In this letter I propose to... [make] such observations as have occurred to me during the last twenty-five years, in which I have been more or less conversant with the subject...
      [The Chinese] botanical arrangements (if indeed they deserve the name) are extremely defective.  No attempt has been made by them to form genera and species; the place of growth, the use, and the like, being with them the only distinguishing marks of plants.  It therefore cannot be supposed, that anything like a scientific botanical collection exists in China.  With the exception of a scanty Herbal, which was compiled by order of a former emperor, and some meagre articles on plants, in their Encyclopaedias, they have nothing which could have even secured a tolerable uniformity in the names, by which plants are known; so that Europeans (whose intercourse with China has been for the last half century, almost entirely confined to the port of Canton), frequently find that plants are not known by the same names, at the distance sometimes of a very few miles...
      The state of botany in China, may also be pretty correctly understood, by examining the Fa-tee, or flower gardens, situated on the banks of the river, at a short distance from Canton.  To these small nursery gardens strangers used to have access at all times; but for the two or three last years, these visits have been restricted to three days in a month, (say the eighth, eighteenth and twenty-eighth,) and they must besides pay eight dollars for their chop, or permission to go thither.  In these gardens may be seen all the plants, for which a demand exists among the Chinese themselves, but they will be found to consist of a very small variety, comprehending only shewy or odoriferous plants, shrubs, and trees, and such fruit trees as are commonly cultivated in the gardens.  To these may be added, abundance of dwarf trees, which the Chinese greatly admire, and for some of which they are content to pay a very high price.
      The soil of these gardens, and indeed of the banks of the river, to a considerable distance, consists of a strong alluvial clay.  The plants are either kept in the ground, or they are placed in pots, which are invariably filled with the same clay as the soil.  The clay, at lesst on the surface, both of the ground, and of the pots, is broken into small cubical pieces of about half an inch in size.  This soil has, no doubt, many advantages in the climate of Canton, where violent rains, and droughts of long continuance, are very common: for the small lumps of clay do not readily coalesce and run together, in consequence of the rains, or necessary waterings; but it is very obvious that plants so prepared, cannot be generally well suited for the purpose of transportation.  The rich Chinese commonly contract with the proprietors of these nurseries, at a certain price by the year, for a succession of plants, when in flower or fruit, which, after being used, are returned, to make room for others, in a fresher state, or of greater beauty.
      It is from these collections which I have described, that Europeans are generally supplied with the plants, which they send, or carry home: and it surely cannot be surprising that plants so treated commonly die on the passage.
      ...in the way of conveying plants safely from China to England ... the following particulars require attention:  the plants should be collected in proper time, so as to enable them to be firmly rooted in the soil in which they are to be transported to England; a proper soil should be obtained, wherein they might be planted; they should be arranged in their chests or boxes, accordingly as they require abundant, frequent, moderate, or slight waterings; when on board, the covers of the chests should be well closed when the spray is flying over the ship, and opened at all times in temperate and fine weather; the plants should be duly watered with good water, and particular attention be paid to them, from the time the ship arrives at her anchorage in the Thames, till they are landed.
      As but few plants can be had at the nursery gardens at Canton, and even these by no means in a proper state for transportation, those who wish to send any home, with a fair chance of success, will do well to procure them at least six months before the time of the ship's sailing, in order that they may be in a proper condition for the voyage.
      Being without data for a correct calculation, I must content myself with the nearest approximation I am able to make, from my own knowledge and observation: from these I am of opinion, that one thousand plants have been lost, for one, which survived the voyage to England.  Plants purchased at Canton, including their chests and other necessary charges, cost six shillings and eight pence sterling each, on a fair average; consequently every plant now in England, must have been introduced at the enormous expense of upwards of £300.  It surely, then, becomes a matter of importance to attempt some more certain method of gratifying the English horticulturist and botanist, with the plants of China. 1
 

      However much a correct taste may depreciate the art of dwarfing Trees and Shrubs, no doubt can be well entertained that the subject possesses some attractions to physiologists, since it may, in several respects, extend our information regarding the laws of organic life...
      [Air layering] is the general method now practised in China for obtaining by far the greatest number of fruit trees and shrubs.  It is extended also to many of the forest trees which they cultivate; and it is a preliminary step in the formation of nearly all [sic] their dwarf trees and shrubs.  Of the origin of the practice no record seems to have been preserved.  It was probably very remote, since we see, on the oldest specimens of porcelain, the same figures of dwarf trees that the Chinese admire at the present day...
      ... the same process [of propagation by cuttings] is so modified, that, instead of a full-formed beautiful tree, the bough is tortured [sic] into a grotesque dwarf.
      Dr. Morrison informs me, that the Chinese call dwarf trees Koo-Shoo, ancient trees; and that they express the rearing of them, by terms signifying bending down, or repressing ancient trees, which means much the same as dwarfing.
      When the dwarfing process is intended, the branch [using earth layering] which has pushed radicles into the surrounding composition in sufficient abundance, and for a sufficient length of time, is separated from the tree, and planted in a shallow earthenware flower pot, of an oblong square shape; it is sometimes made to rest upon a flat stone.  The pot is then filled with small pieces of alluvial clay, which, in the neighbourhood of Canton, is broken into bits, of about the size of common beans, being just sufficient to supply the scanty nourishment which the particular nature of the tree and the process require.  In addition to a careful regulation of the quantity and quality of the earth, the quantity of water, and the management of the plants, with respect to sun and shade, recourse is had to a great variety of mechanical contrivances, to produce the desired shape.  The containing flower pot is so narrow, that the roots pushing out towards the sides are pretty effectually cramped.  No radicle can descend, consequently it is only those which run towards the ends, or upwards, that can serve to convey nourishment properly, and it is easy to regulate those by cutting, burning, &c. so as to cramp the growth at pleasure.  Every succeeding formation of leaves becomes more and more stunted, the buds and radicles become diminished in the same proportion, till at length the balance between the root and leaves is obtained which suits the character of the dwarf required.  In some trees, this is accomplished in two or three years, but in others it requires at least twenty years.
      ... The Chinese say that all trees and shrubs may be made from boughs, and that these, under proper management, may be made dwarfs, but the larger the branches the better they are for the dwarfing process.
      The manipulations may be thus enumerated in detail.
      1st. The bark must be removed quite round the branch, to the breadth of about half its diameter.
      2d. The denuded part is to be covered with a composition similar to that used in England for grafting.  For common use the Chinese consider clay tempered with almost any kind of earth sufficient.
      3d. For large branches the Elm, Mitchellia Champacca, &c. a covering of straw, coarse cloth, &c. is used; for the Orange, Peach, &c. the composition itself is sufficient.  No contrivance for the application of water is ever seen in this part of China.  On this point I have made diligent enquiry among the best gardeners, and have always been assured that the sap of the boughs is sufficient to keep up a proper supply of moisture.
      4th. The time of separating the branch varies in different trees from about six weeks to three months.  Success in a great measure depends on the radicles being sufficiently abundant, and having acquired the necessary degree of tenacity to bear planting: in this the eye of the gardener is his only sure guide.
      5th. When it has been ascertained that the root is sufficient to preserve the living system, any exuberance of growth is at first carefully repressed by cutting off the extreme points of the branches, as well as rubbing off part of the buds and leaves.
      6th. The branches are next bent and contorted by wires, and various other mechanical means, to suit the taste of the operator.
      7th. After which, the various means already alluded to are practiced to complete the work; and which vary in different places.  It is the custom in the province of Fo-kien [Fujian], where the best dwarfs are said to be formed [sic], to entice ants to destroy the heart wood, into which openings are made to introduce sugar and the like... 
 (The Capsicum Cerasiforme, and some other species of the same genus, are planted in pots of not more than two or three inches in diameter.  The plants are reduced thereby to the most diminutive size, with fruit scarcely larger than our Currant.  The Dracaena ferrea, and some species of Bambusa, are made to grow in corners of their rock work, chiefly by means of water, to the height of only a few inches.  The Gardenia florida, by the same mode of treatment, is equally stunted.)
      ...dwarf trees being much in demand in China, and bringing a price in proportion to the crop of fruit which they bear, especially Oranges, Finger fruits, Peaches, Carambolas, Grapes, c. the artists select invariably the branches which are most loaded with flowers, for abscission [removal], which operation does not take place till the fruit is nearly ripe.  In this state they are exposed for sale.  The fruit is sufficiently well tasted, but is never intended for use by the Chinese.  They are contented with the handsome appearance of the miniature fruit tree, on which the fruit is also usually more permanent than on trees of full size.  In succeeding seasons, the dwarf tree flowers well, and the flowers, for the most part, remain longer, than on the old or large trees; little fruit, however, comes to maturity, and that little is not good.  The Peaches are without juice, and the Plums remain hard.
      Of fruit trees the Chinese admire most the dwarf Plum tree.  Should it be hollow, with its boughs twisted and contorted into all kinds of fantastic forms, it is so much the more esteemed; for such a dwarf tree a very large price will be cheerfully given.  Of forest trees the Elm seems to be the most approved, and it is the most common; but being much more easily formed than the Plum, its value is never so considerable. 2
 

1     Livingstone, John   "Observations on the Difficulties which have existed in the Transportation of Plants from China to England, and suggestions for obviating them," Transactions of the Horticultural Society of London, Vol. III, 1819, pp. 421-429;

Also partially cited in Bretschneider, Emil, M.D.  History of European Botanical Discoveries in China (Leipzig: Zentral-Antiquariat; 1981, reprint of the original 1898 edition), pg. 266-267;

cf. Bretschneider Botanicon sinicum (as Article III in "Journal of the North-China Branch of the Royal Asiatic Society," 1881, New Series Vol. XVI, Part I, Shanghai, 1882, specifically pp. 18-230, "Notes on Chinese Botany from Native and Western Sources"), pg. 106: "Although the common cultivated plants are known under the same Chinese names all over the Empire, many other plants, especially drugs, go under different local names in different properties.  Li Shi chen, the author of the [Ben Cao Gang Mu (1596), the most important native Chinese work ever compiled on Materia medica and natural History], and other authors before him, have attempted to bring these synonyms together, but perhaps they have not always been correct in their identifications.  In some instances also the same Chinese names are applied to distinct plants in different parts of China."

2     Livingstone, John "Account of the Method of Dwarfing Trees and Shrubs, as practised by the Chinese, including their Plan of Propagation from Branches,"Transactions of the Horticultural Society of London, Vol. IV, 1822, pp. 224-231;

Cited in Bretschneider  History, pg. 267, which states the article is "dated Macao, 1820."  In History, pg. 225 and Bretschneider Botanicon sinicum, pg. 121 it is stated that Rev. Rob Morrison was one of the first English sinologists, a resident of Macao since 1807.  He accompanied the 1816-17 Embassy to Peking as principal interpreter.  In 1822 Morrison published an English-Chinese dictionary. 

The use of ants is also mentioned in Hansen, Albert A. "Dwarfed plants, they are normal in all but size," Nature Magazine, August 1930, pg. 130.  Note: termites are sometimes also known as "white ants."