phoses in almost every cultivated flower can be compared to nothing but the alterations caused in the plumage of birds, or the hairs of animals by domestication. No cause has ever been assigned to these phenomena, neither has any attempt been made to determine the cause in plants. We are, however, in possession of the knowledge of some of the laws under which change of colour is effected. A blue flower will change to white or red, but not to bright yellow; a bright yellow flower will become white or red, but never blue. Thus the hyacinth, of which the primitive colour is blue, produces abundance of white and red varieties, but nothing that can be compared to bright yellow; the yellow hyacinths, as they are called, being a sort of pale yellow ochre colour verging to green. Again, the ranunculus, which is originally of an intense yellow, sports into scarlet, red, purple, and almost any colour but blue. White flowers, which have a tendency to produce red, will never sport to blue, although they will to yellow; the rose, for example, and Chrysanthemums. It is also probable that white flowers, with a tendency to produce blue, will not vary to yellow; but of this I have no instance at hand.

Smell varies in degree rather than in nature; some plants, which are but slightly perfumed, as the common China rose, acquire a powerful fragrance when converted to the variety called "the sweet-scented;" but I am not acquainted with any case among flowers in which a positive difference of smell exists in two varieties of the same species.

Metamorphoses of the fruit are very common, and administer largely to the wants of mankind. They consist of alteration in colour, size, flavour, smell, and structure. The wild blue sloe of our hedges has, in the course of ages, by successive domestication, been converted into the purple, white, and yellow plums of our desserts. The wild crab is the original from which have sprung the many coloured, Proteus-like variety of the apple; some of which are destitute of smell, others scented like the pine apple, and a few partaking of the perfume of the rose. In peas the parchment-like lining of the pod occasionally disappears, and the whole substance of the seed vessel consists of lax cellular tissue. In the orange a second fruit

is sometimes produced in the inside, agreeing in all respects with the outer fruit, even in peel; this is doubtless due to an attempt at producing a second series of pistilla. In a variety of citrons called the fingered shaddock, well known in China, this tendency to form a second row of pistilla is not only in excess, but the cells of the fruit, in attempting to separate themselves into the simple individuals of which the fruit of the shaddock is ordinarily composed, divide it into distinct lobes irregularly arranged round a common axis.

Having thus passed in review the irregular metamorphoses of plants through all the different parts, there still remains a subject on which it is requisite to say a few words. This is the permanency of such metamorphoses, or their capability of being perpetuated by seeds. It is a general law of nature, that seeds will perpetuate a species but not a variety; and this is no doubt true, if rightly considered: and yet it may be urged, if this be so, how have the varieties, well known to gardeners and agriculturists, for many years been unceasingly carried on from generation to generation without change? The long, red, and round white radishes of the markets, for instance, have been known from time immemorial in the same state in which they now exist. The answer is this. A species will perpetuate itself from seed for ever under any circumstances, and left to the simple aid of nature: but accidental varieties cannot be so perpetuated; if suffered to become wild, they very soon revert to the form from which they originally sprung. It is necessary that they should be cultivated with the utmost care; that seed should be saved from those individuals only in which the marks of the variety are most distinctly traced; and all plants that indicate any disposition to cast off their peculiar characteristics should be rejected. If this is carefully done, the existence of any variety of annual or perennial plant may undoubtedly be prolonged through many generations; but in woody plants this scarcely happens, it being a rare occurrence to find any variety of tree or shrub producing its like when increased by seed.

EXPLANATION OF THE PLATES.

N. B. All the figures in the plates, of which the following is an explanation, are more or less magnified: the drawings from which they have been prepared are in all cases original, except where it is stated to the contrary.

PLATE I.

Fig. 1. A small portion of a section of the cellular tissue of the pith of Calycanthus floridus, showing the pore-like spots upon the membrane. Fig. 2. A section of the leaf of Lilium candidum, after A. Brongniart; a, cuticle of the upper surface; b, ditto of the lower surface; c, stomata cut through in different directions; these last are seen to open into cavities in the parenchyma; d, upper layer of parenchyma; e, intermediate ditto; f, lower ditto. Fig. 3. Cubical cellular tissue, passing gradually into prismatical, from the stem of the gourd, cut vertically; after Kieser.

Fig. 4. Fibres forming arches in the endothecium of Linaria cymbalaria; after Purkinje.

Fig. 5. Fusiform cellules in the wood of a young branch of Viscum album; after Kieser; a, common hexagonal cells of the pith, with grains of amydon sticking to their sides; b, fusiform cellules, considered by Kieser to be pierced with holes; c, other cells of the same figure, with lines of dots spirally arranged on the membrane; d, others, in which the dots are run into lines; e, f, others, in which the cellules have all the appearance of short spiral vessels. Kieser considers these not as spiral vessels but as cellules of a peculiar kind, replacing spiral vessels in the viscum.

Fig. 6. A portion of the cuticle of Billbergia amæna, with the membrane torn on one side, showing that it does not tear with an even edge, but breaks into little teeth.

Fig. 7. Muriform cellular tissue, forming the medullary processes of Platanus occidentalis. Each cellule contains particles of brownish matter of very irregular size and form.

Fig. 8. a, Glandular hairs of the peduncle of Primula sinensis; 1, the glandular apex more highly magnified, with a particle of the viscid secretion of the species on its point; 2, the apex of another hair, showing that the end is open, a conical piece of the viscid secretion lying in the orifice: b, a hair of Dorstenia, showing the cellular base from which it arises, and that it consists of a single hollow conical curved cell.

Fig. 9. A branched hair from the cilia of the leaf of a species of Verbascum.
Fig. A. A simple coloured hair in Dichorizandra rufa.

Fig. B. A hair with tumid articulations from the leaf of Gesneria tuberosa.