Origin of Cultivated Plants

The aspect of the question as to the origin of the melon has completely changed since the experiments of Naudin. The paper which he published in 1859, in the Annales des Sciences Naturelles, 4th series, vol. ii., on the genus Cucumis, is as remarkable as that on the genus Cucurbita. He gives an account of the observations and experiments of several years on the variability of forms and the crossed fecundation of a multitude of species, breeds, or varieties coming from all parts of the world. I have already spoken (p. 250) of the physiological principle on which he believes it possible to distinguish those groups of forms which he terms species, although certain exceptions have occurred which render the criterion of fertilization less absolute. In spite of these exceptional cases, it is evident that if nearly allied forms can be easily crossed and produce fertile individuals, as we see, for example, in the human species, they must be considered as constituting a single species.

In this sense Cucumis Melo, according to the experiments and observations made by Naudin upon about two thousand living plants, constitutes a species which comprehends an extraordinary number of varieties and even of breeds; that is to say, forms which are preserved by heredity. These varieties or races can be fertilized by each other, and yield varied and variable products. They are classed by the author into ten groups, which he calls canteloups, melons brodés, sucrins, melons d’hiver, serpents, forme de concombre, Chito, Dudaim, rouges de Perse, and sauvages, each containing varieties or nearly allied races. These have been named in twenty-five or thirty different ways by botanists, who, without noticing transitions of form, the faculty of crossing or of change under cultivation, have distinguished as species all the varieties which occur in a given time or place.

Hence it results that several forms found wild, and which have been described as species, must be the types and sources of the cultivated forms; and Naudin makes the very just observation that these wild forms, which differ more or less the one from the other, may have produced different cultivated varieties. This is the more probable that they sometimes inhabit countries remote from each other as Southern Asia and tropical Africa, so that differences in climate and isolation may have created and consolidated varieties.

The following are the forms which Naudin enumerates as wild: 1. Those of India, which are named by Wildenow Cucumis pubescens, and by Roxburgh C. turbinatus or C. maderas-patanus. The whole of British India and Beluchistan is their natural area. Its natural wildness is evident even to non-botanical travellers.[1273 - Gardener’s Chronicle, articles signed “I. H. H.,” 1857, p. 153; 1858, p. 130.] The fruit varies from the size of a plum to that of a lemon. It is either striped or barred, or all one colour, scented or odourless. The flesh is sweet, insipid, or slightly acid, differences which it has in common with the cultivated Cantelopes. According to Roxburgh the Indians gather and have a taste for the fruits of C. turbinatus and of C. maderas-patanus, though they do not cultivate it.

Referring to the most recent flora of British India, in which Clarke has described the Cucurbitaceæ (ii. p. 619), it seems that this author does not agree with M. Naudin about the Indian wild forms, although both have examined the numerous specimens in the herbarium at Kew. The difference of opinion, more apparent than real, arises from the fact that the English author attributes to a nearly and certainly wild allied species, C. trigonus, Roxburgh, the varieties which Naudin classes under C. Melo. Cogniaux,[1274 - Cogniaux, Monogr. Phanér., iii. p. 485.] who afterwards saw the same specimens, attributes only C. turbinatus to trigonus. The specific difference between C. Melo and C. trigonus is unfortunately obscure, from the characters given by these three authors. The principal difference is that C. Melo is an annual, the other perennial, but this duration does not appear to be very constant. Mr. Clarke says himself that C. Melo is perhaps derived by cultivation from C. trigonus; that is to say, according to him, from the forms which Naudin attributes to C. Melo.

The experiments made during three consecutive years by Naudin[1275 - Naudin, Ann. Sc. Nat., 4th series, vol. xviii. p. 171.] upon the products of Cucumis trigonus, fertilized by C. Melo, seem in favour of the opinion which admits a specific diversity; for if fertilization took place the products were of different forms, and often reverted to one or other of the original parents.

2. The African forms. Naudin had no specimens in sufficiently good condition, or of which the wild state was sufficiently certain to assert positively the habitation of the species in Africa. He admits it with hesitation. He includes in the species cultivated forms, or other wild ones, of which he had not seen the fruit. Sir Joseph Hooker[1276 - Hooker, in Oliver, Fl. of Trop. Afr., ii. p. 546.] subsequently obtained specimens which prove more. I am not speaking of those from the Nile Valley,[1277 - Schweinfurth and Ascherson, Aufzählung, p. 267.] which are probably cultivated, but of plants gathered by Barter in Guinea in the sands on the banks of the Niger. Thonning[1278 - Schumacher and Thonning, Guineiske Planten., p. 426.] had previously found, in sandy soil in Guinea, a Cucumis to which he had given the name arenarius; and Cogniaux,[1279 - Cogniaux, in de Candolle, Monogr. Phanér., p. 483.] after having seen a specimen brought home by this traveller, had classed it with C. Melo, as Sir J. Hooker thought. The negroes eat the fruit of the plant found by Barter. The smell is that of a fresh green melon. In Thonning’s plant the fruit is ovoid, the size of a plum. Thus in Africa as in India the species bears small fruit in a wild state, as we might expect. The Dudaim among cultivated varieties is allied to it.

The majority of the species of the genus Cucumis are found in Africa; a small minority in Asia or in America. Other species of Cucurbitaceæ are divided between Asia and America, although as a rule, in this family, the areas of species are continuous and restricted. Cucumis Melo was once perhaps, like Citrullus Colocynthis of the same family, wild from the west coast of Africa as far as India without any break.

I formerly hesitated to admit that the melon was indigenous in the north of the Caucasus, as it is asserted by ancient authors – an assertion which has not been confirmed by subsequent botanists. Hohenacker, who was said to have found the species near Elisabethpolis, makes no mention of it in his paper upon the province of Talysch. M. Boissier does not include Cucumis Melo in his Oriental flora. He merely says that it is easily naturalized on rubbish-heaps and waste ground. The same thing has been observed elsewhere, for instance in the sands of Ussuri, in Eastern Asia. This would be a reason for mistrusting the locality of the sands of the Niger, if the small size of the fruit in this case did not recall the wild forms of India.

The culture of the melon, or of different varieties of the melon, may have begun separately in India and Africa.

Its introduction into China appears to date only from the eighth century of our era, judging from the epoch of the first work which mentions it.[1280 - Bretschneider, letter of Aug. 26, 1881.] As the relations of the Chinese with Bactriana, and the north-west of India by the embassy of Chang-kien, date from the second century, it is possible that the culture of the species was not then widely diffused in Asia. The small size of the wild fruit offered little inducement. No Sanskrit name is known, but there is a Tamul name, probably less ancient, molam,[1281 - Piddington, Index.] which is like the Latin Melo.

It is not proved that the ancient Egyptians cultivated the melon. The fruit figured by Lepsius[1282 - See the copy in Unger’s Pflanzen des Alten Ægyptens, fig. 25.] is not recognizable. If the cultivation had been customary and ancient in that country, the Greeks and Romans would have early known it. Now, it is doubtful whether the Sikua of Hippocrates and Theophrastus, or the Pepon of Dioscorides, or the Melopepo of Pliny, was the melon. The passages referring to it are brief and insignificant; Galen[1283 - Galen, De Alimentis, l. 2, c. 5.] is less obscure, when he says that the inside of the Melopepones is eaten, but not of the Pepones. There has been much discussion about those names,[1284 - See all the Vergilian floras, and Naudin, Ann. Sc. Nat., 4th series, vol. xii. p. 111.] but we want facts more than words. The best proof which I have been able to discover of the existence of the melon among the Romans is a very accurate representation of a fruit in the beautiful mosaic of fruits in the Vatican. Moreover, Dr. Comes certifies that the half of a melon is represented in a painting at Herculaneum.[1285 - Comes, Ill. Piante nei Dipinti Pompeiani, in 4to, p. 20, in the Museo Nation., vol. iii. pl. 4.] The species was probably introduced into the Græco-Roman world at the time of the Empire, in the beginning of the Christian era. It was probably of indifferent quality, to judge from the silence or the faint praise of writers in a country where gourmets were not wanting. Since the Renaissance, an improved cultivation and relations with the East have introduced better varieties into our gardens. We know, however, that they often degenerate either from cold or bad conditions of soil, or by crossing with inferior varieties of the species.

Water-Melon—Citrullus vulgaris, Schrader; Cucurbita Citrullus, Linnæus.

The origin of the water-melon was long mistaken or unknown. According to Linnæus, it was a native of Southern Italy.[1286 - Habitat in Apulia, Calabria, Sicilia (Linnæus, Species, edit. 1763, p. 1435).] This assertion was taken from Matthiole, without observing that this author says it was a cultivated species. Seringe,[1287 - Seringe, in Prodromus, iii. p. 301.] in 1828, supposed it came from India and Africa, but he gives no proof. I believed it came from Southern Asia, because of its very general cultivation in this region. It was not known in a wild state. At length it was found indigenous in tropical Africa, on both sides of the equator, which settles the question.[1288 - Naudin, Ann. sc. Nat., 4th series, vol. xii. p. 101; Sir J. Hooker, in Oliver, Flora of Trop. Afr., ii. p. 549.] Livingstone[1289 - French trans., p. 56.] saw districts literally covered with it, and the savages and several kinds of wild animals eagerly devoured the wild fruit. They are sometimes, but not always, bitter, and this cannot be detected from the appearance of the fruit. The negroes strike it with an axe, and taste the juice to see whether it is good or bad. This diversity in the wild plant, growing in the same climate and in the same soil, is calculated to show the small value of such a character in cultivated Cucurbitaceæ. For the rest, the frequent bitterness of the water-melon is not at all extraordinary, as the most nearly allied species is Citrullus Colocynthis. Naudin obtained fertile hybrids from crossing the bitter water-melon, wild at the Cape, with a cultivated species which confirms the specific unity suggested by the outward appearance.

The species has not been found wild in Asia.

The ancient Egyptians cultivated the water-melon, which is represented in their paintings.[1290 - Unger has copied the figures from Lepsius’ work in his memoir Die Pflanzen des Alten Ægyptens, figs. 30, 31, 32.] This is one reason for believing that the Israelites knew the species, and called it abbatitchim, as is said; but besides the Arabic name, battich, batteca, evidently derived from the Hebrew, is the modern name for the water-melon. The French name, pastèque, comes through the Arabic from the Hebrew. A proof of the antiquity of the plant in the north of Africa is found in the Berber name, tadelaât,[1291 - Dictionnaire Français-Berber, at the word pastèque.] which differs too widely from the Arabic name not to have existed before the Conquest. The Spanish names zandria, cindria, and the Sardinian sindria,[1292 - Moris, Flora Sardoa.] which I cannot connect with any others, show also an ancient culture in the eastern part of the Mediterranean basin. Its cultivation early spread into Asia, for there is a Sanskrit name, chayapula,[1293 - Piddington, Index.] but the Chinese only received the plant in the tenth century of the Christian era. They call it si-kua, that is melon of the West.[1294 - Bretschneider, Study and Value, etc., p. 17.]

As the water-melon is an annual, it ripens out of the tropics wherever the summer is sufficiently hot. The modern Greeks cultivate it largely, and call it carpousia or carpousea,[1295 - Heldreich, Pflanz. d. Attisch. Ebene., p. 591; Nutzpfl. Griechenl., p. 50.] but this name does not occur in ancient authors, nor even in the Greek of the decadence and of the Middle Ages.[1296 - Langkavel, Bot. der Spät. Griechen.] It is the same as the karpus of the Turks of Constantinople,[1297 - Forskal, Flora Ægypto-Arabica, part i. p. 34.] which we find again in the Russian arbus,[1298 - Nemnich, Polyg. Lexic., i. p. 1309.] and in Bengali and Hindustani as tarbuj, turbouz.[1299 - Piddington, Index; Pickering, Chronol. Arrang., p. 72.] Another Constantinople name, mentioned by Forskal, chimonico, recurs in Albanian chimico.[1300 - Heldreich, Nutzpfl., etc., p. 50.] The absence of an ancient Greek name which can with certainty be attributed to this species, seems to show that it was introduced into the Græco-Roman world about the beginning of the Christian era. The poem Copa, attributed to Virgil and Pliny, perhaps mentions it (lib. 19, cap. 5), as Naudin thinks, but it is doubtful.

Europeans have introduced the water-melon into America, where it is now cultivated from Chili to the United States. The jacé of the Brazilians, of which Piso and Marcgraf have a drawing, is evidently introduced, for the first-named author says it is cultivated and partly naturalized.[1301 - “Sativa planta et tractu temporis quasi nativa facta” (Piso, edit. 1658, p. 233).]

Cucumber—Cucumis sativus, Linnæus.

In spite of the very evident difference between the melon and cucumber, which both belong to the genus Cucumis, cultivators suppose that the species may be crossed, and that the quality of the melon is thus sometimes spoilt. Naudin[1302 - Naudin, in Ann. Sc. Nat., 4th series, vol. xi. p. 31.] ascertained by experiments that this fertilization is not possible, and has also shown that the distinction of the two species is well founded.

The original country of Cucumis sativus was unknown to Linnæus and Lamarck. In 1805, Wildenow[1303 - Wildenow, Species, iv. p. 615.] asserted it was indigenous in Tartary and India, but without furnishing any proof. Later botanists have not confirmed the assertion. When I went into the question in 1855, the species had not been anywhere found wild. For various reasons deduced from its ancient culture in Asia and in Europe, and especially from the existence of a Sanskrit name, soukasa,[1304 - Piddington, Index.] I said, “Its original habitat is probably the north-west of India, for instance Cabul, or some adjacent country. Everything seems to show that it will one day be discovered in these regions which are as yet but little known.”

This conjecture has been realized if we admit, with the best-informed modern authors, that Cucumis Hardwickii, Royle, possesses the characteristics of Cucumis sativus. A coloured illustration of this cucumber found at the foot of the Himalayas may be seen in Royle’s Illustrations of Himalayan Plants, p. 220, pl. 47. The stems, leaves, and flowers are exactly those of C. sativus. The fruit, smooth and elliptical, has a bitter taste; but there are similar forms of the cultivated cucumber, and we know that in other species of the same family, the water-melon, for instance, the pulp is sweet or bitter. Sir Joseph Hooker, after describing the remarkable variety which he calls the Sikkim cucumber,[1305 - Bot. Mag., pl. 6206.] adds that the variety Hardwickii, wild from Kumaon to Sikkim, and of which he has gathered specimens, does not differ more from the cultivated plant than certain varieties of the latter differ from others; and Cogniaux, after seeing the plants in the herbarium at Kew, adopts this opinion.[1306 - Cogniaux, in de Candolle, Monogr. Phanér., iii. p. 499.]

The cucumber, cultivated in India for at least three thousand years, was only introduced into China in the second century before Christ, when the ambassador Chang-kien returned from Bactriana.[1307 - Bretschneider, letters of Aug. 23 and 26, 1881.] The species spread more rapidly towards the West. The ancient Greeks cultivated the cucumber under the name of sikuos,[1308 - Theophrastus, Hist., lib. 7, cap. 4; Lenz, Bot. der Alten, p. 492.] which remains as sikua in the modern language. The modern Greeks have also the name aggouria, from an ancient Aryan root which is sometimes applied to the water-melon, and which recurs for the cucumber in the Bohemian agurka, the German Gurke, etc. The Albanians (Pelasgians?) have quite a different name, kratsavets,[1309 - Heldreich, Nutzpfl. Griechen., p. 50.] which we recognize in the Slav Krastavak. The Latins called the cucumber cucumis. These different names show the antiquity of the species in Europe. There is even an Esthonian name, uggurits, ukkurits, urits.[1310 - Nemnich, Polygl. Lex., i. p. 1306.] It does not seem to be Finnish, but to belong to the same Aryan root as aggouria. If the cucumber came into Europe before the Aryans, there would perhaps be some name peculiar to the Basque language, or seeds would have been found in the lake-dwellings of Switzerland and Savoy; but this is not the case. The peoples in the neighbourhood of the Caucasus have names quite different to the Greek; in Tartar kiar, in Kalmuck chaja, in Armenian karan.[1311 - Nemnich, ibid.] The name chiar exists also in Arabic for a variety of the cucumber.[1312 - Forskal, Fl. Ægypt., p. 76.] This is, therefore, a Turanian name anterior to the Sanskrit, whereby its culture in Western Asia would be more than three thousand years old.

It is often said that the cucumber is the kischschuim, one of the fruits of Egypt regretted by the Israelites in the desert.[1313 - Rosenmüller, Biblische Alterth., i. p. 97; Hamilton, Bot. de la Bible, p. 34.] However, I do not find any Arabic name among the three given by Forskal which can be connected with this, and hitherto no trace has been found of the presence of the cucumber in ancient Egypt.

West Indian Gherkin—Cucumis Anguria, Linnæus.

This small species of cucumber is designated in the Bon Jardinier under the name of the cucumber Arada. The fruit, of the size of an egg, is very prickly. It is eaten cooked or pickled. As the plant is very productive, it is largely cultivated in the American colonies. Descourtilz and Sir Joseph Hooker have published good coloured illustrations of it, and M. Cogniaux a plate with a detailed analysis of the flower.[1314 - Descourtilz, Fl. Méd. des Antilles, v. pl. 329; Hooker, Bot. Mag., t. 5817; Cogniaux, in Fl. Brasil., fasc. 78, pl. 2.]

Several botanists affirm that it is wild in the West Indies. P. Browne,[1315 - Browne, Jamaica, edit. 2, p. 353.] in the last century, spoke of the plant as the “little wild cucumber” (in Jamaica). Descourtilz said, “The cucumber grows wild everywhere, and principally in the dry savannahs and near rivers, whose banks afford a rich vegetation.” The inhabitants call it the “maroon cucumber.” Grisebach[1316 - Grisebach, Fl. of Brit. W. India Is., p. 288.] saw specimens in several other West India Isles, and appears to admit their wild character. M. E. André found the species growing in the sand of the sea-shore at Porto-Cabello, and Burchell in a similar locality in Brazil, and Riedel near Rio di Janeiro.[1317 - Cogniaux, ubi supra.] In the case of a number of other specimens gathered in the east of America from Brazil to Florida, it is unknown whether they were wild or cultivated. A wild Brazilian plant, badly drawn by Piso,[1318 - Guanerva-oba, in Piso, Brasil., edit. 1658, p. 264; Marcgraf, edit. 1648, p. 44, without illustration, calls it Cucumis sylvestris Brasiliæ.] is mentioned as belonging to the species, but I am very doubtful of this.

Botanists from Tournefort down to our own day have considered the Anguria to be of American origin, a native of Jamaica in particular. M. Naudin[1319 - Naudin, Ann. Sc. Nat., 4th series, vol. ii. p. 12.] was the first to point out that all the other species of Cucumis are of the old world, and principally African. He wondered whether this one had not been introduced into America by the negroes, like many other plants which have become naturalized. However, unable to find any similar African plant, he adopted the general opinion. Sir Joseph Hooker, on the contrary, is inclined to believe that C. Anguria is a cultivated and modified form of some African species nearly allied to C. prophetarum and C. Figarei, although these are perennial. In favour of this hypothesis, I may add: (1) The name maroon cucumber, given in the French West India Islands, indicates a plant which has become wild, for this is the meaning of the word maroon as applied to the negroes; (2) its extended area in America from Brazil to the West Indies, always along the coast where the slave trade was most brisk, seems to be a proof of foreign origin. If the species grew in America previous to its discovery, it would, with such an extensive habitat, have been also found upon the west coast of America, and inland, which is not the case.

The question can only be solved by a more complete knowledge of the African species of Cucumis, and by experiments upon fertilization, if any have the patience and ability necessary to do for the genus Cucumis what Naudin has done for the genus Cucurbita.

Lastly, I would point out the absurdity of a common name for the Anguria in the United States —Jerusalem Cucumber.[1320 - Darlington, Agric. Bot., p. 58.] After this, is it possible to take popular names as a guide in our search for origins?

White Gourd-melon, or Benincasa—Benincasa hispida, Thunberg; Benincasa cerifera, Savi.

This species, which is the only one of the genus Benincasa, is so like the pumpkins that early botanists took it for one,[1321 - Cucurbita Pepo of Loureiro and Roxburgh.] in spite of the waxy efflorescence on the surface of the fruit. It is very generally cultivated in tropical countries. It was, perhaps, a mistake to abandon its cultivation in Europe after having tried it, for Naudin and the Bon Jardinier both recommend it.

It is the cumbalam of Rheede, the camolenga of Rumphius, who had seen it cultivated in Malabar and the Sunda Islands, and give illustrations of it.

From several works, even recent ones,[1322 - Clarke, in Fl. of Brit. Ind., ii. p. 616.] it might be supposed that it had never been found in a wild state, but if we notice the different names under which it has been described we shall find that this is not the case. Thus Cucurbita hispida, Thunberg, and Lagenaria dasystemon, Miquel, from authentic specimens seen by Cogniaux,[1323 - Cogniaux, in de Candolle, Monogr. Phanér., iii. p. 513.] are synonyms of the species, and these plants are wild in Japan.[1324 - Thunberg, Fl. Jap., p. 322; Franchet and Savatier, Enum. Pl. Jap., i. p. 173.]Cucurbita littoralis, Hasskarl,[1325 - Hasskarl, Catal. Horti. Bogor. Alter., p. 190; Miquel, Flora Indo-Batav.] found among shrubs on the sea-shore in Java, and Gymnopetalum septemlobum, Miquel, also in Java, are the Benincasa according to Cogniaux. As are also Cucurbita vacua, Mueller,[1326 - Mueller, Fragm., vi. p. 186; Forster, Prodr. (no description); Seemann, Jour. of Bot., ii. p. 50.] and Cucurbita pruriens, Forster, of which he has seen authentic specimens found at Rockingham, in Australia, and in the Society Islands. Nadeaud[1327 - Nadeaud, Plan. Usu. des Taitiens, Enum. des Pl. Indig. à Taiti.] does not mention the latter. Temporary naturalization may be suspected in the Pacific Isles and in Queensland, but the localities of Java and Japan seem quite certain. I am the more inclined to believe in the latter, that the cultivation of the Benincasa in China dates from the remotest antiquity.[1328 - Bretschneider, letter of Aug. 26, 1881.]

Towel Gourd—Momordica cylindrica, Linnæus; Luffa cylindrica, Rœmer.

Naudin[1329 - Naudin, Ann. Sc. Nat., 4th series, vol. xii. p. 121.] says, “Luffa cylindrica, which in some of our colonies has retained the Indian name pétole, is probably a native of Southern Asia, and perhaps also of Africa, Australia, and Polynesia. It is cultivated by the peoples of most hot countries, and it appears to be naturalized in many places where it doubtless did not exist originally.” Cogniaux[1330 - Cogniaux, Monogr. Phanér., iii. p. 458.] is more positive. “An indigenous species,” he says, “in all the tropical regions of the old world; often cultivated and half wild in America between the tropics.” In consulting the works quoted in these two monographs, and herbaria, its character as a wild plant will be found sometimes conclusively certified.

With regard to Asia,[1331 - Rheede, Hort. Malab., viii. p. 15, t. 8; Roxburgh, Fl. Ind., iii. p. 714, as L. clavata; Kurz, Contrib., ii. p. 100; Thwaites, Enum.] Rheede saw it in sandy places, in woods and other localities in Malabar; Roxburgh says it is wild in Hindustan; Kurz, in the forests of Burmah; Thwaites, in Ceylon. I have specimens from Ceylon and Khasia. There is no Sanskrit name known, and Dr. Bretschneider, in his work On the Study and Value of Chinese Botanical Works, and in his letters mentions no luffa either wild or cultivated in China. I suppose, therefore, that its cultivation is not ancient even in India.

The species is wild in Australia, on the banks of rivers in Queensland,[1332 - Mueller, Fragmenta, iii. p. 107; Bentham, Fl. Austr., iii. p. 317, under names which Naudin and Cogniaux regard as synonyms of L. cylindrica.] and hence it is probable it will be found wild in the Asiatic Archipelago, where Rumphius, Miquel, etc., only mention it as a cultivated plant.

Herbaria contain a great number of specimens from tropical Africa, from Mozambique to the coast of Guinea, and even as far as Angola, but collectors do not appear to have indicated whether they were cultivated or wild plants. In the Delessert herbarium, Heudelot indicates it as growing in fertile ground in the environs of Galam. Sir Joseph Hooker[1333 - Hooker, in Oliver, Fl. of Trop. Afr., ii. p. 530.] quotes this without affirming anything. Schweinfurth and Ascheron,[1334 - Schweinfurth and Ascheron, Aufzählung, p. 268.] who are always careful in this matter, say the species is only a cultivated one in the Nile Valley. This is curious, because the plant was seen in the seventeenth century in Egyptian gardens under the Arabian name of luff,[1335 - Forskal, Fl. Ægypt., p. 75.] whence the genus was called Luffa, and the species Luffa ægyptica. The ancient Egyptian monuments show no trace of it. The absence of a Hebrew name is another reason for believing that its cultivation was introduced into Egypt in the Middle Ages. It is now grown in the Delta, not only for the fruit but also for the export of the seed, from which a preparation is made for softening the skin.

The species is cultivated in Brazil, Guiana, Mexico, etc., but I find no indication that it is indigenous in America. It appears to have been here and there naturalized, in Nicaragua for instance, from a specimen of Levy’s.

In brief, the Asiatic origin is certain, the African very doubtful, that of America imaginary, or rather the effect of naturalization.

Angular Luffa—Luffa acutangula, Roxburgh.

The origin of this species, cultivated like the preceding one in all tropical countries, is not very clear, according to Naudin and Cogniaux.[1336 - Naudin, Ann. Sc. Nat., 4th series, vol. xii. p. 122; Cogniaux, in de Candolle, Monogr. Phanér., iii. p. 459.] The first gives Senegal, the second Asia, and, doubtfully, Africa. It is hardly necessary to say that Linnæus[1337 - Linnæus, Species, p. 1436, as Cucumis acutangulus.] was mistaken in indicating Tartary and China. Clarke, in Sir Joseph Hooker’s flora, says without hesitation that it is indigenous in British India. Rheede[1338 - Rheede, Hort. Malab., viii. p. 13, t. 7.] formerly saw the plant in sandy soil in Malabar. Its natural area seems to be limited, for Thwaites in Ceylon, Kurz in British Burmah, and Loureiro in China and Cochin-China,[1339 - Thwaites, Enum. Ceylan, p. 126; Kurz, Contrib., ii. p. 101; Loureiro, Fl. Cochin., p. 727.] only give the species as cultivated, or growing on rubbish-heaps near gardens. Rumphius[1340 - Rumphius, Amboin, v. p. 408, t. 149.] calls it a Bengal plant. No luffa has been long cultivated in China, according to a letter of Dr. Bretschneider. No Sanskrit name is known. All these are indications of a comparatively recent culture in Asia.

A variety with bitter fruit is common in British India[1341 - Clarke, in Fl. Brit. Ind., ii. p. 614.] in a wild state, since there is no inducement to cultivate it. It exists also in the Sunda Islands. It is Luffa amara, Roxburgh, and L. sylvestris, Miquel. L. subangulata, Miquel, is another variety which grows in Java, which M. Cogniaux also unites with the others from authentic specimens which he saw.

M. Naudin does not say what traveller gives the plant as wild in Senegambia; but he says the negroes call it papengaye, and as this is the name of the Mauritius planters,[1342 - Bojer, Hort. Maurit.] it is probable that the plant is cultivated in Senegal, and perhaps naturalized near dwellings. Sir Joseph Hooker, in the Flora of Tropical Africa, gives the species, but without proof that it is wild in Africa, and Cogniaux is still more brief. Schweinfurth and Ascheron[1343 - Schweinfurth and Ascherson, Aufzählung, p. 268.] do not mention it either as wild or cultivated in Egypt, Nubia, and Abyssinia. There is no trace of its ancient cultivation in Egypt.

The species has often been sent from the West Indies, New Granada, Brazil, and other parts of America, but there is no indication that it has been long in these places, nor even that it occurs at a distance from gardens in a really wild state.

The conditions or probabilities of origin, and of date of culture, are, it will be seen, identical for the two cultivated species of luffa. In support of the hypothesis that the latter is not of African origin, I may say that the four other species of the genus are Asiatic or American; and as a sign that the cultivation of the luffa is not very ancient, I will add that the form of the fruit varies much less than in the other cultivated cucurbitacea.

Snake Gourd—Trichosanthes anguina, Linnæus.

An annual creeping Cucurbitacea, remarkable for its fringed corolla. It is called petole in Mauritius, from a Java name. The fruit, which is something like a long fleshy pod of some leguminous plants, is eaten cooked like a cucumber in tropical Asia.

As the botanists of the seventeenth century received the plant from China, they imagined that the plant was indigenous there, but it was probably cultivated. Dr. Bretschneider[1344 - Bretschneider, Study and Value, etc., p. 17.] tells us that the Chinese name, mankua, means “cucumber of the southern barbarians.” Its home must be India, or the Indian Archipelago. No author, however, asserts that it has been found in a distinctly wild state. Thus Clarke, in Hooker’s Flora of British India, ii. p. 610, says only, “India, cultivated.” Naudin,[1345 - Naudin, Ann. Sc. Nat., 4th series, vol. xviii. p. 190.] before him, said, “Inhabits the East Indies, where it is much cultivated for its fruits. It is rarely found wild.” Rumphius[1346 - Rumphius, Amboin, v. pl. 148.] is not more positive for Amboyna. Loureiro and Kurz in Cochin-China and Burmah, Blume and Miquel in the islands to the south of Asia, have only seen the plant cultivated. The thirty-nine other species of the genus are all of the old world, found between China or Japan, the west of India and Australia. They belong especially to India and the Malay Archipelago. I consider the Indian origin as the most probable one.