Known facts and historical and philological probabilities tend to show that the species was wild from Cochin-China to the south of Japan and to Java when the ancient inhabitants of this region began to cultivate it at a very remote period, to use it for food in various ways, and to obtain from it varieties of which the number is remarkable, especially in Japan.

Pigeon-Pea—Cajanus indicus, Sprengel; Cytisus Cajan, Linnæus.

This leguminous plant, often grown in tropical countries, is a shrub, but it fruits in the first year, and in some countries it is grown as an annual. Its seed is an important article of the food of the negroes and natives, but the European colonists do not care for it unless cooked green like our garden-pea. The plant is easily naturalized in poor soil round cultivated plots, even in the West India Islands, where it is not indigenous.1672

In Mauritius it is called ambrevade; in the English colonies, doll, pigeon-pea; and in the French Antilles, pois d’Angola, pois de Congo, pois pigeon.

It is remarkable that, though the species is diffused in three continents, the varieties are not numerous. Two are cited, based only upon the yellow or reddish colour of the flower, which were formerly regarded as distinct species; but a more attentive examination has resulted in their being classed as one, in accordance with Linnæus’ opinion.1673 The small number of variations obtained even in the organ for which the species is cultivated is a sign of no very ancient culture. Its habitation previous to culture is uncertain. The best botanists have sometimes supposed it to be a native of India, sometimes of tropical Africa. Bentham, who has made a careful study of the leguminous plants, believed in 1861 in the African origin; in 1865 he inclined rather to Asia.1674 The problem is, therefore, an interesting one. There is no question of an American origin. The cajan was introduced into the West Indies from the coast of Africa by the slave trade, as the common names quoted above show,1675 and the unanimous opinion of authors or American floras. It has also been taken to Brazil, Guiana, and into all the warm parts of the American continent.

The facility with which the species is naturalized would alone prevent attaching great importance to the statements of collectors, who have found it more or less wild in Asia or in Africa; and besides, these assertions are not precise, but are usually doubtful. Most writers on the flora of continental India have only seen the plant cultivated,1676 and none, to my knowledge, affirms that it exists wild. For the island of Ceylon Thwaites says,1677 “It is said not to be really wild, and the country names seem to confirm this.” Sir Joseph Hooker, in his Flora of British India, says, “Wild (?) and cultivated to the height of six thousand feet in the Himalayas.” Loureiro1678 gives it as cultivated and non-cultivated in China and Cochin-China. Chinese authors do not appear to have spoken of it, for the species is not named by Bretschneider in his work On the Study, etc. In the Sunda Isles it is mentioned as cultivated, and that rarely, at Amboyna at the end of the eighteenth century, according to Rumphius.1679 Forster had not seen it in the Pacific Isles at the time of Cook’s voyages, but Seemann says that it has been recently introduced by missionaries into the Fiji Isles.1680 All this argues no very ancient extension of cultivation to the east and south of the continent of Asia. Besides the quotation from Loureiro, I find the species indicated on the mountain of Magelang, Java;1681 but, supposing this to be a true and ancient wild growth in both cases, it would be very extraordinary not to find the species in many other Asiatic localities.

The abundance of Indian and Malay names1682 shows a somewhat ancient cultivation. Piddington even gives a Sanskrit name, arhuku, which was not known to Roxburgh, but he gives no proof in support of his assertion. The name may have been merely supposed from the Hindu and Bengali names urur and orol. No Semitic name is known.

In Africa the cajan is often found from Zanzibar to the coast of Guinea.1683 Authors say it is cultivated, or else make no statement on this head, which would seem to show that the specimens are sometimes wild. In Egypt this cultivation is quite modern, of the nineteenth century.1684

Briefly, then, I doubt that the species is really wild in Asia, and that it has been grown there for more than three thousand years. If more ancient peoples had known it, it would have come to the knowledge of the Arabs and Egyptians before our time. In tropical Africa, on the contrary, it is possible that it has existed wild or cultivated for a very long time, and that it was introduced into Asia by ancient travellers trading between Zanzibar and India or Ceylon.

The genus Cajanus has only one species, so that no analogy of geographical distribution leads us to believe it to be rather of Asiatic than African origin, or vice versâ.

Carob Tree1685—Ceratonia siliqua, Linnæus.

The seeds and pods of the carob are highly prized in the hotter parts of the Mediterranean basin, as food for animals and even for man. De Gasparin1686 has given interesting details about the raising, uses, and habitation of the species as a cultivated tree. He notes that it does not pass the northern limit beyond which the orange cannot be grown without shelter. This fine evergreen tree does not thrive either in very hot countries, especially where there is much humidity. It likes the neighbourhood of the sea and rocky places. Its original country, according to Gasparin, is “probably the centre of Africa. Denham and Clapperton found it in Burnou.” This proof seems to me insufficient, for in all the Nile Valley and in Abyssinia the carob is not wild nor even cultivated.1687 R. Brown does not mention it in his account of Denham and Clapperton’s journey. Travellers have seen it in the forests of Cyrenaica between the high-lands and the littoral; but the able botanists who have drawn up the catalogue of the plants of this country are careful to say,1688 “perhaps indigenous.” Most botanists merely mention the species in the centre and south of the Mediterranean basin, from Spain and Marocco to Syria and Anatolia, without inquiring closely whether it is indigenous or cultivated, and without entering upon the question of its true country previous to cultivation. Usually they indicate the carob tree, as “cultivated and subspontaneous, or nearly wild.” However, it is stated to be wild in Greece by Heldreich, in Sicily by Gussone and Bianca, in Algeria by Munby;1689 and these authors have each lived long enough in the country for which each is quoted to form an enlightened opinion.

Bianca remarks, however, that the carob tree is not always healthy and productive in those restricted localities where it exists in Sicily, in the small adjacent islands, and on the coast of Italy. He puts forward the opinion, moreover, based upon the similarity of the Italian name carrubo with the Arabic word, that the species was anciently introduced into the south of Europe, the species being of Syrian or north African origin. He maintains as probable the theory of Hœfer and Bonné,1690 that the lotus of the lotophagi was the carob tree, of which the flower is sweet and the fruit has a taste of honey, which agrees with the expressions of Homer. The lotus-eaters dwelt in Cyrenaica, so that the carob must have been abundant in their country. If we admit this hypothesis we must suppose that Pliny and Herodotus did not know Homer’s plant, for the one describes the lotus as bearing a fruit like a mastic berry (Pistacia lentiscus), the other as a deciduous tree.1691

An hypothesis regarding a doubtful plant formerly mentioned by a poet can hardly serve as the basis of an argument upon facts of natural history. After all, Homer’s lotus plant perhaps existed only in the fabled garden of Hesperides. I return to more serious arguments, on which Bianca has said a few words.

The carob has two names in ancient languages—the one Greek, keraunia or kerateia;1692 the other Arabic, chirnub or charûb. The first alludes to the form of the pod, which is like a slightly curved horn; the other means merely pod, for we find in Ebn Baithar’s1693 work that four other leguminous plants bear the same name, with a qualifying epithet. The Latins had no special name; they used the Greek word, or the expression siliqua, siliqua græca (Greek pod).1694 This dearth of names is the sign of a once restricted area, and of a culture which probably does not date from prehistoric time. The Greek name is still retained in Greece. The Arab name persists among the Kabyles, who call the fruit kharroub, the tree takharrout,1695 and the Spaniards algarrobo. Curiously enough, the Italians also took the Arab name currabo, carubio, whence the French caroubier. It seems that it must have been introduced after the Roman epoch by the Arabs of the Middle Ages, when there was another name for it. These details are all in favour of Bianca’s theory of a more southern origin than Sicily. Pliny says the species belonged to Syria, Ionia, Cnidos, and Rhodes, but he does not say whether it was wild or cultivated in these places. Pliny also says that the carob tree did not exist in Egypt. Yet it has been recognized in monuments belonging to a much earlier epoch than that of Pliny, and Egyptologists even attribute two Egyptian names to it, kontrates or jiri.1696 Lepsius gives a drawing of a pod which appears to him to be certainly a carob, and the botanist Kotschy made certain by microscopic investigation that a stick taken from a sarcophagus was made from the wood of the carob tree.1697 There is no known Hebrew name for the species, which is not mentioned in the Old Testament. The New Testament speaks of it by the Greek name in the parable of the prodigal son. It is a tradition of the Christians in the East that St. John Baptist fed upon the fruit of the carob in the desert, and hence came the names given to it in the Middle Ages—bread of St. John, and Johannis brodbaum.

Evidently this tree became important at the beginning of the Christian era, and it spread, especially through the agency of the Arabs, towards the West. If it had previously existed in Algeria, among the Berbers, and in Spain, older names would have persisted, and the species would probably have been introduced into the Canaries by the Phœnicians.

The information gained on the subject may be summed up as follows:—

The carob grew wild in the Levant, probably on the southern coast of Anatolia and in Syria, perhaps also in Cyrenaica. Its cultivation began within historic time. The Greeks diffused it in Greece and Italy; but it was afterwards more highly esteemed by the Arabs, who propagated it as far as Marocco and Spain. In all these countries the tree has become naturalized here and there in a less productive form, which it is needful to graft to obtain good fruit.

The carob has not been found in the tufa and quaternary deposits of Southern Europe. It is the only one of its kind in the genus Ceratonia, which is somewhat exceptional among the Leguminosæ, especially in Europe. Nothing shows that it existed in the ancient tertiary or quaternary flora of the south-west of Europe.

Common Haricot Kidney Bean—Phaseolus vulgaris, Savi.

When, in 1855, I wished to investigate the origin of the genera Phaseolus and Dolichos,1698 the distinction of species was so little defined, and the floras of tropical countries so rare, that I was obliged to leave several questions on one side. Now, thanks to the works of Bentham and Georg von Martens,1699 completing the previous labours of Savi,1700 the Leguminæ of hot countries are better known; lastly, the seeds discovered quite recently in the Peruvian tombs of Ancon, examined by Wittmack, have completely modified the question of origin.

I will speak first of the common haricot bean, afterwards of some other species, without, however, enumerating all those which are cultivated, for several of these are still ill defined.

Botanists held for a long time that the common haricot was of Indian origin. No one had found it wild, nor has it yet been found, but it was supposed to be of Indian origin, although the species was also cultivated in Africa and America, in temperate and hot regions, at least in those where the heat and humidity are not excessive. I called attention to the fact that there is no Sanskrit name, and that sixteenth-century gardeners often called the species Turkish bean. Convinced, moreover, that the Greeks cultivated this plant under the names fasiolos and dolichos, I suggested that it came originally from Western Asia, and not from India. Georg von Martens adopted this hypothesis.

However, the meaning of the words dolichos of Theophrastus, fasiolos of Dioscorides, faseolus and phaseolus of the Romans,1701 is far from being sufficiently defined to allow them to be attributed with certainty to Phaseolus vulgaris. Several cultivated Leguminosæ are supported by the trellises mentioned by authors, and have pods and seeds of a similar kind. The best argument for translating these names by Phaseolus vulgaris is that the modern Greeks and Italians have names derived from fasiolus for the common haricot. In modern Greek it is fasoulia, Albanian (Pelasgic?) fasulé, in Italian fagiolo. It is possible, however, that the name has been transferred from a species of pea or vetch, or from a haricot formerly cultivated, to our modern haricot. It is rather bold to determine a species of Phaseolus from one or two epithets in an ancient author, when we see how difficult is the distinction of species to modern botanists with the plants under their eyes. Nevertheless, the dolichos of Theophrastus has been definitely referred to the scarlet runner, and the fasiolos to the dwarf haricot of our gardens, which are the two principal modern varieties of the common haricot, with an immense number of sub-varieties in the form of the pods and seed. I can only say it may be so.

If the common haricot was formerly known in Greece, it was not one of the earliest introductions, for the faseolos did not exist at Rome in Cato’s time, and it is only at the beginning of the empire that Latin authors speak of it. Virchow brought from the excavations at Troy the seeds of several leguminæ, which Wittmack1702 has ascertained to belong to the following species: broad bean (Faba vulgaris), garden-pea (Pisum sativum), ervilla (Ervum ervilia), and perhaps the flat-podded vetchling (Lathyrus Cicera), but no haricot. Nor has the species been found in the lake-dwellings of Switzerland, Savoy, Austria, and Italy.

There are no proofs or signs of its existence in ancient Egypt. No Hebrew name is known answering to the Phaseolus or Dolichos of botanists. A less ancient name, for it is Arabic, loubia, exists in Egypt for Dolichos lubia, and in Hindustani as loba for Phaseolus vulgaris.1703 As regards the latter species, Piddington only gives two names in modern languages, and those both Hindustani, loba and bakla. This, together with the absence of a Sanskrit name, points to a recent introduction into Southern Asia. Chinese authors do not mention P. vulgaris,1704 which is a further indication of a recent introduction into India, and also into Bactriana, whence the Chinese have imported plants from the second century of our era.

All these circumstances incline me to doubt whether the species was known in Asia before the Christian era. The argument based upon the modern Greek and Italian names for the haricot, derived from fasiolos, needs some support. It may be said in its favour that it was used in the Middle Ages, probably for the common haricot. In the list of vegetables which Charlemagne commanded to be sown in his farms, we find fasiolum,1705 without explanation. Albertus Magnus describes under the name faseolus a leguminous plant which appears to be our dwarf haricot.1706 I notice, on the other hand, that writers in the fifteenth century, such as Pierre Crescenzio1707 and Macer Floridus,1708 mention no faseolus or similar name. On the other hand, after the discovery of America, from the sixteenth century all authors publish descriptions and drawings of Phaseolus vulgaris, with a number of varieties.

It is doubtful that its cultivation is ancient in tropical Africa. It is indicated there less often than that of other species of the Dolichos and Phaseolus genera.

It had not occurred to any one to seek the origin of the haricot in America till, quite recently, some remarkable discoveries of fruits and seeds were made in Peruvian tombs at Ancon, near Lima. Rochebrune1709 published a list of the species of different families from the collection made by Cossac and Savatier. Among the number are three kinds of haricot, none of which, says the author, is Phaseolus vulgaris; but Wittmack,1710 who studied the leguminæ brought from these same tombs by Reiss and Stubel, says he made out several varieties of the common haricot among other seeds belonging to Phaseolus lunatus, Linnæus. He had identified them with the varieties of P. vulgaris called by botanists Oblongus purpureus (Martens), Ellipticus præcox (Alefeld), and Ellipticus atrofuscus (Alefeld), which belong to the category of dwarf or branchless haricots.

It is not certain that the tombs in question are all anterior to the advent of the Spaniards. The work of Reiss and Stubel, now in the press, will perhaps give some information on this head; but Wittmack admits, on their authority, that some of the tombs are not ancient. I notice a fact, however, which has passed without observation. The fifty species of Rochebrune are all American. There is not one which can be suspected to be of European origin. Evidently these plants and seeds were either deposited before the conquest, or, in certain tombs which perhaps belong to a subsequent epoch, the inhabitants took care not to put species of foreign origin. This was natural enough according to their ideas, for the custom of depositing plants in the tombs was not a result of the Catholic religion, but was an inheritance from the customs and opinions of the natives. The presence of the common haricot among exclusively American plants seems to me important, whatever the date of the tombs.

It may be objected that the seeds are insufficient ground for determining the species of a phaseolus, and that several species of this genus which are not yet well known were cultivated in South America before the arrival of the Spaniards. Molina1711 speaks of thirteen or fourteen species (or varieties?) cultivated formerly in Chili alone.

Wittmack insists upon the general and ancient use of the haricot in several parts of South America. This proves at least that several species were indigenous and cultivated. He quotes the testimony of Joseph Acosta, one of the first writers after the conquest, who says that “the Peruvians cultivated vegetables which they called frisoles and palares, and which they used as the Spaniards use garbanzos (chick-pea), beans and lentils. I have not found,” he adds, “that these or other European vegetables were found here before the coming of the Europeans.” Frisole, fajol, fasoler, are Spanish names for the common haricot, corruptions of the Latin faselus, fasolus, faseolus. Paller is American.

I may take this opportunity of explaining the origin of the French name haricot. I sought for it formerly in vain;1712 but I noticed that Tournefort1713 (Instit., p. 415) was the first to use it. I called attention also to the existence of the word arachos (Greek: arachos) in Theophrastus, probably for a kind of vetch, and of the Sanskrit word harenso for the common pea. I rejected as improbable the notion that the name of a vegetable could come from the dish called haricot or laricot of mutton, as suggested by an English author, and criticized Bescherelle, who derived the word from Keltic, while the Breton words are totally different, and signify small bean (fa-munno) or kind of pea (pis-ram). Lettré, in his dictionary, also seeks the etymology of the word. Without any acquaintance with my article, he inclines to the theory that haricot, the plant, comes from the ragout, seeing that the latter is older in the language, and that a certain resemblance may be traced between the haricot bean and the morsels of meat in the ragout, or else that this bean was suitable to the making of the dish. It is certain that this vegetable was called in French faséole or fazéole, from the Latin name, until nearly the end of the seventeenth century; but chance has led me to discover the real origin of the word haricot. An Italian name, araco, found in Durante and Matthioli, in Latin Aracus niger,1714 was given to a leguminous plant which modern botanists attribute to Lathyrus ochrus. It is not surprising that an Italian seventeenth-century name should be transported by French cultivators of the following century to another leguminous plant, and that ara should have been ari. It is the sort of mistake which is common now. Besides, aracos or arachos has been attributed by commentators to several Leguminosœ of the genera Lathyrus, Vicia, etc. Durante gives the Greek arachos as the synonym for his araco, whereby we see the etymology. Père Feuillée1715 wrote in French aricot; before him Tournefort spelt it haricot, in the belief, perhaps, that the Greek word was written with an aspirate, which is not the case; at least in the best authors.

I may sum up as follows:—(1) Phaseolus vulgaris has not been long cultivated in India, the south-west of Asia, and Egypt; (2) it is not certain that it was known in Europe before the discovery of America; (3) at this epoch the number of varieties suddenly increased in European gardens, and all authors commenced to mention them; (4) the majority of the species of the genus exist in South America; (5) seeds apparently belonging to the species have been discovered in Peruvian tombs of an uncertain date, intermixed with many species, all American.

I do not examine whether Phaseolus vulgaris existed in both hemispheres previous to cultivation, because examples of this nature are exceedingly rare among non-aquatic phanerogamous plants of tropical countries. Perhaps there is not one in a thousand, and even then human agency may be suspected.1716 To open this question in the case of Ph. vulgaris, it should at least be found wild in both old and new worlds, which has not happened. If it had occupied so vast an area, we should see signs of it in individuals really wild in widely separate regions on the same continent, as is the case with the following species, Ph. lunatas.

Scimetar-podded Kidney Bean, or Sugar Bean.—Phaseolus lunatus, Linnæus; Phaseolus lunatus macrocarpus; Bentham, Ph. inamœnus, Linnæus.

This haricot, as well as that called Lima, is so widely diffused in tropical countries, that it has been described under different names.1717 All these forms can be classed in two groups, of which Linnæus made different species. The commonest in our gardens is that which has been called since the beginning of the century the Lima haricot. It may be distinguished by its height, by the size of its pods and beans. It lasts several years in countries which are favourable to it.

Linnæus believed that his Ph. lunatus came from Bengal and the other from Africa, but he gives no proof. For a century his assertions were repeated. Now, Bentham,1718 who is careful about origins, believes the species and its variety to be certainly American; he only doubts about its presence as a wild plant both in Africa and Asia. I see no indication whatever of ancient existence in Asia. The plant has never been found wild, and it has no name in the modern languages of India or in Sanskrit.1719 It is not mentioned in Chinese works. Anglo-Indians call it French bean,1720 like the common haricot, which shows how modern is its cultivation.

It is cultivated in nearly all tropical Africa. However, Schweinfurth and Ascherson1721 do not mention it for Abyssinia, Nubia, or Egypt. Oliver1722 quotes a number of specimens found in Guinea and the interior of Africa, without saying whether they were wild or cultivated. If we suppose the species of African origin or of very early introduction, it would have spread to Egypt and thence to India.

The facts are quite different for South America. Bentham mentions wild specimens from the Amazon basin and Central Brazil. They belong especially to the large variety (macrocarpus), which abounds also in the Peruvian tombs of Ancon, according to Wittmack.1723 It is evidently a Brazilian species, diffused by cultivation, and perhaps long since naturalized here and there in tropical America. I am inclined to believe it was introduced into Guinea by the slave trade, and that it spread thence into the interior and the coast of Mozambique.

Moth, or Aconite-leaved Kidney Bean—Phaseolus aconitifolius, Willdenow.

An annual species grown in India as fodder, and of which the seeds are eatable, though but little valued. The Hindustani name is mout, among the Sikhs moth. It is somewhat like Ph. trilobus, which is cultivated for the seed. Ph. aconitifolius is wild in British India from Ceylon to the Himalayas.1724 The absence of a Sanskrit name, and of different names in modern Indian languages, points to a recent cultivation.

Three-lobed Kidney Bean—Phaseolus trilobus, Willdenow.

One of the most commonly cultivated species in India;1725 at least in the last few years, for Roxburgh,1726 at the end of the eighteenth century, had only seen it wild. All authors agree in considering it as wild from the foot of the Himalayas to Ceylon. It also exists in Nubia, Abyssinia, and Zambesi;1727 it is not said whether wild or cultivated. Piddington gives a Sanskrit name, and several names in modern Indian languages, which shows that the species has been cultivated, or at least known for three thousand years.

Green Gram, or Múng—Phaseolus mungo, Linnæus.

A species commonly cultivated in India and in the Nile Valley. The considerable number of varieties, and the existence of three different names in the modern languages of India, point to a cultivation of one or two thousand years, but there is no Sanskrit name.1728 In Africa it is probably recent. Anglo-Indian botanists agree that it is wild in India.

Lablab, or Wall—Dolichos Lablab, Linnæus.

This species is much cultivated in India and tropical Africa. Roxburgh counts as many as seven varieties with Indian names. Piddington quotes in his Index a Sanskrit name, schimbi, which recurs in modern languages. Its culture dates perhaps from three thousand years. Yet the species was not anciently diffused in China, or in Western Asia and Egypt; at least, I can find no trace of it. The little extension of these edible Leguminosæ beyond India in ancient times is a singular fact. It is possible that their cultivation is not of ancient date.

The lablab is undoubtedly wild in India, and also, it is said, in Java.1729 It has become naturalized from cultivation in the Seychelles.1730 The indications of authors are not positive enough to say whether it is wild in Africa.1731

Lubia—Dolichos Lubia, Forskal.

This species, cultivated in Europe under the name of lubia, loubya, loubyé, according to Forskal and Delile,1732 is little known to botanists. According to the latter author it exists also in Syria, Persia, and India; but I do not find this in any way confirmed in modern works on these two countries. Schweinfurth and Ascherson1733 admit it as a distinct species, cultivated in the Nile Valley. Hitherto no one has found it wild. No Dolichos or Phaseolus is known in the monuments of ancient Egypt. We shall see from the evidence of the common names that these plants were probably introduced into Egyptian agriculture after the time of the Pharaohs.

The name lubia is used by the Berbers, unchanged, and by the Spaniards as alubia for the common haricot, Phaseolus vulgaris. Although Phaseolus and Dolichos are very similar, this is an example of the little value of common names as a proof of species. Loba is, as we have seen, one of the Hindustani names for Phaseolus vulgaris,1734 and lobia that of Dolichos sinensis in the same language.1735 Orientalists should tell us whether lubia is an old word in Semitic languages. I do not find a similar name in Hebrew, and it is possible that the Armenians or the Arabs took lubia from the Greek lobos (λοβος), which means any projection, like the lobe of the ear, a fruit of the nature of a pod, and more particularly, according to Galen, Ph. vulgaris. Lobion (Λοβιον) in Dioscorides is the fruit of Ph. vulgaris, at least in the opinion of commentators.1736 It remains as loubion in modern Greek, with the same meaning.1737

Bambarra Ground Nut—Glycine subterranea, Linnæus, junr.; Voandzeia subterranea, Petit Thouars.

The earliest travellers in Madagascar remarked this leguminous annual, cultivated by the natives for the pod or seed, dressed like peas, French beans, etc. It resembles the earth, particularly in that the flower-stem curves downwards, and plunges the young fruit or pod into the earth. Its cultivation is common in the gardens of tropical Africa, and it is found, but less frequently, in those of Southern Asia.1738 It seems that it is not much grown in America,1739 except in Brazil, where it is called mandubi di Angola.1740

Early writers on Asia do not mention it; its origin must, therefore, be sought in Africa. Loureiro1741 had seen it on the eastern coast of this continent, and Petit Thouars in Madagascar, but they do not say that it was wild. The authors of the flora of Senegambia1742 described it as “cultivated and probably wild” in Galam. Lastly, Schweinfurth and Ascherson1743 found it wild on the banks of the Nile from Khartoum to Gondokoro. In spite of the possibility of naturalization from cultivation, it is extremely probable that the plant is wild in tropical Africa.

Buckwheat—Polygonum fagopyrum, Linnæus; Fagopyrum esculentum, Mœnch.

The history of this species has been completely cleared up in the last few years. It grows wild in Mantschuria, on the banks of the river Amur,1744 in Dahuria, and near Lake Baikal.1745 It is also indicated in China and in the mountains of the north of India,1746 but I do not find that in these regions its wild character is certain. Roxburgh has only seen it in a cultivated state in the north of India, and Bretschneider1747 thinks it doubtful that it is indigenous in China. Its cultivation is not ancient, for the first Chinese author who mentions it lived in the tenth or eleventh century of the Christian era.

Buckwheat is cultivated in the Himalayas under the names ogal or ogla and kouton.1748 As there is no Sanskrit name for this species nor for the two following, I doubt the antiquity of their cultivation in the mountains of Central Asia. It was certainly unknown to the Greeks and Romans. The name fagopyrum is an invention of modern botanists from the similarity in the shape of the seed to a beech-nut, whence also the German buchweitzen1749 (corrupted in English into buckwheat) and the Italian faggina.

The names of this plant in European languages of Aryan origin have not a common root. Thus the western Aryans did not know the species any more than the Sanskrit-speaking Orientals, a further sign of the nonexistence of the plant in the mountains of Central Asia. Even at the present day it is probably unknown in the north of Persia and in Turkey, since floras do not mention it.1750 Bosc states, in the Dictionnaire d’Agriculture, that Olivier had seen it wild in Persia, but I do not find this in this naturalist’s published account of his travels.

The species came into Europe in the Middle Ages, through Tartary and Russia. The first mention of its cultivation in Germany occurs in a Mecklenburg register of 1436.1751 In the sixteenth century it spread towards the centre of Europe, and in poor soil, as in Brittany, it became important. Reynier, who, as a rule, is very accurate, imagined that the French name sarrasin was Keltic;1752 but M. le Gall wrote to me formerly that the Breton names simply mean black wheat or black corn, ed-du and gwinis-du. There is no original name in Keltic languages, which seems natural now that we know the origin of the species.1753

When the plant was introduced into Belgium and into France, and even when it became known in Italy, that is to say in the sixteenth century, the name blé sarrasin (Saracen wheat) or sarrasin was commonly adopted. Common names are often so absurd, and so unthinkingly bestowed, that we cannot tell in this particular case whether the name refers to the colour of the grain which was that attributed to the Saracens, or to the supposed introduction from the country of the Arabs or Moors. It was not then known that the species did not exist in the countries south of the Mediterranean, nor even in Syria and Persia. It is also possible that the idea of a southern origin was taken from the name sarrasin, which was given from the colour. This origin was admitted until the end of the last and even in the present century.1754 Reynier was, fifty years ago, the first to oppose it.

Buckwheat sometimes escapes from cultivation and becomes quasi-wild. The nearer we approach its original country the more often this occurs, whence it results that it is hard to define the limit of the wild plant on the confines of Europe and Asia, in the Himalayas, and in China. In Japan these semi-naturalizations are not rare.1755

Tartary Buckwheat—Polygonum tataricum, Linnæus; Fagopyrum tataricum, Gærtner.

Less sensitive to cold than the common buckwheat, but yielding a poorer kind of seed, this species is sometimes cultivated in Europe and Asia—in the Himalayas,1756 for instance; but its culture is recent. Authors of the sixteenth and seventeenth centuries do not mention it, and Linnæus was one of the first to speak of it as of Tartar origin. Roxburgh and Hamilton had not seen it in Northern India in the beginning of this century, and I find no indication of it in China and Japan.

It is undoubtedly wild in Tartary and Siberia, as far as Dauria;1757 but Russian botanists have not found it further east, in the basin of the river Amur.1758

As this plant came from Tartary into Eastern Europe later than the common buckwheat, it is the latter which bears in several Slav languages the names tatrika, tatarka, or tattar, which would better suit the Tartary buckwheat.

It seems that the Aryan peoples must have known the species, and yet no name is mentioned in the ancient Indo-European languages. No trace of it has hitherto been found in the lake-dwellings of Switzerland or of Savoy.

Notch-seeded Buckwheat—Polygonum emarginatum, Roth; Fagopyrum emarginatum, Meissner.

This third species of buckwheat is grown in the highlands of the north-east of India, under the name phaphra or phaphar,1759 and in China.1760 I find no positive proof that it has been found wild. Roth only says that it “inhabits China,” and that the grain is used for food. Don,1761 who was the first of Anglo-Indian botanists to mention it, says that it is hardly considered wild. It is not mentioned in floras of the Amur valley, nor of Japan. Judging from the countries where it is cultivated, it is probably wild in the Eastern Himalayas and the north-west of China.

The genus Fagopyrum has eight species, all of temperate Asia.

Quinoa—Chenopodium quinoa, Willdenow.

The quinoa was a staple food of the natives of New Granada, Peru, and Chili, in the high and temperate parts at the time of the conquest. Its cultivation has persisted in these countries from custom, and on account of the abundance of the product.

From all time the distinction has existed between the quinoa with coloured leaves, and the quinoa with green leaves and white seed.1762 The latter was regarded by Moquin1763 as a variety of a little known species, believed to be Asiatic; but I believe that I showed conclusively that the two American quinoas are two varieties, probably very ancient, of a single species.1764 The less coloured, which is also the most farinaceous, is probably derived from the other.

The white quinoa yields a grain which is much esteemed at Lima, according to information furnished by the Botanical Magazine, where a good drawing may be seen (pl. 3641). The leaves may be dressed in the same manner as spinach.1765

No botanist has mentioned the quinoa as wild or semi-wild. The most recent and complete work on one of the countries where the species is cultivated, the Flora of Chili, by Cl. Gay, speaks of it only as a cultivated plant. Père Feuillée and Humboldt said the same for Peru and New Granada. It is perhaps due to the insignificance of the plant and its aspect of a garden weed that collectors have neglected to bring back wild specimens.

Kiery—Amarantus frumentaceus, Roxburgh.

This annual is cultivated in the Indian peninsula for its small farinaceous grain, which is in some localities the principal food of the natives.1766 Fields of this species, of a red or golden colour, produce a beautiful effect.1767 From Roxburgh’s account, Dr. Buchanan “discovered it on the hills of Mysore and Coimbatore,” which seems to indicate a wild condition. Amarantus speciosus, cultivated in gardens and figured on pl. 2227 of the Botanical Magazine, appears to be the same species. Hamilton found it in Nepal.1768 A variety or allied species, Amarantus anardana, Wallich,1769 is grown on the slopes of the Himalayas, but has been hitherto ill defined by botanists. Other species are used as vegetables (see p. 100, Amarantus gangeticus).

Chestnut—Castanea vulgaris, Lamarck.

The chestnut, belonging to the order Cupuliferæ, has an extended but disjunctive natural area. It forms forests and woods in mountainous parts of the temperate zone from the Caspian Sea to Portugal. It has also been found in the mountains of Edough in Algeria, and more recently towards the frontier of Tunis (Letourneux). If we take into account the varieties japonica and americana, it exists also in Japan and in the temperate region of North America.1770 It has been sown or planted in several parts of the south and west of Europe, and it is now difficult to know if it is wild or cultivated. However, cultivation consists chiefly in the operation of grafting good varieties on the trees which yield indifferent fruit. For this purpose the variety which produces but one large kernel is preferred to those which bear two or three, separated by a membrane, which is the natural state of the species.

The Romans in Pliny’s time1771 already distinguished eight varieties, but we cannot discover from the text of this author whether they possessed the variety with a single kernel (Fr. marron). The best chestnuts came from Sardis in Asia Minor, and from the neighbourhood of Naples. Olivier de Serres,1772 in the sixteenth century, praises the chestnuts Sardonne and Tuscane, which produced the single-kernelled fruit called the Lyons marron.1773 He considered that these varieties came from Italy, and Targioni1774 tells us that the name marrone or marone was employed in that country in the Middle Ages (1170).

Wheat and Kindred Species.—The innumerable varieties of wheat, properly so called, of which the ripened grain detaches itself naturally from the husk, have been classed into four groups by Vilmorin,1775 which form distinct species, or modifications of the common wheat according to different authors. I am obliged to distinguish them in order to study their history, but this, as will be seen, supports the opinion of a single species.1776

1. Common Wheat—Triticum vulgare, Villars; Triticum hybernum and T. æstivum, Linnæus.

According to the experiments of the Abbé Rozier, and later of Tessier, the distinction between autumn and spring wheats has no importance. “All wheats,” says the latter,1777 “are either spring or autumn sown, according to the country. They all pass with time from the one state to the other, as I have ascertained. They only need to be gradually accustomed to the change, by sowing the autumn wheat a little later, spring wheat a little earlier, year by year.” The fact is that among the immense number of varieties there are some which feel the cold of the winter more than others, and it has become the custom to sow them in the spring.1778 We need take no note of this distinction in studying the question of origin, especially as the greater number of the varieties thus obtained date from a remote period.

The cultivation of wheat is prehistoric in the old world. Very ancient Egyptian monuments, older than the invasion of the shepherds, and the Hebrew Scriptures show this cultivation already established, and when the Egyptians or Greeks speak of its origin, they attribute it to mythical personages, Isis, Ceres, Triptolemus.1779 The earliest lake-dwellings of Western Switzerland cultivated a small-grained wheat, which Heer1780 has carefully described and figured under the name Triticum vulgare antiquorum. From various facts, taken collectively, we gather that the first lake-dwellers of Robenhausen were at least contemporary with the Trojan war, and perhaps earlier. The cultivation of their wheat persisted in Switzerland until the Roman conquest, as we see from specimens found at Buchs. Regazzoni also found it in the rubbish-heaps of the lake-dwellers of Varese, and Sordelli in those of Lagozza in Lombardy.1781 Unger found the same form in a brick of the pyramid of Dashur, Egypt, to which he assigns a date, 3359 B.C. (Unger, Bot. Streifzüge, vii.; Ein Ziegel, etc., p. 9), Another variety (Triticum vulgare compactum muticum, Heer) was less common in Switzerland in the earliest stone age, but it has been more often found among the less ancient lake-dwellers of Western Switzerland and of Italy.1782 A third intermediate variety has been discovered at Aggtelek in Hungary, cultivated in the stone age.1783 None of these is identical with the wheat now cultivated, as more profitable varieties have taken their place.

The Chinese, who grew wheat 2700 B.C., considered it a gift direct from heaven.1784 In the annual ceremony of sowing five kinds of seed, instituted by the Emperor Shen-nung or Chin-nong, wheat is one species, the others being rice, sorghum, Setaria italica, and soy.

The existence of different names for wheat in the most ancient languages confirms the belief in a great antiquity of cultivation. The Chinese name is mai, the Sanskrit sumana and gôdhûma, the Hebrew chittah, Egyptian br, Guancho yrichen, without mentioning several names in languages derived from the primitive Sanskrit, nor a Basque name, ogaia or okhaya, which dates perhaps from the Iberians,1785 and several Finn, Tartar, and Turkish names, etc.,1786 which are probably Turanian. This great diversity might be explained by a wide natural area in the case of a very common wild plant, but this is far from being the case of wheat. On the contrary, it is difficult to prove its existence in a wild state in a few places in Western Asia, as we shall see. If it had been widely diffused before cultivation, descendants would have remained here and there in remote countries. The manifold names of ancient languages must, therefore, be attributed to the extreme antiquity of its culture in the temperate parts of Europe, Asia, and Africa—an antiquity greater than that of the most ancient languages. We have two methods of discovering the home of the species previous to cultivation in the immense zone stretching from China to the Canaries: first, the opinion of ancient authors; second, the existence, more or less proved, of wheat in a wild state in a given country.

According to the earliest of all historians, Berosus, a Chaldean priest, fragments of whose writings have been preserved by Herodotus, wild wheat (Frumentum agreste1787) might be seen growing in Mesopotamia. The texts of the Bible alluding to the abundance of wheat in Canaan prove no more than that the plant was cultivated there, and that it was very productive. Strabo,1788 born 50 B.C., says that, according to Aristobulus, a grain very similar to wheat grew wild upon the banks of the Indus on the 25th parallel of latitude. He also says1789 that in Hircania the modern Mazanderan) the grains of wheat which fell from the ear sowed themselves. This may be observed to some degree at the present day in all countries, and the author says nothing upon the important question whether this accidental sowing reproduced itself in the same place from generation to generation. According to the Odyssey,1790 wheat grew in Sicily without the help of man. But it is impossible to attach great importance to the words of a poet, and of a poet whose very existence is contested. Diodorus Siculus at the beginning of the Christian era says the same thing, and deserves greater confidence, since he is a Sicilian. Yet he may easily have been mistaken as to the wild character, as wheat was then generally cultivated in Sicily. Another passage in Diodorus1791 mentions the tradition that Osiris found wheat and barley growing promiscuously with other plants at Nisa, and Dureau de la Malle has proved that this town was in Palestine. Among all this evidence, that of Berosus and that of Strabo for Mesopotamia and Western India alone appear to me of any value.

The five species of seed of the ceremony instituted by Chin-nong are considered by Chinese scholars to be natives of their country,1792 and Bretschneider adds that communication between China and Western Asia dates only from the embassy of Chang-kien in the second century before Christ. A more positive assertion is needed, however, before we can believe wheat to be indigenous in China; for a plant cultivated in western Asia two or three thousand years before the epoch of Chin-nong, and of which the seeds are so easily transported, may have been introduced into the north of China by isolated and unknown travellers, as the stones of peaches and apricots were probably carried from China into Persia in prehistoric time.

Botanists have ascertained that wheat is not wild in Sicily at the present day.1793 It sometimes escapes from cultivation, but it does not persist indefinitely.1794 The plant which the inhabitants call wild wheat, Frumentu sarvaggiu, which covers uncultivated ground, is Ægilops ovata, according to Inzenga.1795

A zealous collector, Balansa, believed that he had found wheat growing on Mount Sipylus, in Asia Minor, under circumstances in which it was impossible not to believe it wild;1796 but the plant he brought back is a spelt, Triticum monococcum, according to a very careful botanist, to whom it was submitted for examination.1797 Olivier,1798 before him, when he was on the right bank of the Euphrates, to the north-west of Anah, a country unfit for cultivation, “found in a kind of ravine, wheat, barley, and spelt, which,” he adds, “we have already seen several times in Mesopotamia.”

Linnæus says,1799 that Heintzelmann found wheat in the country of the Baschkirs, but no one has confirmed this statement, and no modern botanist has seen the species really wild in the neighbourhood of the Caucasus or the north of Persia. Bunge,1800 whose attention was drawn to this point, declares that he has seen no indication which leads him to believe that cereals are indigenous in that country. It does not even appear that wheat has a tendency in these regions to spring up accidentally outside cultivated ground. I have not discovered any mention of it as a wild plant in the north of India, in China, or Mongolia.

It is remarkable that wheat has been twice asserted to be indigenous in Mesopotamia, at an interval of twenty three centuries, once by Berosus, and once by Olivier in our own day. The Euphrates valley lying nearly in the middle of the belt of cultivation which formerly extended from China to the Canaries, it is infinitely probable that it was the principal habitation of the species in very early prehistoric times. The area may have extended towards Syria, as the climate is very similar, but to the east and west of Western Asia wheat has probably never existed but as a cultivated plant; anterior, it is true, to all known civilization.

2. Turgid, and Egyptian Wheat—Triticum turgidum and T. compositum, Linnæus.

Among the numerous common names of the varieties which come under this head, we find that of Egyptian wheat. It appears that it is now much cultivated in that country and in the whole of the Nile valley. A. P. de Candolle says1801 that he recognized this wheat amongst seeds taken from the sarcophagi of ancient mummies, but he had not seen the ears. Unger1802 thinks it was cultivated by the ancient Egyptians, yet he gives no proof founded on drawings or specimens. The fact that no Hebrew or Armenian name1803 can be attributed to the species seems to me important. It proves at least that the remarkable forms with branching ears, commonly called wheat of miracle, wheat of abundance, did not exist in antiquity, for they would not have escaped the knowledge of the Israelites. No Sanskrit name is known, nor even any modern Indian names, and I cannot discover any Persian name. The Arab names which Delile1804 attributes to the species belong perhaps to other varieties of wheat. There is no Berber name.1805 From all this it results, I think, that the plants united under the name of Triticum turgidum, and especially the varieties with branching ears, are not ancient in the north of Africa or in the west of Asia.

Oswald Heer,1806 in his curious paper upon the plants of the lake-dwellers of the stone age in Switzerland, attributes to T. turgidum two non-branched ears, the one bearded, the other almost without beard, of which he gives drawings. Later, in an exploration of the lake-dwellings of Robenhausen, Messicommer did not find it, although there was abundant store of grain.1807 Strœbel and Pigorini said they found wheat with grano grosso duro (T. turgidum), in the lake-dwellings of Parmesan.1808 For the rest, Heer1809 considers this to be a variety or race of the common wheat, and Sordelli inclines to the same opinion.

Fraas thinks that the krithanias of Theophrastus was T. turgidum, but this is absolutely uncertain. According to Heldreich,1810 the great wheat is of modern introduction into Greece. Pliny1811 spoke briefly of a wheat with branching ears, yielding one hundred grains, which was most likely our miraculous wheat.

Thus history and philology alike lead us to consider the varieties of Triticum turgidum as modifications of the common wheat obtained by cultivation. The form with branching ears is not perhaps earlier than Pliny’s time.

These deductions would be overthrown by the discovery of the T. turgidum in a wild state, which has not hitherto been made with certainty. In spite of C. Koch,1812 no one admits that it grows, outside cultivation, at Constantinople and in Asia Minor. Boissier’s herbarium, so rich in Eastern plants, has no specimen of it. It is given as wild in Egypt by Schweinfurth, and Ascherson, but this is the result of a misprint.1813