain. Positive 
 identification of biologic forms in such preparations may be said to in- 
 crease in a geometric progression with the number of such forms present. 
 
1919] Brandes: Banana Wilt 377 
 
 The writer searched such preparations diligently for a whole day and was 
 rewarded by finding two Fusarium macrospores, one of which had an in- 
 tercalary chlamydospore, several microspores and numerous bodies that 
 might have been detached chlamydospores. No vegetative structures 
 were seen whatsoever. These preparations were saved. There can be 
 no doubt as to the identity of the macrospores, that is, they are unques- 
 tionably forms of a Fusarium. The other findings would seem to indi- 
 cate that chlamydospores may be formed. The absence of vegetative 
 structures does not prove that the organism cannot grow in the soil. 
 Laboratory experiments suggest strongly that it can. Large test tubes 
 containing sterile, moist, clay loam were inoculated at the top. After one 
 month the bottom of the tube was broken and a fragment of the soil 
 plated out, with the result that the organism was recovered at this point. 
 Later U-tubes of soil were inoculated at the top of one arm, and the 
 organism was subsequently recovered from the top of the other arm. 
 
 The organism then is capable of growing in the soil. In what form it 
 may be is still somewhat obscure, but it is evident that if healthy bulbs 
 are planted in such soil, they probably can become infected. This leads 
 to consideration of another method of distribution of the fungus. Mud, 
 carried on the feet of men and animals, or on the wheels of vehicles would 
 serve as a medium for dissemination of the fungus for great distances. 
 
 Infected banana leaves are frequently found among the "trash" used 
 for protecting bunches when they are packed in cars for hauling to the 
 steamers. This trash is often carelessly thrown from the train at various 
 points along the route. It may be carried for great distances on the floor 
 of cars and probably has aided in distributing the fungus to new regions. 
 
 'Insects may act as carriers, but there is no definite information on this 
 point, and in the opinion of the writer it is of little importance. 
 
 3. Infection courts. It has been noted (p. 355) that the disease pro- 
 gresses from below the surface of the ground upward, originating either 
 at the large wound on the rhizome caused by separating the sucker from 
 the parent stool, or at the side of the bulb where a root is given off. At- 
 tempts to infect any of the above ground parts, either by spraying spores 
 on the unwounded surface, or by inserting the inoculum into wounds have 
 never given rise to the typical disease. Limited local growth of mycelium 
 in the vessels may follow inoculating into the tissues of pseudostem, stem 
 or leaf, but the method is not uniformly successful, and has never been 
 known to cause death or even serious symptoms. That the unwounded 
 young roots and wounded rhizome are the natural infection courts was 
 proved by the following experiments. 
 
 A young healthy plant was selected and the roots carefully exposed by 
 washing the soil away with a jet of water. Drops of a thin spore sus- 
 
378 Phytopathology [Vol. 9 
 
 pension were placed on the uninjured thread-like roots about six or eight 
 inches from the bulb. Sterile moist cotton was wrapped about the point, 
 and the soil replaced. In twelve hours the root was removed and the in- 
 oculated portion killed and fixed in chrom-acetic acid. 
 
 Similarly the rhizome was partly exposed and a cut made in the stele 
 with a sharp razor. The flat cut surface was inoculated in the same way, 
 protected with cotton, and the soil replaced. After twelve hours, cubes of 
 the stelar tissue were removed from the inoculated surface, in such a way 
 that the latter were included as one face of the cube. These were also 
 killed and fixed in the chrom-acetic acid solution. All of the tissues 
 were then washed, dehydrated, infiltrated with paraffin in the usual way, 
 imbedded, and radial, transverse tangential, and serial sections, were cut 
 with the microtome. The sections were stained twelve hours in safranin 
 and three minutes in Delafield's hematoxylin, and mounted in balsam. 
 Microscopic examination of these slides showed that in the case of the cut 
 stelar tissue, the mycelial threads had in twelve hours penetrated the cut 
 ends of vessels, and the parenchymatous starch containing cells in great 
 abundance. In some cases the fungus was demonstrated by transverse 
 sections to have passed through eleven cell walls. The average depth of 
 penetration was from eight to ten cells from the cut surface. Beyond 
 this point there was no invasion and the tissue was normal. 
 
 In the case of the thread like roots, there was no evidence of penetra- 
 tion whatever. The spores had germinated and grown among the root 
 hairs, but were not seen to penetrate them, or the epidermal cells of the 
 root at all. The roots used in this case were some distance from the bulb 
 and the cells were mature, i.e., the tip of the root had not been used for in- 
 oculation. It was determined to repeat the experiment, using young roots 
 to see what effect the fungus might have on meristematic root tissue and 
 young cells just back of the root cap. Several of the large fleshy roots, 
 just emerging from the cortex of the rhizome were selected for this pur- 
 pose. They were not more than § inch long at the time, and were possibly 
 f of an inch in diameter at the base. They were exposed and inoculated 
 without injuring them, in the same manner as previously described for 
 the thread like roots. The same culture was used as in the former in- 
 oculations. After twelve hours they were removed, killed, fixed, and 
 stained, serial sections prepared as before described. Upon microscopic 
 examination a quite different condition was found -to exist (plate XXXI, 
 fig. 2). The epidermal cells and cortex had been deeply penetrated. 
 The root hairs (fragments of which are shown at the surface in the photo- 
 •micrograph) were not attacked, and do not serve as a means of entrance 
 for the parasite. 
 
1919] Brandes: Banana Wilt 379 
 
 The xylem elements of the young roots were not at all lignified at this 
 stage of development (as indicated by their not taking the red of the 
 safranin stain), and would not offer any resistance to invasion by the 
 fungus. Proof, then, that the young fleshy roots, and injured stelar 
 tissue act as infection courts is considered to be amply furnished by these 
 experiments. 
 
 4- Type of penetration. Penetration by both macroconidia and micro- 
 conidia is accomplished directly by means of a germ tube. The germ 
 tube may arise from any cell of the macroconidium. It is quite thick, at 
 least equal in diameter to the conidium from which it arises. This is in 
 great contrast to the germ-tubes of conidia germinating in water, where 
 they are very attenuated and thread-like (plate XXXII, fig. 1). It is not 
 known by what means the tip of the germ tube penetrates the outer cell 
 wall of the epidermal cells, but it is probably accomplished by the excre- 
 tion of some dissolving enzyme. These cell walls in the young root tip 
 are not especially thick, and do not seem to be cutinized or cuticularized. 
 With the stains used, they did not react differently than the other cell 
 walls of the cortical tissue. 
 
 5. Growth of the parasite in the host. Immediately upon penetration of 
 the epidermal cells, a food relationship is established with the host, and 
 the germ tube becomes rapidly growing mycelium. There are no special 
 absorption organs. The contents of the host cells is evidently absorbed 
 at any point in the undifferentiated cell walls of the parasite. The my- 
 celium is septate and profusely branched from the start (plate XXXI, fig. 
 2) . Growth is intracellular and the cell walls seem to offer little resistance 
 to the passage of the hyphae. When the xylem elements are reached, the 
 fungus grows in the lumen of the trachae, proceeding towards the stele. 
 The most abundant vegetable growth found anywhere in the vascular 
 tissues of the host occurs in the root vessels, close to the point of infection. 
 Evidently growth proceeds more rapidly in the vessels than elsewhere, for 
 a cross section of the root where it passes through the fleshy cortex of the 
 rhizome, will reveal a still abundant growth of mycelium in the lumen of 
 the vessels (plate XXXII, fig. 2), but none can be detected in other cells of 
 the vascular bundle. The mycelium now continues its growth through 
 the continuous system of vessels, entering the stele (plate XXXIII, fig. 1), 
 thence passing again through the cortex at the upper part of the rhizome, 
 where the vascular bundles form the leaf-traces, and so on up into the 
 leaves, but apparently always being confined to the vessels. Soon after 
 entering the young xylem elements in the root, when growth has proceeded 
 a little distance towards the rhizome, the fungus finds itself in the older, 
 lignified tracheae. Probably there would be more difficulty in penetrating 
 the walls of these old lignified vessels, if indeed it were not impossible, so the 
 
380 Phytopathology [Vol. 9 
 
 fungus follows the line of least resistance upward through the lumen, unim- 
 peded by cross walls of any kind. At least it is a fact, that once the fungus 
 enters the vessels of the root, it does not escape from the vessels until 
 fructification takes place high up in the leaves. It has been remarked 
 that the mycelium in the vessels of the pseudostem is very scant (plate 
 XXXIII, fig. 2). It has never been observed to be abundant in vessels 
 anywhere excepting in the infected root close to the point of inoculation. A 
 logical explanation of this is to be found in the meager supply of organic 
 food in the vessels. Studies on the nutrition of heterotrophic plants show 
 us that a luxuriant growth is not to be expected in the absence of an ap- 
 preciable quantity of organic food. Where would this be obtained in the 
 vessels? Unless the roots have absorbed some organic substance in solu- 
 tion (20) which may be possible in some plants, the available food would 
 consist of certain salts in solution, and possibly some material obtained 
 by the partial digestion of the spiral or reticulated secondary thickening 
 of the vessels. There is no known abundant source of food in this situ- 
 ation, and it is not surprising that the vegetative growth of the fungus 
 is not luxuriant. This paucity of food may account for the production 
 of microconidia in the vessels of the pseudostem, noted by Dr. Erwin F. 
 Smith in 1910 (28). 
 
 It is known that with many fungi, fructification takes place more readily 
 after the medium has become impoverished. 
 
 The scant mycelium extends in the late stages of the disease to the leaf 
 stalk and even to the midrib of the leaf. By the time that this has oc- 
 curred, the leaf has exhibited external symptoms of disease. It has 
 become yellow, and possibly wilted, so that it hangs limply by the side 
 of the pseudostem. At this point, probably due to the weakened con- 
 dition of the host, the fungus escapes from the vessels, and is found in 
 the sieve tubes, and in fact in all of the parenchymatous cells of the leaf 
 stalk. It is particularly abundant in the epidermal cells and subepidermal 
 cells (plate XXXIV, fig. 1). The mycelium is here very abundant branch- 
 ing and intracellular. The subsequent development of sporodochia and 
 the production of conidia have been discussed under "morphology of 
 the causal organism" and need not be repeated here. 
 
 7. ECOLOGY 
 
 The severity as well as the spread of banana wilt are strikingly corre- 
 lated with certain well defined weather conditions. Wet weather, es- 
 pecially if it is characterized by dashing rains with movement of surface 
 water, is apparently a contributing factor of considerable importance in 
 the dissemination of the pathogene (cf. p. 376). This assumption is based 
 
1919] Brandes: Banana Wilt 381 
 
 i 
 on field observations in Porto Rico. There, it may readily be seen that 
 shortly after the rainy season has set in, there is a decided increase in the 
 number of apparently new cases, and an unmistakably more rapid progress 
 of the disease in plants already infected. The first of these phenomena 
 may be accounted for in either or both of two ways. Laboratory exper- 
 iments (p. 375) indicate that the conidia are very subject to injury by 
 desiccation. During the dry season the surface of the soil is very hot and 
 dry. The is little likelihood that conidia deposited thereon would sur- 
 vive. In addition to that fact, a certain amount of atmospheric humidity 
 seems essential to the development of sporodochia, at least they are more 
 abundantly produced when a high degree of humidity prevails. The 
 other explanation is that the young, fleshy roots, which are known to be 
 especially susceptible infection courts, are not developed during the dry 
 season, but push out from the rhizome in large numbers after the rains 
 have set in. The explanation for the other observed phenomenon, 
 namely, that previously infected plants go down with the disease much 
 more rapidly in wet weather, is more obscure. The answer may be sought 
 in the fact that there is a retarded transpiration a*nd consequently less 
 conduction of water and solutes in the vessels when the soil is dry, so that 
 the toxin assumed to be excreted by the fungus would not be carried up- 
 ward into the pseudostem and leaves in such large amounts. In wet 
 weather on the contrary, due to the abundance of soil moisture and the 
 hot dry winds that prevail during a part of the day, transpiration would 
 be more active, and conditions providing for a more plentiful upward 
 distribution of the water and its inclusions would prevail. 6 
 
 It has been briefly mentioned that in arid regions, where irrigation is 
 necessary for the successful production of bananas, the disease is unknown. 
 The writer has never seen nor heard of a case of banana wilt in the arid 
 regions of southern Porto Rico, southern Jamaica or northern Colombia, 
 with the exception of one plant observed in St. Catherine Parish, Jamaica, 
 which was evidently accidentally inoculated with material brought in 
 for experimental purposes. In such regions the intense heat and dryness 
 of the surface of the soil would make the survival of wind-born conidia 
 impossible. In southern Jamaica, the soil is so hot and dry that planters 
 invariably set the bulbs with about 1 foot of the old pseudostem still 
 attached, so that the succulent latter portion when it rots will furnish 
 moisture for the young "peeper" and prevent its death by "boiling" as 
 it is termed. "Baking" would be a more descriptive term. Attempts 
 to isolate the organism from such soils failed, but it was readily isolated 
 
 6 It was not proved experimentally that there is actually more transpiration in 
 the wet than in the dry season. 
 
382 Phytopathology [Vol. 9 
 
 from soil in the vicinity of a diseased plant in northern Jamaica where 
 rains are frequent. It is believed that the disease need never be feared 
 in these arid regions. It is probable that in times past infected bulbs 
 must have been imported into such areas, but if they were, it is evident 
 that under the prevailing conditions the plants either outgrew the disease 
 or were killed, and the suckers were not used for propagation. 
 
 In regions where there are no well demarked wet and dry seasons, that 
 is, where the rain is more or less evenly distributed throughout the year, 
 optimum conditions for dissemination of the fungus exist. This is at- 
 tested by the virulence of the disease in the banana districts of Panama 
 and Costa Rica. Even there one finds an occasional drought, but no well 
 defined annual dry season. In Nicaragua, Guatemala, and Honduras, 
 conditions are about the same, so it is to be expected that a repetition of 
 the history of the malady in Panama will take place in those countries 
 unless extraordinary steps are taken to prevent it. 
 
 In Surinam, where the annual precipitation is very heavy, and the rains 
 are distributed over practically the whole year, the disease spread over 
 the entire country in' the short space of four years. In such regions it is 
 not only highly infectious, but extremely virulent and aggressive. At- 
 tacked plants succumb to the disease very quickly, whereas in dryer 
 regions the plants may survive for many months, sometimes nearly a 
 year after becoming infected. 
 
 8. CONTROL 
 
 Introduction 
 
 A scrutiny of the life history of the pathogene will convince one that 
 its elimination as a source of injury to the banana plantations in the 
 countries where conditions are favorable to it, cannot be based oh an 
 attempt to eradicate the parasite. It is true that a rational system of 
 sanitation, based on the findings of this paper, would aid in suppressing 
 the disease somewhat, and if universally practiced for some years might 
 materially reduce the amount of loss. Any direct method of attack, 
 however, such as protecting the plants by the application of fungicides or 
 eradication of the parasite in the soil by any method now in use is obvi- 
 ously out of the question. Aside from the fact that fungicides which 
 migh be used to treat the bulbs before planting would soon become in- 
 effective, the very susceptible young fleshy roots, as they pushed out into 
 the soil would be absolutely unprotected. Sterilization of the soil is not 
 practicable considering the present market price of bananas. Selection 
 of disease-free bulbs is of value only where they are to be planted in soil 
 which is not already infested. It is manifest that in view of the different 
 
1919] • Brandes: Banana Wilt 383 
 
 climatic and soil conditions and other factors, the question of control 
 measures presents separate and distinct problems in the several banana- 
 growing regions. Experiments on the control of banana wilt in badly 
 infested areas at the Porto Rico Agricultural Experiment Station have 
 up to the present led to no very definite conclusions. It is considered of 
 some value, however, to briefly mention a few of the methods of attack, 
 even though for the most part they gave negative results. Control 
 measures for this disease, as for other diseases of plants would naturally 
 group themselves under one or more of four headings, namely exclusion, 
 protection, eradication and immunization. 7 
 
 1. Exclusion (quarantine) 
 
 Exclusion of diseased plants, plant parts or other infected material by 
 legislation would be of value only: (1) in countries where the disease is 
 not present but in which conditions are such as favor the disease, (2) in 
 similar regions where the disease is not yet firmly established, or (3) where 
 it is partly held in check by climatic conditions. 
 
 Honduras and parts of Guatemala would be included under the first 
 of these divisions. It is probable that a few isolated cases of the disease 
 exist even there. Under the second division would be the balance of 
 Guatemala and Nicaragua. The last division would include Jamaica, 
 Porto Rico and perhaps Cuba. Quarantine in connection with this dis- 
 ease may be general, local, or both, depending on various factors. It is 
 believed that in Honduras strictly enforced legislation providing for the 
 exclusion of banana plants or plant parts which might be sent from any 
 other banana region with the possible exception of the arid north coast 
 of Columbia would be a practical and profitable measure. In addition, 
 regulatory laws in regard to the importation of tools, machinery, etc., 
 formerly used in the cultivation or handling of bananas in other regions, 
 providing for their sterilization by fungicides is not considered an extreme 
 precaution. Such laws already exist in Jamaica. Orders have been 
 issued by the Governor of Jamaica in accordance with the laws, calculated 
 to prevent the introduction into the island of the inoculum in any form. 
 Furthermore, in Jamaica, this disease has been declared notifiable, that 
 is, it is required by law that any person occupying land on which plants 
 infected with banana wilt exist, must give notice of the same to the Di- 
 rector of Agriculture, who then directs that the treatment prescribed by 
 law be carried out. This treatment consists of the destruction by fire of 
 the diseased banana plant or plants, and all other plants surrounding them 
 
 7 These terms are used by Prof. H. H. Whetzel, head of the Department of 
 Plant Pathology, Cornell University, in his lectures on plant disease control. 
 
384 Phytopathology [Vol. 9 
 
 within a distance of one chain (22 yards), after which the infected area 
 is fenced in and a local quarantine established for one year or until such 
 time as it may be lifted at the discretion of the Director of Agriculture. 
 Several hundred cases were reported and treated in 1915 and 1916 but 
 less than a score were found in 1917, two years after the law became ef- 
 fective. These far-sighted laws were passed at a time when practically 
 nothing was known of the life history of the pathogene. Investigation 
 of the latter shows that they were amply justified. Certain features of 
 the law could doubtless be improved upon, but the leveling and destruc- 
 tion of all affected plants, which means that they could not be used for 
 propagation, or serve as a source of wind- or rain-born conidia, and the 
 exclusion of traffic by man and animals through the infected area, has un- 
 questionably served in a large measure to prevent the spread of the disease. 
 It is believed that the progress of the disease would be greatly retarded 
 in Honduras or other countries where only isolated cases occur, if similar 
 steps were taken. It is thought also, that the same measures would be 
 practical in regions where the disease has gained more of a foothold, but 
 where its progress is somewhat retarded owing to the fact that optimum 
 conditions k>r dissemination of the pathogene are lacking. This of course 
 has been abundantly proved for Jamaica conditions. Porto Rico and 
 Cuba are included in the regions of this character. It is realized that 
 this procedure does not constitute absolute control of the disease. Even 
 in places where it is applicable, it appeals to the writer merely as a tempo- 
 rary measure calculated to alleviate the situation, to be used pending the 
 time when banana wilt may be eliminated by some other method. 
 
 2. Protection 
 
 Protection is defined as the interposition of some effective barrier between 
 the susceptible part of the host and the inoculum of the pathogene. It 
 has been found that the susceptible parts of the host are below the surface 
 of the ground. This means that the necessary renewal of any substance 
 toxic to the fungus would be impossible. Additional applications of such 
 substances are necessary to provide protection for the increased new areas 
 brought about by growth, as well as for the purpose of renewing the in- 
 hibiting substance on old mature parts as it gradually weakens due to 
 washing and leaching away. Any method of attack, similar in principle 
 to the spraying of above ground parts is impracticable. That the appli- 
 cation of such .a principle is unsound is apparent to any one now that we 
 have positive knowledge of the life history of the organism, nevertheless 
 cases have come Under the observation of the writer where investigators 
 have attempted to control this disease both by the application of fungi- 
 
19191 Brandes: Banana Wilt 385 
 
 cides to the leaves and pseudostems and by dipping the bulbs in fungi- 
 cides previous to planting. The fallacy of attempting control measures 
 before completion of the study of the organism is recognized, yet in the 
 hope of saving time, much useless expenditure of energy and money is 
 often made. The writer was exceedingly mortified by the result of an 
 experiment on the control of this malady, started at a time when evidence 
 seemed to point to dissemination of the pathogene through the medium 
 of the soil. It is here mentioned in support of the above statement. Iso- 
 lation of the organism from the soil had been accomplished, and field 
 observations on the origin of new cases led to the belief that the fungus 
 spread through the soil by vegetative growth. A field which had been 
 growing sugar cane for the previous three years, and vegetables for some 
 years before it was put into cane was selected for the experiment. A 
 badly diseased plantation of bananas adjoined the field on two sides. 
 Trenches 2 feet deep were excavated in rows 10 feet apart both ways. 
 The plat was about 100 feet from the edge of the banana plantation. 
 Examination of the soil did not reveal the presence of the fungus. Healthy 
 banana bulbs were planted in the center of each 10 foot square made *by 
 the intersection of the two parallel series of trenches, the idea being that 
 an effective barrier to the growth of the fungus would be made by re- 
 moving the top layer of humus containing soil. One after another of the 
 plants fell victim to the disease. One year after the experiment was 
 started 60 per cent of them were affected. We now know that it was 
 perfectly possible for such infections to have resulted from wind borne 
 conidia. Such a method of protection is of no more value than the 
 application of fungicides. 
 
 3. Eradication 
 
 Practically all of the efforts so far made in attempting to control this 
 disease have been directed along the lines of eradicating the parasite. 
 
 The first principle of crop sanitation is to avoid planting infected bulbs 
 which invariably give rise to diseased plants. The necessity for selecting 
 healthy bulbs for planting cannot be too strongly impressed upon banana 
 growers. It is of the utmost importance in recently opened banana land 
 where new plantations are being layed out. The practice should be fol- 
 lowed even where the disease has been established, but it must be re- 
 membered that when clean bulbs are planted in infected soil, many of them 
 will become diseased. When examining bulbs for evidence of disease, it 
 is essential that the machete be sterilized by fire or by carefully wiping 
 it with a cloth soaked in some fungicidal solution every time a fresh cut 
 is made. Every bulb which has the appearance shown in plate XXIII, 
 fig. 2, in whatever degree the discoloration may be present should be 
 
386 Phytopathology [Vol. 9 
 
 rejected or destroyed by fire. Only bulbs which look like the one figured 
 in plate XXII, fig. 2, are suitable for planting. These should be kept 
 carefully separated from the diseased bulbs after sorting. 
 
 When a diseased individual is detected in the field, it should be imme- 
 diately rooted out and destroyed since it soon becomes a menace to sur- 
 rounding plants on account of the production and dispersal of conidia. 
 Where only isolated cases occur, and firewood is available the entire plant 
 should be removed from the ground, cut into thin slices with a machete 
 and burned. On account of the succulent nature of the banana plant 
 this entails much labor, but heedless neglect of such plants is bound to 
 reap a harvest of new cases. 
 
 Eradication of the pathogene in the soil by allowing the land to lie 
 fallow, rotation of crops, disinfection of the soil, liming, mulching, flooding 
 and other methods have been tried, but none have yet been found practical. 
 
 The longevity of the organism in the soil is not definitely, known, but 
 it has been isolated from land which had not been in bananas for five 
 years. Bulbs planted in this field quickly became diseased. 
 
 In connection with studies on the etiology of the disease it was found 
 that steam sterilization of the soil was very effective, but it is needless 
 to say that under present conditions this method is impracticable. Disin- 
 fection of the soil in small plats by drenching with copper sulphate, car- 
 bolineum and formaldehyde was not only unsuccessful, but the expense 
 prohibitive. Other attempts to eradicate or render innocuous the parasite 
 in the soil were worse than useless. 
 
 4. Immunization 
 
 The successful measures described up to the present are at best only 
 palliative, and are applicable only in restricted areas where the disease 
 has not yet become rampant. It remains for some method of control to 
 be developed whereby the vast areas of fertile banana land in Panama, 
 Costa Rica and Surinam which have been absolutely abandoned owing 
 to the ravages of the banana wilt pathogene may once again be made 
 productive. The writer is convinced that there is only one solution to 
 the problem, and that lies in the development of resistant strains of the 
 desirable varieties of bananas. Attempts have been made to substitute 
 resistant varieties for the ones now on the market, but all of them have 
 had one or more fatal defects which have prevented their meeting the 
 requirements of the market, the shippers or the growers. 
 
 Experiments have been started at the Porto Rico Agricultural Ex- 
 periment Station in the hope of obtaining resistant strains of the Cha- 
 maluco banana. The process of selection is necessarily long and tedious 
 
1919] Br andes: Banana Wilt 387 
 
 with a crop that requires more than a year to mature. It is too early to 
 make predictions as to the outcome, but the indications are such that a 
 successful termination is awaited. It was observed that in a large planta- 
 tion of Chamaluco bananas on the experiment station grounds which had 
 been a veritable hotbed of infection for some years, an occasional plant 
 would resist the disease and make a normal growth. These plants pro- 
 duced good bunches of fruit, and gave rise to healthy suckers. Apparently 
 optimum conditions for infection existed. The healthy stools had been 
 surrounded by diseased plants for many generations. The progeny of 
 these healthy plants were removed on June 23, 1916, to a specially pre- 
 pared field which was artificially inoculated as heavily as possible. Dis- 
 eased plants of this variety had been cut into small pieces distributed 
 evenly over the surface of the field and worked into the ground. It was 
 thus proposed to give the parasite every chance for infection. The prog- 
 eny of the stools that survive will be subjected to the same treatment. 8 
 If they in turn survive, it is believed that they can withstand the disease 
 under any conditions, and they will be propagated and distributed as 
 immune strains. 9 It is strongly recommended that similar experiments 
 be started with commercial varieties in the countries where bananas are 
 grown for export. 
 
 Bureau of Plant Industry 
 Washington, D. C. 
 
 8 Since these experiments were started Edgerton (12) has outlined a method of 
 obtaining wilt resistant tomatoes by first sterilizing the soil, and then inoculating 
 it heavily with pure cultures of the organism. 
 
 9 In April, 1919, the writer was in Porto Rico and paid a visit to this field. Of 
 the one hundred and five selected plants, sixty had survived the treatment. Suckers 
 from the latter healthy plants were set out in the same field after it had been plowed 
 up and reinoculated in the same manner as above described. 
 
38S Phytopathology [Vol. 9 
 
 LITERATURE CITED 
 
 (1) Ashby, S. F. : Banana diseases in Jamaica. Bui. Dept. Agr. Jamaica, s. 2, 2: 
 
 95-128, pi. 21-28. 1913. 
 
 (2) Ashby, S. F. : Banana diseases. Jamaica Dept. Ag. Publication. 1913. 
 
 (3) Atwater, W. O., and Bryant, A. P.: The chemical composition of American 
 
 food materials. U. S. Dept. Agr. Off. Expt. Stas. Bui. 28, rev. ed., 87 p. 
 4 fig. 1906. 
 
 (4) Basu, S. K. : Report of the banana disease of Chinsurah. Quart. Jour. Dept. 
 
 Agr. Bengal 4: 196-198. 1911. 
 
 (5) Brandes, E. W. : Report of the plant pathologist. Porto Rico Agr. Expt. Sta: 
 
 Rpt, 1915:34-35. 1916. 
 
 (6) Brandes, E. W. : Banana wilt (Panama disease). Porto Rico Agr. Expt. Sta. 
 
 Rpt. 1916: 29-31, pi. 4-5. 1918. 
 
 (7) Butler, E. J.: Plant pathology. Ann. Rpt. Bd. Sci. Advice India 1915-16: 
 
 103-109. 1917. 
 
 (8) Carpenter, C. W. : Banana diseases. Hawaii Agr. Expt. Sta. Rpt. 1917: 
 
 40-42, pi. 5. 1918. 
 
 (9) Coons, G. H.: Report of the plant pathologist. Michigan Agr. Expt. Sta. 
 
 29th Ann. Rpt. 1915-16: 265-271. 1916. 
 
 (10) Dash, J. S. : Report .... on the entomological and mycological work. 
 
 .... Banana. Rpt. Dept. Agr. Barbados 1913-14: 44-45. 1915. 
 
 (11) Drost, A. W. : De surinaamsche panamaziekte in de gros michel bacoven. 
 
 Dept. Landb. Suriname Bui. 26, 45 p., 11 pi. (1 col.) 1912. 
 
 (12) Earle, F. S. : Report on a trip to Jamaica, Jour. N. Y.Bot.Gard.4: 1-10. 1903. 
 
 (13) Edgerton, C. W. : A study of wilt resistance in the seed bed. Phytopathology, 
 
 8:5-14, 4 fig. 1918. 
 
 (14) Essed, E.: The Panama disease. III. Ann. Bot. 25:343-361, pi. 28-29. 
 
 1911. 
 
 (15) Fawcett, G. L. : Report of the plant pathologist. Porto Rico Agr. Expt. Sta. 
 
 Ann. Rpt, 1910: 35-36. 1911. 
 
 (16) Fawcett, G. L. : A Porto Rican disease of bananas. Porto Rico Agr. Expt. 
 
 Sta. Rpt, 1915:36-41. 1916. 
 
 (17) Higgins, J. E.: The banana in Hawaii. Hawaii Agr. Expt. Sta. Bui. 7, 53 p., 
 
 illus. 1904. 
 
 (18) Johnston, J. R. : La enfermedad del platano en Cuba. Cuba Estac. Expt. 
 
 Agron. Circ. 47, 13 p., 7 pi. 1915. 
 
 (19) Lathrop, E. C. : The generation of aldehydes by Fusarium cubense. Phyto- 
 
 pathology, 7: 14-16. 1917. 
 
 (20) Laurent, J. : Recherches sur la nutrition carbonee des plantes vertes a l'aide 
 
 de matieres organiques. 126 p., 7 pi. Lille. 1903. 
 Reprinted from Rev. Gen. Bot,, v. 16. 
 
 (21) Levy, H. Q. : The banana disease and other enemies of the plant in Costa 
 
 Rica. Jour. Jamaica Agr. Soc. 14:241-247. 1910. 
 ^22) McCubbin, W. A. : Dispersal distance of urediniospores of Cronartium ribicola 
 rate of fall in still air. Phytopathology 8: 53- 
 
1919] Brandes: Banana Wilt 389 
 
 (25) Prescott, S. C. : The banana: A food of exceptional value. Sci. No. 6: 65-75. 
 
 1918. 
 
 (26) Rijks, A. B. : Report over een Onderzoeck naar de pisangsterfte op de Saleier- 
 
 eilanden. Dept. Landb., Nijv. en Handel [Dutch East Indies], Meded. 
 Lab. Plantenziekten, no. 21, 16 p. 1915. 
 
 (27) Shaw, J. F.: Mycology. Ann. Rept. Bd. Sci. Advice India 1914-15: 104-109. 
 
 1916. 
 
 (28) Smith, E..F.: A Cuban banana disease. -(Abstract.) Science, n. s. 31:754- 
 
 755. 1910. 
 
 (29) Tryon, Henry: The banana in Queensland. Diseases of the banana. Queens- 
 
 land Agr. Jour. 28: 116-119. 1912. 
 
 (30) Werkenthin, F. C. : Fungus flora of Texas soils. Phytopathology 6: 241-253. 
 
 1916. 
 
 (31) Wilcox, E. V.: Summary of investigations. Hawaii Agr. Exp. St$. Rpt. 
 
 1914:7-24. 1915. 
 
 (32) Wilson, J. K.: Calcium hypochlorite as a seed sterilizer. Amer. Jour. Bot. 
 
 '2:420-427. 1915. 
 Bibliography, p. 426^27. 
 
 (33) Wollenweber, H. W. : Studies on the fusarium problem. Phytopathology 
 
 3:24-50., pi. 5. 1913. 
 
Plate XXII 
 
 Fig. 1. Cross section of healthy banana pseudostem. 
 Fig. 2. Cross section of healthy banana rhizome. 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXII 
 
 Brandes: Banana Wilt 
 
Plate XXIII 
 
 Fig. 1. Cross section of diseased banana plant. The plane of the section 
 includes the upper part of the rhizome (at the center) and a few of the sheathing 
 leaf bases which form the pseudostem (at the periphery). 
 
 Fig. 2. Cross section of diseased banana rhizome. 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXIII 
 
 Brandes : Banana Wilt 
 
Plate XXIV 
 
 Fig. 1. Banana plant killed by wilt organism after starting to produce a bunch 
 of fruit. Apparently healthy sucker arising from the old stool. 
 
 Fig. 2. Last stage of the disease. Plants of the Chamaluco variety in Porto 
 Rico. 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXIV 
 
 SMsk 
 
 Br4ndes: Banana Wilt 
 
Plate XXV 
 
 Fig. 1. Longitudinal splitting of the pseudostem. Manzana variety in Cuba. 
 Fig. 2. Longitudinal splitting of the pseudostem. Gros Michel variety in 
 Costa Rica. 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXV 
 
 K 
 P 
 
 < 
 M 
 
 P3 
 
Plate XXVI 
 
 Fig. 1. Longitudinal splitting of the pseudostem. Chainaluco variety in Porto 
 Rico. 
 
 Fig. 2. Longitudinal splitting of pseudostem produced by artifical inoculation. 
 Chamaluco variety in Porto Rico. 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXVI 
 
 «i 
 
 Brandes: Banana Wilt 
 
Plate XXVII 
 
 Fig. 1. Longitudinal section of banana rhizome and pseudostem showing dis- 
 colored leaf traces in the cortex, and diseased roots passing from the diseased stele 
 across the otherwise healthy cortex and out into the soil. 
 
 Fig. 2. Longitudinal section of banana rhizome and pseudostem, showing in- 
 dividual diseased vascular bundle passing from the stele at A, through the endo- 
 dermis at B, and traversing the cortex B to C, thence passing up into the leaf base. 
 Notice that the cortical tissue other than the diseased strand is not invaded, the 
 fungus being confined to the vessels of the vascular bundles. 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXVII 
 
 < 
 
 Q 
 
 < 
 « 
 
 23 
 
Plate XXVIII 
 
 Fig. 1. Transverse section of rhizome showing general infection. Gros Michel 
 variety, Costa Rica. 
 
 Fig. 2. Transverse section of rhizome showing late stage of the disease. 
 Secondary rots have set in. Notice the still healthy appearance of the cortex. 
 Chamaluco variety, Porto Rico. 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXVIII 
 
 Brandes: Banana Wilt 
 
Plate XXIX 
 
 Fig. 1. Transverse section of pseudostem o~ Chamaluco variety in Porto Rico 
 Notice freedom from disease of central and outer portions. The diseased portion is 
 represented by the band of white mycelium concentric with the periphery. This 
 piece of tissue was kept in a moist chamber for three days before the photograph 
 was taken. 
 
 Fig. 2. Transverse section of rhizome showing local infection of stele through 
 a root. Gros Michel variety, Jamaica. 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXIX 
 
 
Plate XXX 
 
 Result of experiment on inoculation of soil with pure cultures of Fusarium 
 cubense. Row on left inoculated, row on right not inoculated. Time, eight months 
 after planting. 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXX 
 
 I . 
 
Plate XXXI 
 
 Fig. 1. Macroconidia and microconidia of Fusarium cubense stained' with [Bis- 
 marck brown. 
 
 Fig. 2. Radial section of young fleshy root showing method of penetration of the 
 fungus. At the top are fragments of root hairs which do not lie in the plane of the 
 section. 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXXI 
 
 Brandes: Banana Wilt 
 
Plate XXXII 
 
 Fig. 1. Macroconidia germinating in distilled water. The germ tubes arise from 
 any cell and are here long and attenuated. 
 
 Fig. 2. Transverse section of a vessel in the root. Notice luxuriant vegetative 
 growth of the fungus. 
 
PHYTOPATHOLOGY, IX 
 
 
 PLATE XXXII 
 
 \ 
 
 V 
 
 I --^PM 
 
 Brandes: Banana Wilt 
 
Plate XXXIII 
 
 Fig. 1. Transverse section of vessel in the stele of a diseased rhizome. The 
 fungus is not so abundant here as in the vessels of the root near the point of 
 infection. 
 
 Fig. 2. Longitudinal section of a diseased vessel in the pseudostem. Notice the 
 scant mvcelium. 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXXIII 
 
 Brandes: Banana Wilt 
 
Plate XXXIV 
 
 Fig. 1. Tangential section of leaf stalk showing subepidermal cells filled with 
 mycelium of the parasite. A substomatal cavity filled with pseudoparenchymatous 
 tissue is present in the section. 
 
 Fig. 2. Radial section of a sporodolchium at surface of the leaf stalk. 
 
PHYTOPATHOLOGY, IX 
 
 PLATE XXXIV 
 
 
 
 Brandes - Banana Wilt 
 

 LIBRARY OF CONGRESS 
 
 Mil III I II II III II 
 
 015 793 543 4