Short URL for this page:

[image ALT: Much of my site will be useless to you if you've got the images turned off!]
Bill Thayer

[image ALT: Cliccare qui per una pagina di aiuto in Italiano.]

[Link to a series of help pages]
[Link to the next level up]
[Link to my homepage]

This webpage reproduces an article in the
of the American Philological Association

Vol. 64 (1933), pp66‑76.

The text is in the public domain.

This page has been carefully proofread
and I believe it to be free of errors.
If you find a mistake though,
please let me know!

 p66  Notes on Ancient Grafting
Arthur Stanley Pease
Harvard University

The more serious ancient instances of grafting between plants of different families were presumably the products, not of scions, but of "grafting by approach."

In modern practice the process of grafting is commonly limited​1 to the insertion of grafts of a given species upon stocks of other forms or strains of the same species (as in the cultivation of rubber, apples, pears, peaches, and roses) or upon stocks of another species of the same genus (for example, pear and apple, both species of the genus Pyrus), so that this familiar operation is confined within limits somewhat similar to those of the quite distinct process of hybridization. The principle of limitation is recognized by ancient writers, for Theophrastus remarks:​2 "It is also easy to grow like kinds on like; the 'eye' [i.e. 'graft'] is practically something growing on its kind. The reciprocal relation, too, is easily understood — especially in plants whose bark [i.e. cambium] is similar, for there is the least need of adaptation between congeners — and the process amounts practically to a mere transposition. For it happens that at the same time the juices  p67 and the entire trees are eager for germination, so that whenever there is similarity of kind and of feeling for fruit on the part of both, it is reasonable that there should be speedy growth, the more so in each part, when their kinds and habitats and seasons of growth are the more similar." Or, again:​3 "It is reasonable to suppose that graftage will produce more beautiful fruit, especially if the cultivated is grafted on the wild; for there results a richer supply of nourishment because of the strength of the supporting plant. Wild olive we are accordingly counselled to sow and then to bud or graft on it later; for a better start is obtained from a stronger stock, and more nourishment is drawn up to make the tree bear fine fruit."​4 Varro moreover asserts that we must consider the trees from which and upon which the grafts are made, as well as the season and fashion of grafting.​5 "For the oak does not admit [a graft of] the pear, even though the apple does admit the pear." Virgil,​6 too, in a poetical case of the rhetorical figure known as the ἀδύνατον,​7 makes Damon pray that the wolf may flee from the sheep, the hard oaks bear golden apples, the alder bloom with the narcissus, and the bark of the tamarisk ooze with rich amber. The common belief was in 1763 formulated by Michel Adanson in his Familles naturelles des plantes, who limited grafts to species within a single genus, a principle later extended to members of the same family.8

 p68  Yet beside this tradition of intrageneric grafts another theory also appears. The pseudo-Aristotelian de Plantis, a work perhaps by Nicolaus of Damascus and probably dating from the Augustan period,​9 declares — I quote Forster's (1913) translation of the Latin translation made in the thirteenth century by the Englishman Alfredus from an Arabic translation (now lost) of the Greek original–: "Grafting of one on another is better in the case of trees which are similar and have the same proportions; the best results are obtained in the grafting, for instance, of apple on pear, fig on fig, or vine on vine. Sometimes grafting of different species is resorted to, bay,​10 for example, on wild plane,​11 olive-trees on terebinth,​12 mulberries on a number of different trees, and wild trees on garden trees." Almost contemporaneously Virgil in two places in the second Georgic alludes to the same belief: "And often we see the boughs of one turn lightly into another's, and the changed pear-tree bear her grafted apples, and the plums redden on the stony cornel";​13 "By grafting the rough arbutus yields the walnut, and the barren planes carry sturdy apple-boughs; the mountain-ash silvers with white pear-blossom, the beech with chestnut blooms, and swine crush acorns beneath the elm."​14 Columella in his fourth book​15 has much to say of grafting, budding, and the like, and describes in detail what is nowadays known as "grafting by approach" or "inarching," a method of grafting, e.g. the olive on the fig,​16 in  p69 which the two trees are planted alongside, a branch of the olive is bound close to the stock of the fig, and, after some years' interval, becomes grafted to the main trunk of the fig. When these have fully grown together the connection with the olive is severed, and in this way, Columella says, any kind of graft may be inserted upon any kind of tree, though the ancients (antiqui) have wrongly maintained that the possibility of grafting was limited to trees with similar bark, fruit, and fruiting season. Pliny, who in the Natural History gives much of interest on the subject of grafting, asserts​17 that he had himself seen a tree loaded with every kind of fruit, on one bough with nuts, on another with olives, on other sides with the vine and with pears, figs, and pomegranates, but he adds that the tree was short-lived. He goes on to say​18 that the plane is thought the most tolerant of all kinds of grafts and next after it the oak. Plutarch, in his Quaestiones Convivales,​19 devotes a whole dialogue to inquiring why the fir, pine, and cypress cannot be grafted,​20 and speaks of being entertained in gardens near the Cephisus in which his host showed the guests trees diversified with many sorts of grafts. "From mastich-trees​21 we saw olive-trees shooting forth and pomegranates from myrtles; and there were also oaks producing  p70 excellent pears and planes grafted with apples, fig-trees with mulberries, etc." The company express their surprise at these monstrosities, suggestive of sphinxes and chimaeras, and the host explains that such limits as there are to the process are due to the different nature (φύσις) of the species in question and their sap, the diverse character of the cambium layer of their bark, and their inconsistent habits in other respects. A letter of Julian​22 ascribes to Theophrastus commendation of the fig-tree as able to receive different kinds of grafts, "and as the only one of them all that easily bears a growth of any other sort, if you cut out every one of its boughs and then break off and insert a different engrafted stock into each of the cleft stumps; hence to look at it is often equivalent to a complete garden, since it returns to you the variegated and manifold splendours of other fruits, as happens in the loveliest orchard." The passage of Theophrastus, however, to which he seems to refer,​23 merely remarks that the fig-tree will, better than any other tree, strike its roots and grow by any method of propagation. Noteworthy in the fourth century is the fourteenth book of Palladius, which is devoted to grafting, and in the introductory poem of which the author announces

quae quibus hospitium praestent virgulta docebo,

quae sit adoptivis arbor onusta comis.​24

In what follows we hear of a large variety of grafts, including pear on apple (line 59), ash (60), almond (61), mountain-ash (64), quince (65), chestnut (67), and medlar (69); apple on pear (78), plum (81), sorbus (83), willow (85), plane (87 f.), peach (89), poplar (90), medlar (91), and chestnut (93 f.); citrus on mulberry (109 f.) and pear (111 f.); plum on chestnut (116); fig on mulberry (119) and plane (123); mulberry on fig (127), ash (129), beech (131), chestnut (132), and terebinth (135 f.); cherry on laurel (143), plane (145), plum (145), and poplar (147); almond on plum (150), peach (151), and chestnut  p71 (155 f.) pistachio on almond (157) and terebinth (159); chestnut on willow (161 f.); and walnut on arbutus (163 f.; cf. II.15, 19).​a Doubtless some of this material may be due to tradition, to inference from supposedly analogous cases, or even to poetic invention rather than to autopsy, yet the less imaginative prose parts of the work of Palladius contain similar statements: citrus on pear, "ut quidam," and mulberry (IV.10, 16); medlar on pear and apple (IV.10, 21); fig on mulberry and plane (IV.10, 32); pistachio on terebinth and almond (according to some; XI.12, 3); cherry on plum, plane, or poplar ("ut alii"; XI.12.6), and many others too numerous to recite.​25 Finally the Geoponica,​26 in a chapter on grafting, mention many examples of intergeneric or interfamily grafts. The chapter is by its title ascribed to Diophanes, apparently the native of Nicaea, who sent to King Deiotarus an abridgement of Cassius Dionysius's translation of the Carthaginian Mago's work on agriculture.​27 If the ascription be correct, the allusions in the chapter to this type of grafting might be carried back to the time of Diophanes himself, i.e. to the Ciceronian period. Doubt, however, arises from the fact that sections 9 and 10 of the chapter quote Didymus​28 and Florentinus,​29 authors of four to five centuries later, so that we can hardly feel confidence that any such principle was known earlier than the time of Virgil, and should probably consider either the de Plantis or Virgil's Georgics as containing the earliest certain reference to intergeneric grafting.

In a modern horticultural treatise​30 I find this statement: "Those who are fond of oddities can, with the assistance of grafting, have on the same thorn stock at the same time fruiting branches of the pear, the medlar, the beam-tree [Pyrus  p72 Aria], the service-tree, the mountain-ash, the European and Japanese quince, and also see there the flowers of the double and red thorns, the Cotoneaster, and the Pyracantha. They may gather from the same plum stock plums, apricots, peaches, nectarines, almonds, the corymbs of the Canadian cherry, and flower garlands of the Chinese and Japanese plum. But these whimsicalities are unworthy the attention of cultivators." At first sight these words suggest the descriptions I have quoted from Pliny and Plutarch, but the botanist will at once note that, although some of these grafts involve the union of plants of two genera, none oversteps the barriers between two families. Further, disbelief in the possibility of grafts between two families has appeared from time to time ever since the antiqui to whom Columella refers.​31 Modern editors​32 have been somewhat agnostic in this regard and some modern scientists decidedly dogmatic.​33 Donald Grant Mitchell (Ik Marvel), to be sure, asks:​34 "Is it remotely possible that these old gentlemen understood the physiology of plants better than we?" Baltet​35 cites the credulity of de Caylus, Mme. de Genlis, and the Abbé Rozier, and remarks: "The ancients are not the only persons guilty of falsification in the matter of grafting. There have been many instances of it in our time, and we shall long continue to hear of black roses being produced from a black currant stock, etc." Typical of the opinion of many modern horticulturalists is the statement of Liberty H. Bailey:​36 "It  p73 seems to have been a popular misconception that any kind of plant will grown on any other. Pliny​37 asserts that the art of grafting was taught to man by nature. Birds swallow seeds, and these seeds, falling in 'some cleft in the bark of a tree,' germinate and make plants. 'Hence it is that we see the cherry growing upon the willow, the plane upon the laurel, the laurel upon the cherry, and fruits of various tints and hues all springing from the same tree at once.' This, of course," continues Professor Bailey, "is not grafting at all, but the implanting of seeds in earth-filled chinks and cracks, in which the plants find a convenient foothold and soil."38

The phenomenon under discussion was for me strikingly illustrated half a dozen years ago in Japan, where, in the Kasuga-no‑miya shrine at Nara, in an enclosure by itself, there stands a tree called the Yadorigi (or Yadoriki). This has a short but very thick trunk of the species called isu (a sort of Ilex?) from which grow boughs of six other species, namely maple, cherry, wisteria, camellia, heavenly bamboo (nanten), and niwatoko (a kind of grape).​39 The tree is evidently very old and the object of much veneration, and it seems to form a fairly close parallel to those described by Pliny and Plutarch. Later, at Nikko, I observed a Cryptomeria tree almost six feet in diameter, from the trunk of which, about twelve feet up, grew a thick bough of oak, which my Japanese companion thought was not grafted but sprung from  p74 a seed lodged on the bark, difficult as this appeared on the vertical and unbranched trunk of the Cryptomeria. An inquiry from Professor M. Miyoshi of the Botanical Institute of the Imperial University at Tokyo brought a courteous reply, stating that the Yadorigi​40 "is a natural product, originated by the growth of one or more trees of different species, sprouted from the seeds carried by wind or brought by birds in hollow cavities of the trunk of an old large tree. They take nourishment and water from decayed matter and soil accumulated in the cavities, sending often roots through the hollow space down to the earth." He adds that he has seen similar trees, bearing several species on the trunk, in different parts of Japan, as well as a Cryptomeria with a cherry growing from the upper part of its trunk. "Other apparent graftings," he concludes, "are in fact instances of merely space parasites but not of true graftings."

That this explanation may account for many cases I readily concede, yet the numerous allusions in ancient authors — both poets and more practical prose writers — may well cause us to pause, and I should, accordingly, like to adduce some further modern evidence. Gardner, Bradford, and Hooker remark, as stated above,​41 that horticultural varieties of exogenous plants generally may be intergrafted freely, species somewhat less so, genera only occasionally,​42 and families very rarely. More specific is Sahut,​43 who, like the pseudo-Aristotelian de  p75 Plantis already cited, emphasizes the importance of likeness in vegetative characters, which do not always parallel taxonomic attributes, and states (p203) that Carrière in 1858 succeeded in grafting Garrya elliptica (of the Garryaceae) upon Aucuba japonica (of the Cornaceae), a combination thereafter commonly employed in France and Belgium.​44 This interfamily grafting, then, is rare but occasional, and the process was carried yet further by a French scientist, M. Daniel, who states​45 that he has effected successful grafts of beans (Leguminosae) on Xanthium (Compositae) and Ricinus (Euphorbiaceae); sunflowers (Compositae) on melons (Cucurbitaceae); Coleus (Labiatae), Cineraria and Zinnia (Compositae), and various Cruciferae on the tomato (Solanaceae); Jerusalem artichoke (Compositae) on black nightshade (Solanaceae); Coleus (Labiatae) on Achyranthes (Amarantaceae); Aster (Compositae) on Phlox (Polemoniaceae); and, among woody plants, of maple (Aceraceae) on lilac (Oleaceae). All these, he asserts, were successful, though maple on lilac only when the plants were very young.46

A clue to the whole problem is, I think, to be found in the statement of Daniel​47 that two plants cannot ordinarily be grafted by scions unless the two are of the same family, but  p76 that this restriction does not apply in the case of "grafting by approach" or "inarching," in which plants of widely separated families can at times be united. If we recall that Columella knew and described this method and observe that it is also known to the Japanese,​48 we may not unfairly conclude that of the ancient cases already noted some (mainly within a genus or a family) belong to ordinary grafting by scions, some doubtless illustrate space parasitism, others are the products of grafting by approach, while those still remaining may well be due to mistaken analogies and enthusiastic exaggerations of amateurs, whether poets or prose writers. In each of these types we have to deal, not with some ancient "lost art," but with phenomena and horticultural practices known to antiquity, recognized today, and doubtless continuous through the intervening centuries.49

The Author's Notes:

1 Gardner, Bradford, and Hooker, Fundamentals of Fruit Production (1922), 552: "Botanical relation­ship, as understood by closeness in the system of classification, is a fair guide to probable congeniality, but it is by no means infallible. Horticultural varieties of exogenous plants generally may be intergrafted freely, species somewhat less so, genera only occasionally, and families only rarely." Cf. note 38, infra; also Baltet, The Art of Grafting and Budding, 6th Engl. ed. (1910), 3:º "The laws of the affinities of species are almost unknown. The observations hitherto made have been undertaken in a practical rather than a purely scientific spirit, as in the fertilizing of plants. . . . No theory has as yet been deduced . . . except that kinds to be united by grafting must be of the same botanic family."

2 C. P. I.6.2, the translation being that of Dengler (1927).

3 C. P. I.6.10.

4 Cf. Palladius V.2; XIV.53 f. for grafting olive on oleaster; id. IV.10, 32, for the fig on the caprificus; Varro, R. R. I.40.5, for grafting pears upon wild pears (less advisable than upon cultivated stock). For olive on wild olive cf. Romans 11.17‑24; Clem. Strom. VI.15: αὐτίκα ἡ ἀγριέλαιος ἐγκεντρίζεται εἰς τὴν πιότητα τῆς ἐλαίας καὶ δὴ καὶ φύεται ὁμοειδῶς ταῖς ἡμέροις ἐλαίαις· χρῆται γὰρ τὸ ἐμφυτευόμενον ἀντὶ γῆς τῷ δένδρῳ τῷ ἐν ᾧ φυτεύεται. Cf. also Poll. I.12.241 (olive on wild olive); Steph. Byz. s.v. Γύαρος: ἡ ἄχερδος ἐκεῖ θανάσιμόν ἐστι κἂν εἰς ἄλλο δένδρον ἐμπήξῃς ἀφαυαίνει; Hor. Epod. 2.19; 16.44.

5 R. R. I.40.5.

6 Ecl. 8.52‑54. Line 318 of an anonymous mediaeval poet published by Hagen in Fleckeisen's Jahrb. XCVII (1868), 725, mentions quercus pomiferas, vimina plena piris as an ἀδύνατον. Cf. also Matthew 7.16.

7 Cf. Canter in A. J. P. LI (1930), 32‑41.

8 Daniel, Les variations spécifiques dans la greffe (1902), 8.

9 Cf. Christ-Schmid-Stählin, Gesch. d. gr. Litt. I6 (1912), 736; II.16 (1920), 375.

10 Artemsia,º probably the Arabic al‑damasit (= laurus). This and the next two notes I borrow from the footnotes of Forster's translation.

11 Adul, probably the Arabic ad‑dulb (= platanus).

12 Botam, probably the Arabic botham (= terebinthus).

13 Geor. II.32‑34 (Mackail's translation).

14 Geor. II.69‑72; in 73‑82 the technique of grafting is described. Pliny's remembrance (N. H. XV.57) of the Virgilian passage varies somewhat from the original: Vergilius insitam nucibus arbutum, malis platanum, cerasis ulmum dicat, and suggests that in other passages to be mentioned some liberties may have been taken with the species named.

15 Especially chap. 11.

16 Plin. N. H. XVII.137 remarks: est etiamnum nova inserendi ratio, ne quid sciens quidem praeteream quod usquam invenerim, Columellae excogitata, ut adfirmat ipse, qua vel diversae insociabilesque arborum naturae copulentur, ut fici atque oleae. After further details he concludes: quarto anno abscisum totum adoptantis esse, nondum volgata ratione aut mihi certe satis comperta. This is the process now known as inarching; cf. Bailey in Encycl. Brit. 11 ed., XIII (1910), 756; Baltet, op. cit. 39º and passim.

17 XVII.120. The tree is described as iuxta Tiburtes Thulias, but for Thulias there are variant readings.

18 XVII.121: capacissima insitorum omnium ducitur platanus, postea robur; verum utraque sapores conrumpit. quaedam omni genere inseruntur, ut ficus, ut puniceae; vitis non recipit emplastra, nec quibus tenuis aut caducus rimosusque cortex, neque inoculationem siccae aut umoris exigui.

19 II.6.

20 Grafts of deciduous plants on evergreens are usually unsuccessful, according to Baltet, op. cit. 4,º yet various grafts of Coniferae upon Coniferae are noted by the same writer (op. cit. 175, 185, 216),º and grafts of the evergreen Eriobotrya on the deciduous Cydonia vulgaris and Crataegus oxyacantha are described by Sahut in Revue horticole LVII (1885), 202.

21 σχῖνοι, lentisci.

22 Ep. 80, p391D (III.270 Wright, whose translation I quote).

23 H. P. II.5.6.

24 Lines 19‑20.

25 Cf. II.15.19, III.17.8; 25.6; 25.17; 25.30; 25.32; IV.10.32; V.4.5; XII.7.6; 7.8; 7.15; 7.22; etc.

26 X.76.

27 Varro, R. R. I.1.10; Colum. I.1.10; Wellmann in P.‑W. V (1905), 1049.

28 Probably Didymus of Alexandria, who lived in the fourth or fifth century after Christ and wrote γεωργικά; cf. Wellmann in P.‑W. V (1905), 445.

29 Of the third century after Christ; Wellmann in P.‑W. VI (1909), 2756.

30 Baltet, op. cit. 4.º

31 V.11 (see note 15, supra). Martial VI.49.10 f., may refer to such sceptics: nascetur, licet hoc velis negare, inserta tibi ficus a cupressu.

32 E.g. the Heyne-Wagner 4th edition of Virgil, I (1830), 404 (on Georg. II.69‑72), repeated by Forbiger. The examples, however, which are there cited (pear on mountain-ash — both within the genus Pyrus —, and chestnut on the oak — a combination within one family, recognized as possible today; cf. Sahut in Revue horticole LVII (1885), 203, who says that, though usually short-lived, one at Dijon has lived from 1835 to 1885; Anon. in Journ. of the Royal Hort. Soc. XXIV (1900), 38; Baltet, op. cit. 3)º prove nothing for interfamily grafts.

33 E.g. Miller and Du Hamel (cited by Heyne-Wagner); Baltet, op. cit. 5.º

34 Wet Days at Edgewood (1884), 51 — a quotation which I owe to Rand, The Magical Art of Virgil (1931), 230.

35 Op. cit. 5.º

36 Cyclopedia of Am. Horticulture III (1906), 659.

37 N. H. XVII.99.

38 This disbelief in the existence of interfamily grafts I have found shared by various horticulturists with whom I have discussed the matter, notably several members of the staff of the Arnold Arboretum and of that of the Massachusetts State College. Several, however, of the works later cited in this article have been kindly called to my notice by Professor Orton Clark of the latter institution.

39 This was the list given on the spot by an attendant to my interpreter, Miss Caroline Schereschewsky of Nara. The Official Guide to Eastern Asia (published by the Imperial Railways), II (1914), 300, names maple, cherry, wisteria, camellia, nandina, kusatazu (Sambucus javanica Bl.), and an oak. Cf. also the Imperial Railway circular of Kyoto and Nara, 27. I was unable to get close enough to the tree to diagnose the species myself, and since it was winter some of the branches had lost their leaves, while others were still evergreen.

40 "Parasite," but in fact only a "space parasite," utilizing the space, according to Professor Miyoshi. Here may doubtless be classed many such cases as the tree which is a willow below and an ash above described from Westcapelle in West Flanders by Burvenich in Tijdschrift over Boomteelkunde (1880), 373 (and in Bull. Arb. Belge [1880], 373).

41 Note 1, supra.

42 Many cases may, however, be cited, to which the following articles will direct one: Guillon in Revue horticole LIII (1881), 257; Carrière in Revue horticole LIV (1882), 228; Dawson in Trans. Mass. Hort. Soc. (1895), 123; Anon. in Journ. Royal Hort. Soc. XXIV (1900), 38 (a long list); Lindemuth in Berichte d. deut. bot. Ges. XIX (1901), 515‑529 (including cases of herbaceous with woody species, and annuals with perennials; others have noted evergreens with deciduous types); Gardner, Bradford, and Hooker, loc. cit.

43 In Revue horticole LVII (1885), 201‑204, in an article entitled "Limites de la possibilité de greffage."

44 He cites also Crassula (of the Crassulaceae) and Stapelia (of the Asclepiadaceae) upon Opuntia (of the Cactaceae). Dr. I. M. Johnston of the Arnold Arboretum tells me of his observation at Berlin of a graft of Ephedra (of the Ephedraceae) upon Welwitschia Barnesii (of the Gnetaceae); these, however, are very nearly related families.

45 Comptes Rendus de l'Acad. des Sciences CXXXI (1900), 192 f.: "Sur les limites de possibilité de greffage chez les végétaux." The name of the author — who has published many articles on the scientific aspects of grafting — and the auspices under which this work appears would seem to give its conclusions greater authority than that of the works censured by Baltet (note 35, supra).

46 Some of the species named are annuals, so that the persistence of the graft is not shown. But even Pliny, in the passage cited above (XVII.120), admits that certain grafts were short-lived.

47 P193; cf. his Les variations spécifiques dans la greffe (1902), 28, where he adds that he had in his garden "une greffe par rapprochement, soudée par les parenchymes depuis le printemps 1900, entre le Lilas (Oléacées) et l'Érable (Acérinées)."

48 The New English Dict., s.v. Inarching, cites Alcock, The Capital of the Tycoon 1 (1863), 325: "The Japanese understand, and sometimes practice, the inarching of plants." I should be tempted to suspect this method in the case remarked above (note 39).

49 I have not attempted to collect mediaeval allusions to grafting, and have not chanced upon any helpful treatment of the question.

Thayer's Note:

a The Internet is a wonderful thing. Mohammad Suleman Bajwa alerts me to some most unusual grafts from medieval Spain: squash seed on squill onion, date tree on parsnip! S. K. Jayyusi, The Legacy of Muslim Spain (Brill, 1994), p997.

[image ALT: Valid HTML 4.01.]

Page updated: 17 Apr 20