In foliage we have variegated leaves which are often inherited; dark purple or red leaves, as in the hazel, barberry, and beech, the colour in these two latter trees being sometimes strongly and sometimes weakly inherited (10/151. 'Journal of a Horticultural Tour, by Caledonian Hort. Soc.' 1823 page 107. Alph. De Candolle 'Geograph. Bot.' page 1083. Verlot 'Sur La Production des Varietes' 1865 page 55 for the Barberry.); deeply-cut leaves; and leaves covered with prickles, as in the variety of the holly well called ferox, which is said to reproduce itself by seed. (10/152. Loudon's 'Arboretum et Fruticetum' volume 2 page 508.) In fact, nearly all the peculiar varieties evince a tendency, more or less strongly marked, to reproduce themselves by seed. (10/153. Verlot 'Des Varietes' 1865 page 92.) This is to a certain extent the case, according to Bosc (10/154. Loudon's 'Arboretum et Fruticetum' volume 3 page 1376.), with three varieties of the elm, namely, the broad-leafed, lime-leafed, and twisted elm, in which latter the fibres of the wood are twisted. Even with the heterophyllous hornbeam (Carpinus betulus), which bears on each twig leaves of two shapes, "several plants raised from seed all retained "the same peculiarity." (10/155. 'Gardener's Chronicle' 1841 page 687.) I will add only one other remarkable case of variation in foliage, namely, the occurrence of two sub- varieties of the ash with simple instead of pinnated leaves, and which generally transmit their character by seed. (10/156. Godron 'De l'Espece' tome 2 page 89. In Loudon's 'Gardener's Mag.' volume 12 1836 page 371, a variegated bushy ash is described and figured, as having simple leaves; it originated in Ireland.) The occurrence, in trees belonging to widely different orders, of weeping and fastigiate varieties, and of trees bearing deeply cut, variegated, and purple leaves, shows that these deviations of structure must result from some very general physiological laws.

Differences in general appearance and foliage, not more strongly marked than those above indicated, have led good observers to rank as distinct species certain forms which are now known to be mere varieties. Thus, a plane-tree long cultivated in England was considered by almost every one as a North American species: but is now ascertained by old records, as I am informed by Dr. Hooker, to be a variety. So, again, the Thuja pendula or filiformis was ranked by such good observers as Lambert, Wallich, and others, as a true species; but it is now known that the original plants, five in number, suddenly appeared in a bed of seedlings, raised at Mr. Loddige's nursery, from T. orientalis; and Dr. Hooker has adduced excellent evidence that at Turin seeds of T. pendula have reproduced the parent form, T. orientalis. (10/157. 'Gardener's Chronicle' 1863 page 575.)

Every one must have noticed how certain individual trees regularly put forth and shed their leaves earlier or later than others of the same species. There is a famous horse-chestnut in the Tuileries which is named from leafing so much earlier than the others. There is also an oak near Edinburgh which retains its leaves to a very late period. These differences have been attributed by some authors to the nature of the soil in which the trees grow; but Archbishop Whately grafted an early thorn on a late one, and vice versa, and both grafts kept to their proper periods, which differed by about a fortnight, as if they still grew on their own stocks. (10/158. Quoted from Royal Irish Academy in 'Gardener's Chronicle' 1841 page 767.) There is a Cornish variety of the elm which is almost an evergreen, and is so tender that the shoots are often killed by the frost; and the varieties of the Turkish oak (Q. cerris) may be arranged as deciduous, sub-evergreen, and evergreen. (10/159. Loudon 'Arboretum et Fruticetum:' for Elm see volume 3 page 1376; for Oak page 1846.)

SCOTCH FIR (Pinus sylvestris).

I allude to this tree as it bears on the question of the greater variability of our hedgerow trees compared with those under strictly natural conditions. A well-informed writer (10/160. 'Gardener's Chronicle' 1849?, page 822.) states that the Scotch fir presents few varieties in its native Scotch forests; but that it "varies much in figure and foliage, and in the size, shape, and colour of its cones, when several generations have been produced away from its native locality." There is little doubt that the highland and lowland varieties differ in the value of their timber, and that they can be propagated truly by seed; thus justifying Loudon's remark, that "a variety is often of as much importance as a species, and sometimes far more so." (10/161. 'Arboretum et Fruticetum' volume 4 page 2150.) I may mention one rather important point in which this tree occasionally varies; in the classification of the Coniferae, sections are founded on whether two, three, or five leaves are included in the same sheath; the Scotch fir has properly only two leaves thus enclosed, but specimens have been observed with groups of three leaves in a sheath. (10/162. 'Gardener's Chronicle' 1852 page 693.) Besides these differences in the semi-cultivated Scotch fir, there are in several parts of Europe natural or geographical races, which have been ranked by some authors as distinct species. (10/163. See 'Beitrage zur Kenntniss Europaischer Pinus-arten von Dr. Christ: Flora, 1864.' He shows that in the Ober-Engadin P. sylvestris and montana are connected by intermediate links.) Loudon (10/164. 'Arboretum et Fruticetum' volume 4 pages 2159 and 2189.) considers P. pumilio, with its several sub- varieties, as mughus, nana, etc., which differ much when planted in different soils, and only come "tolerably true from seed," as alpine varieties of the Scotch fir; if this were proved to be the case, it would be an interesting fact as showing that dwarfing from long exposure to a severe climate is to a certain extent inherited.

THE HAWTHORN (Crataegus oxyacantha).

The Hawthorn has varied much. Besides endless slighter variations in the form of the leaves, and in the size, hardness, fleshiness, and shape of the berries, Loudon (10/165. Ibid volume 2 page 830; Loudon's 'Gardener's Mag.' volume 6 1830 page 714.) enumerates twenty-nine well-marked varieties. Besides those cultivated for their pretty flowers, there are others with golden-yellow, black, and whitish berries; others with woolly berries, and others with re-curved thorns. Loudon truly remarks that the chief reason why the hawthorn has yielded more varieties than most other trees, is that nurserymen select any remarkable variety out of the immense beds of seedlings which are annually raised for making hedges. The flowers of the hawthorn usually include from one to three pistils; but in two varieties, named monogyna and sibirica, there is only a single pistil; and d'Asso states that the common thorn in Spain is constantly in this state. (10/166. Loudon's 'Arboretum et Fruticetum' volume 2 page 834.) There is also a variety which is apetalous, or has its petals reduced to mere rudiments. The famous Glastonbury thorn flowers and leafs towards the end of December, at which time it bears berries produced from an earlier crop of flowers. (10/167. Loudon's 'Gardener's Mag.' volume 9 1833 page 123.) It is worth notice that several varieties of the hawthorn, as well as of the lime and juniper, are very distinct in their foliage and habit whilst young, but in the course of thirty or forty years become extremely like each other (10/168. Ibid volume 11 1835 page 503.) thus reminding us of the well-known fact that the deodar, the cedar of Lebanon, and that of the Atlas, are distinguished with the greatest ease whilst young, but with difficulty when old.

FLOWERS.

I shall not for several reasons treat the variability of plants which are cultivated for their flowers alone at any great length. Many of our favourite kinds in their present state are the descendants of two or more species crossed and commingled together, and this circumstance alone would render it difficult to detect the difference due to variation. For instance, our Roses, Petunias, Calceolarias, Fuchsias, Verbenas, Gladioli, Pelargoniums, etc., certainly have had a multiple origin. A botanist well acquainted with the parent-forms would probably detect some curious structural differences in their crossed and cultivated descendant; and he would certainly observe many new and remarkable constitutional peculiarities. I will give a few instances, all relating to the Pelargonium, and taken chiefly from Mr. Beck (10/169. 'Gardener's Chronicle' 1845 page 623.) a famous cultivator of this plant: some varieties require more water than others; some are "very impatient of the knife if too greedily used in making cuttings;" some, when potted, scarcely "show a root at the outside of the ball of the earth;" one variety requires a certain amount of confinement in the pot to make it throw up a flower- stem; some varieties bloom well at the commencement of the season, others at the close; one variety is known (10/170. D. Beaton in 'Cottage Gardener' 1860 page 377. See also Mr. Beck on the habits of Queen Mab in 'Gardener's Chronicle' 1845 page 226.), which will stand "even pine-apple top and bottom heat, without looking any more drawn than if it had stood in a common greenhouse; and Blanche Fleur seems as if made on purpose for growing in winter, like many bulbs, and to rest all summer." These odd constitutional peculiarities would enable a plant in a state of nature to become adapted to widely different circumstances and climates.

Flowers possess little interest under our present point of view, because they have been almost exclusively attended to and selected for their beautiful colour, size, perfect outline, and manner of growth. In these particulars hardly one long-cultivated flower can be named which has not varied greatly. What does a florist care for the shape and structure of the organs of fructification, unless, indeed, they add to the beauty of the flower? When this is the case, flowers become modified in important points; stamens and pistils may be converted into petals, and additional petals may be developed, as in all double flowers. The process of gradual selection by which flowers have been rendered more and more double, each step in the process of conversion being inherited, has been recorded in several instances. In the so-called double flowers of the Compositae, the corollas of the central florets are greatly modified, and the modifications are likewise inherited. In the columbine (Aquilegia vulgaris) some of the stamens are converted into petals having the shape of nectaries, one neatly fitting into the other; but in one variety they are converted into simple petals. (10/171. Moquin-Tandon 'Elements de Teratologie' 1841 page 213.) In the "hose in hose" primulae, the calyx becomes brightly coloured and enlarged so as to resemble a corolla; and Mr. W. Wooler informs me that this peculiarity is transmitted; for he crossed a common polyanthus with one having a coloured calyx (10/172. See also 'Cottage Gardener' 1860 page 133.) and some of the seedlings inherited the coloured calyx during at least six generations. In the "hen-and-chicken" daisy the main flower is surrounded by a brood of small flowers developed from buds in the axils of the scales of the involucre. A wonderful poppy has been described, in which the stamens are converted into pistils; and so strictly was this peculiarity inherited that, out of 154 seedlings, one alone reverted to the ordinary and common type. (10/173. Quoted by Alph. de Candolle 'Bibl. Univ.' November 1862 page 58.) Of the cock's-comb (Celosia cristata), which is an annual, there are several races in which the flower-stem is wonderfully "fasciated" or compressed; and one has been exhibited (10/174. Knight 'Transact. Hort. Soc.' volume 4 page 322.) actually eighteen inches in breadth. Peloric races of Gloxinia speciosa and Antirrhinum majus can be propagated by seed, and they differ in a wonderful manner from the typical form both in structure and appearance.

A much more remarkable modification has been recorded by Sir William and Dr. Hooker (10/175. 'Botanical Magazine' tab. 5160 figure 4; Dr. Hooker in 'Gardener's Chronicle' 1860 page 190; Prof. Harvey in 'Gardener's Chronicle' 1860 page 145; Mr. Crocker in 'Gardener's Chron.' 1861 page 1092.) in Begonia frigida. This plant properly produces male and female flowers on the same fascicles; and in the female flowers the perianth is superior; but a plant at Kew produced, besides the ordinary flowers, others which graduated towards a perfect hermaphrodite structure; and in these flowers the perianth was inferior. To show the importance of this modification under a classificatory point of view, I may quote what Prof. Harvey says, namely, that had it "occurred in a state of nature, and had a botanist collected a plant with such flowers, he would not only have placed it in a distinct genus from Begonia, but would probably have considered it as the type of a new natural order." This modification cannot in one sense be considered as a monstrosity, for analogous structures naturally occur in other orders, as with Saxifragae and Aristolochiaceae. The interest of the case is largely added to by Mr. C.W. Crocker's observation that seedlings from the NORMAL flowers produced plants which bore, in about the same proportion as the parent-plant, hermaphrodite flowers having inferior perianths. The hermaphrodite flowers fertilised with their own pollen were sterile.

If florists had attended to, selected, and propagated by seed other modifications of structure besides those which are beautiful, a host of curious varieties would certainly have been raised; and they would probably have transmitted their characters so truly that the cultivator would have felt aggrieved, as in the case of culinary vegetables, if his whole bed had not presented a uniform appearance. Florists have attended in some instances to the leaves of their plant, and have thus produced the most elegant and symmetrical patterns of white, red, and green, which, as in the case of the pelargonium, are sometimes strictly inherited. (10/176. Alph. de Candolle 'Geograph. Bot.' page 1083; 'Gardener's Chronicle' 1861 page 433. The inheritance of the white and golden zones in Pelargonium largely depends on the nature of the soil. See D. Beaton in 'Journal of Horticulture' 1861 page 64.) Any one who will habitually examine highly- cultivated flowers in gardens and greenhouses will observe numerous deviations in structure; but most of these must be ranked as mere monstrosities, and are only so far interesting as showing how plastic the organisation becomes under high cultivation. From this point of view such works as Professor Moquin-Tandon's 'Teratologie' are highly instructive.

ROSES.

These flowers offer an instance of a number of forms generally ranked as species, namely, R. centifolia, gallica, alba, damascena, spinosissima, bracteata, indica, semperflorens, moschata, etc., which have largely varied and been intercrossed. The genus Rosa is a notoriously difficult one, and, though some of the above forms are admitted by all botanists to be distinct species, others are doubtful; thus, with respect to the British forms, Babington makes seventeen, and Bentham only five species. The hybrids from some of the most distinct forms—for instance, from R. indica, fertilised by the pollen of R. centifolia—produce an abundance of seed; I state this on the authority of Mr. Rivers (10/177. 'Rose Amateur's Guide' T. Rivers 1837 page 21.) from whose work I have drawn most of the following statements. As almost all the aboriginal forms brought from different countries have been crossed and re-crossed, it is no wonder that Targioni- Tozzetti, in speaking of the common roses of the Italian gardens, remarks that "the native country and precise form of the wild type of most of them are involved in much uncertainty." (10/178. 'Journal Hort. Soc.' volume 9 1855 page 182.) Nevertheless, Mr. Rivers in referring to R. indica (page 68) says that the descendants of each group may generally be recognised by a close observer. The same author often speaks of roses as having been a little hybridised; but it is evident that in very many cases the differences due to variation and to hybridisation can now only be conjecturally distinguished.

The species have varied both by seed and by bud; such modified buds being often called by gardeners sports. In the following chapter I shall fully discuss this latter subject, and shall show that bud-variations can be propagated not only by grafting and budding, but often by seed. Whenever a new rose appears with any peculiar character, however produced, if it yields seed, Mr. Rivers (page 4) fully expects it to become the parent-type of a new family. The tendency to vary is so strong in some kinds, as in the Village Maid (Rivers page 16), that when grown in different soils it varies so much in colour that it has been thought to form several distinct kinds. Altogether the number of kinds is very great: thus M. Desportes, in his Catalogue for 1829, enumerates 2562 as cultivated in France; but no doubt a large proportion of these are merely nominal.

It would be useless to specify the many points of difference between the various kinds, but some constitutional peculiarities may be mentioned. Several French roses (Rivers page 12) will not succeed in England; and an excellent horticulturist (10/179. The Rev. W.F. Radclyffe in 'Journal of Horticulture' March 14, 1865 page 207.) remarks, that "Even in the same garden you will find that a rose that will do nothing under a south wall will do well under a north one. That is the case with Paul Joseph here. It grows strongly and blooms beautifully close to a north wall. For three years seven plants have done nothing under a south wall." Many roses can be forced, "many are totally unfit for forcing, among which is General Jacqueminot." (10/180. 'Gardener's Chronicle' 1831 page 46.) From the effects of crossing and variation Mr. Rivers enthusiastically anticipates (page 87) that the day will come when all our roses, even moss-roses, will have evergreen foliage, brilliant and fragrant flowers, and the habit of blooming from June till November. "A distant view this seems, but perseverance in gardening will yet achieve wonders," as assuredly it has already achieved wonders.

It may be worth while briefly to give the well-known history of one class of roses. In 1793 some wild Scotch roses (R. spinosissima) were transplanted into a garden (10/181. Mr. Sabine in 'Transact. Hort. Soc.' volume 4 page 285.); and one of these bore flowers slightly tinged with red, from which a plant was raised with semi-monstrous flowers, also tinged with red; seedlings from this flower were semi-double, and by continued selection, in about nine or ten years, eight sub-varieties were raised. In the course of less than twenty years these double Scotch roses had so much increased in number and kind, that twenty-six well-marked varieties, classed in eight sections, were described by Mr. Sabine. In 1841 (10/182. 'An Encyclop. of Plants' by J.C. Loudon 1841 page 443.) it is said that three hundred varieties could be procured in the nursery-gardens near Glasgow; and these are described as blush, crimson, purple, red, marbled, two-coloured, white, and yellow, and as differing much in the size and shape of the flower.

PANSY OR HEARTSEASE (Viola tricolor, etc.).

The history of this flower seems to be pretty well known; it was grown in Evelyn's garden in 1687; but the varieties were not attended to till 1810- 1812, when Lady Monke, together with Mr. Lee, the well-known nursery-man, energetically commenced their culture; and in the course of a few years twenty varieties could be purchased. (10/183. Loudon's 'Gardener's Magazine' volume 11 1835 page 427; also Journal of Horticulture' April 14, 1863 page 275.) At about the same period, namely in 1813 or 1814, Lord Gambier collected some wild plants, and his gardener, Mr. Thomson, cultivated them, together with some common garden varieties, and soon effected a great improvement. The first great change was the conversion of the dark lines in the centre of the flower into a dark eye or centre, which at that period had never been seen, but is now considered one of the chief requisites of a first-rate flower. In 1835 a book entirely devoted to this flower was published, and four hundred named varieties were on sale. From these circumstances this plant seemed to me worth studying, more especially from the great contrast between the small, dull, elongated, irregular flowers of the wild pansy, and the beautiful, flat, symmetrical, circular, velvet-like flowers, more than two inches in diameter, magnificently and variously coloured, which are exhibited at our shows. But when I came to enquire more closely, I found that, though the varieties were so modern, yet that much confusion and doubt prevailed about their parentage. Florists believe that the varieties (10/184. Loudon's 'Gardener's Magazine' volume 8 page 575: volume 9 page 689.) are descended from several wild stocks, namely, V. tricolor, lutea, grandiflora, amoena, and altaica, more or less intercrossed. And when I looked to botanical works to ascertain whether these forms ought to be ranked as species, I found equal doubt and confusion. Viola altaica seems to be a distinct form, but what part it has played in the origin of our varieties I know not; it is said to have been crossed with V. lutea. Viola amoena (10/185. Sir J.E. Smith 'English Flora' volume 1 page 306. H.C. Watson 'Cybele Britannica' volume 1 1847 page 181.) is now looked at by all botanists as a natural variety of V. grandiflora; and this and V. sudetica have been proved to be identical with V. lutea. The latter and V. tricolor (including its admitted variety V. arvensis) are ranked as distinct species by Babington, and likewise by M. Gay (10/186. Quoted from 'Annales des Sciences' in the Companion to the 'Bot. Mag.' volume 1 1835 page 159.) who has paid particular attention to the genus; but the specific distinction between V. lutea and tricolor is chiefly grounded on the one being strictly and the other not strictly perennial, as well as on some other slight and unimportant differences in the form of the stem and stipules. Bentham unites these two forms; and a high authority on such matters, Mr. H.C. Watson (10/187. 'Cybele Britannica' volume 1 page 173. See also Dr. Herbert on the changes of colour in transplanted specimens, and on the natural variations of V. grandiflora, in 'Transact. Hort. Soc.' volume 4 page 19.) says that, "while V. tricolor passes into V. arvensis on the one side, it approximates so much towards V. lutea and V. Curtisii on the other side, that a distinction becomes scarcely more easy between them."

Hence, after having carefully compared numerous varieties, I gave up the attempt as too difficult for any one except a professed botanist. Most of the varieties present such inconstant characters, that when grown in poor soil, or when flowering out of their proper season, they produced differently coloured and much smaller flowers. Cultivators speak of this or that kind as being remarkably constant or true; but by this they do not mean, as in other cases, that the kind transmits its character by seed, but that the individual plant does not change much under culture. The principle of inheritance, however, does hold good to a certain extent even with the fleeting varieties of the Heartsease, for to gain good sorts it is indispensable to sow the seed of good sorts. Nevertheless, in almost every large seed-bed a few, almost wild seedlings reappear through reversion. On comparing the choicest varieties with the nearest allied wild forms, besides the difference in the size, outline, and colour of the flowers, the leaves sometimes differ in shape, as does the calyx occasionally in the length and breadth of the sepals. The differences in the form of the nectary more especially deserve notice; because characters derived from this organ have been much used in the discrimination of most of the species of Viola. In a large number of flowers compared in 1842 I found that in the greater number the nectary was straight; in others the extremity was a little turned upwards, or downwards, or inwards, so as to be completely hooked; in others, instead of being hooked, it was first turned rectangularly downwards, and then backwards and upwards; in others, the extremity was considerably enlarged; and lastly, in some the basal part was depressed, becoming, as usual, laterally compressed towards the extremity. In a large number of flowers, on the other hand, examined by me in 1856 from a nursery-garden in a different part of England, the nectary hardly varied at all. Now M. Gay says that in certain districts, especially in Auvergne, the nectary of the wild V. grandiflora varies in the manner just described. Must we conclude from this that the cultivated varieties first mentioned were all descended from V. grandiflora, and that the second lot, though having the same general appearance, were descended from V. tricolor, of which the nectary, according to M. Gay, is subject to little variation? Or is it not more probable that both these wild forms would be found under other conditions to vary in the same manner and degree, thus showing that they ought not to be ranked as specifically distinct?

THE DAHLIA

The Dahlia has been referred to by almost every author who has written on the variation of plants, because it is believed that all the varieties are descended from a single species, and because all have arisen since 1802 in France, and since 1804 in England. (10/188. Salisbury in 'Transact. Hort. Soc.' volume 1 1812 pages 84, 92. A semi-double variety was produced in Madrid in 1790.) Mr. Sabine remarks that "it seems as if some period of cultivation had been required before the fixed qualities of the native plant gave way and began to sport into those changes which now so delight us." (10/189. 'Transact. Hort. Soc.' volume 3 1820 page 225.) The flowers have been greatly modified in shape from a flat to a globular form. Anemone and ranunculus-like races (10/190. Loudon's 'Gardener's Mag.' volume 6 1830 page 77.) which differ in the form and arrangement of the florets, have arisen; also dwarfed races, one of which is only eighteen inches in height. The seeds vary much in size. The petals are uniformly coloured or tipped or striped, and present an almost infinite diversity of tints. Seedlings of fourteen different colours (10/191. Loudon's 'Encyclop. of Gardening' page 1035.) have been raised from the same plant; yet, as Mr. Sabine has remarked, "many of the seedlings follow their parents in colour." The period of flowering has been considerably hastened, and this has probably been effected by continued selection. Salisbury, writing 1808, says that they then flowered from September to November; in 1828 some new dwarf varieties began flowering in June (10/192. 'Transact. Hort. Soc.' volume 1 page 91; and Loudon's 'Gardener's Mag.' volume 3 1828 page 179.); and Mr. Grieve informs me that the dwarf purple Zelinda in his garden is in full bloom by the middle of June and sometimes even earlier. Slight constitutional differences have been observed between certain varieties: thus, some kinds succeed much better in one part of England than in another (10/193. Mr. Wildman in 'Gardener's Chronicle' 1843 page 87. 'Cottage Gardener' April 8, 1856 page 33.) and it has been noticed that some varieties require much more moisture than others. (10/194. M. Faivre has given an interesting account of the successive variations of the Chinese primrose, since its introduction into Europe about the year 1820: 'Revue des Cours Scientifiques' June 1869 page 428.)

Such flowers as the CARNATION, COMMON TULIP, and HYACINTH, which are believed to be descended, each from a single wild form, present innumerable varieties, differing almost exclusively in the size, form, and colour of the flowers. These and some other anciently cultivated plants which have been long propagated by offsets, pipings, bulbs, etc., become so excessively variable, that almost each new plant raised from seed forms a new variety, "all of which to describe particularly," as old Gerarde wrote in 1597, "were to roll Sisyphus's stone, or to number the sands."

HYACINTH (Hyacinthus orientalis).

It may, however, be worth while to give a short account of this plant, which was introduced into England in 1596 from the Levant. (10/195. The best and fullest account of this plant which I have met with is by a famous horticulturist, Mr. Paul, of Waltham, in the 'Gardener's Chronicle' 1864 page 342.) The petals of the original flower, says Mr. Paul, were narrow, wrinkled, pointed, and of a flimsy texture; now they are broad, smooth, solid, and rounded. The erectness, breadth, and length of the whole spike, and the size of the flowers, have all increased. The colours have been intensified and diversified. Gerarde, in 1597, enumerates four, and Parkinson, in 1629, eight varieties. Now the varieties are very numerous, and they were still more numerous a century ago. Mr. Paul remarks that "it is interesting to compare the Hyacinths of 1629 with those of 1864, and to mark the improvement. Two hundred and thirty-five years have elapsed since then, and this simple flower serves well to illustrate the great fact that the original forms of nature do not remain fixed and stationary, at least when brought under cultivation. While looking at the extremes, we must not, however, forget that there are intermediate stages which are for the most part lost to us. Nature will sometimes indulge herself with a leap, but as a rule her march is slow and gradual." He adds that the cultivator should have "in his mind an ideal of beauty, for the realisation of which he works with head and hand." We thus see how clearly Mr. Paul, an eminently successful cultivator of this flower, appreciates the action of methodical selection.

In a curious and apparently trustworthy treatise, published at Amsterdam (10/196. 'Des Jacinthes, de leur Anatomie, Reproduction, et Culture' Amsterdam 1768.) in 1768, it is stated that nearly 2,000 sorts were then known; but in 1864 Mr. Paul found only 700 in the largest garden at Haarlem. In this treatise it is said that not an instance is known of any one variety reproducing itself truly by seed: the white kinds, however, now (10/197. Alph. de Candolle 'Geograph. Bot.' page 1082.) almost always yield white hyacinths, and the yellow kinds come nearly true. The hyacinth is remarkable from having given rise to varieties with bright blue, pink, and distinctly yellow flowers. These three primary colours do not occur in the varieties of any other species; nor do they often all occur even in the distinct species of the same genus. Although the several kinds of hyacinths differ but slightly from each other except in colour, yet each kind has its own individual character, which can be recognised by a highly educated eye; thus the writer of the Amsterdam treatise asserts (page 43) that some experienced florists, such as the famous G. Voorhelm, seldom failed in a collection of above twelve hundred sorts to recognise each variety by the bulb alone! This same writer mentions some few singular variations: for instance, the hyacinth commonly produces six leaves, but there is one kind (page 35) which scarcely ever has more than three leaves; another never more than five; whilst others regularly produce either seven or eight leaves. A variety, called la Coryphee, invariably produces (page 116) two flower-stems, united together and covered by one skin. The flower-stem in another kind (page 128) comes out of the ground in a coloured sheath, before the appearance of the leaves, and is consequently liable to suffer from frost. Another variety always pushes a second flower-stem after the first has begun to develop itself. Lastly, white hyacinths with red, purple, or violet centres (page 129) are the most liable to rot. Thus, the hyacinth, like so many previous plants, when long cultivated and closely watched, is found to offer many singular variations.]

In the two last chapters I have given in some detail the range of variation, and the history, as far as known, of a considerable number of plants, which have been cultivated for various purposes. But some of the most variable plants, such as Kidney-beans, Capsicum, Millets, Sorghum, etc., have been passed over; for botanists are not at all agreed which kinds ought to rank as species and which as varieties; and the wild parent- species are unknown. (10/198. Alph. De Candolle 'Geograph. Bot.' page 983.) Many plants long cultivated in tropical countries, such as the Banana, have produced numerous varieties; but as these have never been described with even moderate care, they are here also passed over. Nevertheless, a sufficient, and perhaps more than sufficient, number of cases have been given, so that the reader may be enabled to judge for himself on the nature and great amount of variation which cultivated plants have undergone.

CHAPTER 1.XI.

ON BUD-VARIATION, AND ON CERTAIN ANOMALOUS MODES OF REPRODUCTION AND VARIATION.

BUD-VARIATION IN THE PEACH, PLUM, CHERRY, VINE, GOOSEBERRY, CURRANT, AND BANANA, AS SHOWN BY THE MODIFIED FRUIT. IN FLOWERS: CAMELLIAS, AZALEAS, CHRYSANTHEMUMS, ROSES, ETC. ON THE RUNNING OF THE COLOUR IN CARNATIONS. BUD-VARIATIONS IN LEAVES. VARIATIONS BY SUCKERS, TUBERS, AND BULBS. ON THE BREAKING OF TULIPS. BUD-VARIATIONS GRADUATE INTO CHANGES CONSEQUENT ON CHANGED CONDITIONS OF LIFE. GRAFT-HYBRIDS. ON THE SEGREGATION OF THE PARENTAL CHARACTERS IN SEMINAL HYBRIDS BY BUD- VARIATION. ON THE DIRECT OR IMMEDIATE ACTION OF FOREIGN POLLEN ON THE MOTHER-PLANT. ON THE EFFECTS IN FEMALE ANIMALS OF A PREVIOUS IMPREGNATION ON THE SUBSEQUENT OFFSPRING. CONCLUSION AND SUMMARY.

This chapter will be chiefly devoted to a subject in many respects important, namely, bud-variation. By this term I include all those sudden changes in structure or appearance which occasionally occur in full-grown plants in their flower-buds or leaf-buds. Gardeners call such changes "Sports;" but this, as previously remarked, is an ill-defined expression, as it has often been applied to strongly marked variations in seedling plants. The difference between seminal and bud reproduction is not so great as it at first appears; for each bud is in one sense a new and distinct individual; but such individuals are produced through the formation of various kinds of buds without the aid of any special apparatus, whilst fertile seeds are produced by the concourse of the two sexual elements. The modifications which arise through bud-variation can generally be propagated to any extent by grafting, budding, cuttings, bulbs, etc., and occasionally even by seed. Some few of our most beautiful and useful productions have arisen by bud-variation.

Bud-variations have as yet been observed only in the vegetable kingdom; but it is probable that if compound animals, such as corals, etc., had been subjected to a long course of domestication, they would have varied by buds; for they resemble plants in many respects. For instance, any new or peculiar character presented by a compound animal is propagated by budding, as occurs with differently coloured Hydras, and as Mr. Gosse has shown to be the case with a singular variety of a true coral. Varieties of the Hydra have also been grafted on other varieties, and have retained their character.

I will in the first place give all the cases of bud variations which I have been able to collect, and afterwards show their importance. (11/1. Since the publication of the first edition of this work, I have found that M. Carriere, Chef des Pepinieres au Mus. d'Hist. Nat. in his excellent Essay 'Production et Fixation des Varietes' 1865 has given a list of bud- variations far more extensive than mine; but as these relate chiefly to cases occurring in France I have left my list as it stood, adding a few facts from M. Carriere and others. Any one who wishes to study the subject fully should refer to M. Carriere's Essay.) These cases prove that those authors who, like Pallas, attribute all variability to the crossing either of distinct races, or of distinct individuals belonging to the same race but somewhat different from each other, are in error; as are those authors who attribute all variability to the mere act of sexual union. Nor can we account in all cases for the appearance through bud-variation of new characters by the principle of reversion to long-lost characters. He who wishes to judge how far the conditions of life directly cause each particular variation ought to reflect well on the cases immediately to be given. I will commence with bud-variations, as exhibited in the fruit, and then pass on to flowers, and finally to leaves.

[PEACH (Amygdalus persica).

In the last chapter I gave two cases of a peach-almond and a double- flowered almond which suddenly produced fruit closely resembling true peaches. I have also given many cases of peach-trees producing buds, which, when developed into branches, have yielded nectarines. We have seen that no less than six named and several unnamed varieties of the peach have thus produced several varieties of nectarine. I have shown that it is highly improbable that all these peach-trees, some of which are old varieties, and have been propagated by the million, are hybrids from the peach and nectarine, and that it is opposed to all analogy to attribute the occasional production of nectarines on peach-trees to the direct action of pollen from some neighbouring nectarine-tree. Several of the cases are highly remarkable, because, firstly, the fruit thus produced has sometimes been in part a nectarine and in part a peach; secondly, because nectarines thus suddenly produced have reproduced themselves by seed; and thirdly, because nectarines are produced from peach-trees from seed as well as from buds. The seed of the nectarine, on the other hand, occasionally produces peaches; and we have seen in one instance that a nectarine-tree yielded peaches by bud-variation. As the peach is certainly the oldest or primary variety, the production of peaches from nectarines, either by seeds or buds, may perhaps be considered as a case of reversion. Certain trees have also been described as indifferently bearing peaches or nectarines, and this may be considered as bud-variation carried to an extreme degree.

The grosse mignonne peach at Montreuil produced "from a sporting branch" the grosse mignonne tardive, "a most excellent variety," which ripens its fruit a fortnight later than the parent tree, and is equally good. (11/2. 'Gardener's Chronicle' 1854 page 821.) This same peach has likewise produced by bud-variation the early grosse mignonne. Hunt's large tawny nectarine "originated from Hunt's small tawny nectarine, but not through seminal reproduction." (11/3. Lindley 'Guide to Orchard' as quoted in 'Gardener's Chronicle' 1852 page 821. For the Early mignonne peach see 'Gardener's Chronicle' 1864 page 1251.)

PLUMS.

Mr. Knight states that a tree of the yellow magnum bonum plum, forty years old, which had always borne ordinary fruit, produced a branch which yielded red magnum bonums. (11/4. 'Transact. Hort. Soc.' volume 2 page 160.) Mr. Rivers, of Sawbridgeworth, informs me (January 1863) that a single tree out of 400 or 500 trees of the Early Prolific plum, which is a purple kind, descended from an old French variety bearing purple fruit, produced when about ten years old bright yellow plums; these differed in no respect except colour from those on the other trees, but were unlike any other known kind of yellow plum. (11/5. See also 'Gardener's Chronicle' 1863 page 27.)

CHERRY (Prunus cerasus).

Mr. Knight has recorded (ibid) the case of a branch of a May-Duke cherry, which, though certainly never grafted, always produced fruit, ripening later, and more oblong than the fruit on the other branches. Another account has been given of two May-Duke cherry-trees in Scotland, with branches bearing oblong and very fine fruit, which invariably ripened, as in Knight's case, a fortnight later than the other cherries. (11/6. 'Gardener's Chronicle' 1852 page 821.) M. Carriere gives (page 37) numerous analogous cases, and one of the same tree bearing three kinds of fruit.

GRAPES (Vitis vinifera).

The black or purple Frontignan in one case produced during two successive years (and no doubt permanently), spurs which bore white Frontignan grapes. In another case, on the same footstalk, the lower berries "were well- coloured black Frontignans; those next the stalk were white, with the exception of one black and one streaked berry;" and altogether there were fifteen black and twelve white berries on the same stalk. In another kind of grape, black and amber-coloured berries were produced in the same cluster. (11/7. 'Gardener's Chronicle' 1852 page 629; 1856 page 648; 1864 page 986. Other cases are given by Braun 'Rejuvenescence' in 'Ray Soc. Bot. Mem.' 1853 page 314.) Count Odart describes a variety which often bears on the same stalk small round and large oblong berries; though the shape of the berry is generally a fixed character. (11/8. 'Ampelographie' etc. 1849 page 71.) Here is another striking case given on the excellent authority of M. Carriere (11/9. 'Gardener's Chronicle' 1866 page 970.): "a black Hamburg grape (Frankenthal) was cut down, and produced three suckers; one of these was layered, and after a time produced much smaller berries, which always ripened at least a fortnight earlier than the others. Of the remaining two suckers, one produced every year fine grapes, whilst the other, although it set an abundance of fruit, matured only a few, and these of inferior quality."

GOOSEBERRY (Ribes grossularia).

A remarkable case has been described by Dr. Lindley (11/10. 'Gardener's Chronicle' 1855 pages 597, 612.) of a bush which bore at the same time no less than four kinds of berries, namely, hairy and red,—smooth, small and red,—green,—and yellow tinged with buff; the two latter kinds had a different flavour from the red berries, and their seeds were coloured red. Three twigs on this bush grew close together; the first bore three yellow berries and one red; the second twig bore four yellow and one red; and the third four red and one yellow. Mr. Laxton also informs me that he has seen a Red Warrington gooseberry bearing both red and yellow fruit on the same branch.

CURRANT (Ribes rubrum).

A bush purchased as the Champagne, which is a variety that bears blush- coloured fruit intermediate between red and white, produced during fourteen years on separate branches and mingled on the same branch, berries of the red, white, and champagne kinds. (11/11. 'Gardener's Chronicle' 1842 page 873; 1855 page 646. In the 'Chronicle' page 876 Mr. P. Mackenzie states that the bush still continues to bear the three kinds of fruit, "although they have not been every year alike.) The suspicion naturally arises that this variety may have originated from a cross between a red and white variety, and that the above transformation may be accounted for by reversion to both parent-forms; but from the foregoing complex case of the gooseberry this view is doubtful. In France, a branch of a red-currant bush, about ten years old, produced near the summit five white berries) and lower down, amongst the red berries, one berry half red and half white. (11/12. 'Revue Horticole' quoted in 'Gardener's Chronicle' 1844 page 87.) Alexander Braun (11/13. 'Rejuvenescence in Nature' 'Bot. Memoirs Ray Soc.' 1853 page 314.) also has often seen branches on white currant-trees bearing red berries.

PEAR (Pyrus communis).

Dureau de la Malle states that the flowers on some trees of an ancient variety, the doyenne galeux, were destroyed by frost: other flowers appeared in July, which produced six pears; these exactly resembled in their skin and taste the fruit of a distinct variety, the gros doyenne blanc, but in shape were like the bon-chretien: it was not ascertained whether this new variety could be propagated by budding or grafting. The same author grafted a bon-chretien on a quince, and it produced, besides its proper fruit, an apparently new variety, of a peculiar form with thick and rough skin. (11/14. 'Comptes Rendus' tome 41 1855 page 804. The second case is given on the authority of Gaudichaud ibid tome 34 1852 page 748.)

APPLE (Pyrus malus).

In Canada, a tree of the variety called Pound Sweet, produced (11/15. This case is given in the 'Gardener's Chronicle' 1867 page 403.), between two of its proper fruit, an apple which was well russeted, small in size, different in shape, and with a short peduncle. As no russet apple grew anywhere near, this case apparently cannot be accounted for by the direct action of foreign pollen. M. Carriere (page 38) mentions an analogous instance. I shall hereafter give cases of apple-trees which regularly produce fruit of two kinds, or half-and-half fruit; these trees are generally supposed, and probably with truth, to be of crossed parentage, and that the fruit reverts to both parent-forms.

BANANA (Musa sapientium).

Sir R. Schomburgk states that he saw in St. Domingo a raceme on the Fig Banana which bore towards the base 125 fruits of the proper kind; and these were succeeded, as is usual, higher up the raceme, by barren flowers, and these by 420 fruits, having a widely different appearance, and ripening earlier than the proper fruit. The abnormal fruit closely resembled, except in being smaller, that of the Musa chinensis or cavendishii, which has generally been ranked as a distinct species. (11/16. 'Journal of Proc. Linn. Soc.' volume 2 Botany page 131.)

FLOWERS.

Many cases have been recorded of a whole plant, or single branch, or bud, suddenly producing flowers different from the proper type in colour, form, size, doubleness, or other character. Half the flower, or a smaller segment, sometimes changes colour.

CAMELLIA.

The myrtle-leaved species (C. myrtifolia), and two or three varieties of the common species, have been known to produce hexagonal and imperfectly quadrangular flowers; and the branches producing such flowers have been propagated by grafting. (11/17. 'Gardener's Chronicle' 1847 page 207.) The Pompon variety often bears "four distinguishable kinds of flowers,—the pure white and the red-eyed, which appear promiscuously; the brindled pink and the rose-coloured, which may be kept separate with tolerable certainty by grafting from the branches that bear them." A branch, also, on an old tree of the rose-coloured variety has been seen to "revert to the pure white colour, an occurrence less common than the departure from it." (11/18. Herbert 'Amaryllidaceae' 1838 page 369.)

Crataegus oxyacantha.

A dark pink hawthorn has been known to throw out a single tuft of pure white blossoms (11/19. 'Gardener's Chronicle' 1843 page 391.); and Mr. A. Clapham, nurseryman, of Bedford, informs me that his father had a deep crimson thorn grafted on a white thorn, which during several years, always bore, high above the graft, bunches of white, pink and deep crimson flowers.

Azalea indica.

A. indica is well known often to produce new varieties by buds. I have myself seen several cases. A plant of Azalea indica variegata has been exhibited bearing a truss of flowers of A. ind. gledstanesii "as true as could possibly be produced, thus evidencing the origin of that fine variety." On another plant of A. ind. variegata a perfect flower of A. ind. lateritia was produced; so that both gledstanesii and lateritia no doubt originally appeared as sporting branches of A. ind. variegata. (11/20. Exhibited at Hort. Soc. London. Report in 'Gardener's Chronicle' 1844 page 337.)

HIBISCUS (Paritium tricuspis).

A seedling of this plant, when some years old, produced, at Saharunpore (11/21. Mr. W. Bell 'Bot. Soc. of Edinburgh' May 1863.), some branches "which bore leaves and flowers widely different from the normal form." "The abnormal leaf is much less divided, and not acuminated. The petals are considerably larger, and quite entire. There is also in the fresh state a conspicuous, large, oblong gland, full of a viscid secretion, on the back of each of the calycine segments." Dr. King, who subsequently had charge of these Gardens, informs me that a tree of Paritium tricuspis (probably the very same plant) growing there, had a branch buried in the ground, apparently by accident; and this branch changed its character wonderfully, growing like a bush, and producing flowers and leaves, resembling in shape those of another species, viz., P. tiliaceum. A small branch springing from this bush near the ground, reverted to the parent-form. Both forms were extensively propagated during several years by cuttings and kept perfectly true.

Althaea rosea.

A double yellow Hollyhock suddenly turned one year into a pure white single kind; subsequently a branch bearing the original double yellow flowers reappeared in the midst of the branches of the single white kind. (11/22. 'Revue Horticole' quoted in 'Gardener's Chronicle' 1845 page 475.)

PELARGONIUM.

These highly cultivated plants seem eminently liable to bud-variation. I will give only a few well-marked cases. Gartner has seen (11/23. 'Bastarderzeugung' 1849 s. 76.) a plant of P. zonale with a branch having white edges, which remained constant for years, and bore flowers of a deeper red than usual. Generally speaking, such branches present little or no difference in their flowers: thus a writer (11/24. 'Journal of Horticulture' 1861 page 336.) pinched off the leading shoot of a seedling P. zonale, and it threw out three branches, which differed in the size and colour of their leaves and stems; but on all three branches "the flowers were identical," except in being largest in the green-stemmed variety, and smallest in that with variegated foliage: these three varieties were subsequently propagated and distributed. Many branches, and some whole plants, of a variety called compactum, which bears orange-scarlet flowers, have been seen to produce pink flowers. (11/25. W.P. Ayres 'Gardener's Chronicle' 1842 page 791.) Hill's Hector, which is a pale red variety, produced a branch with lilac flowers, and some trusses with both red and lilac flowers. This apparently is a case of reversion, for Hill's Hector was a seedling from a lilac variety. (11/26. W.P. Ayres ibid.) Here is a better case of reversion: a variety produced from a complicated cross, after having been propagated for five generations by seed, yielded by bud- variation three very distinct varieties which were undistinguishable from plants, "known to have been at some time ancestors of the plant in question." (11/27. Dr. Maxwell Masters 'Pop. Science Review' July 1872 page 250.) Of all Pelargoniums, Rollisson's Unique seems to be the most sportive; its origin is not positively known, but is believed to be from a cross. Mr. Salter, of Hammersmith, states (11/28. 'Gardener's Chronicle' 1861 page 968.) that he has himself known this purple variety to produce the lilac, the rose-crimson or conspicuum, and the red or coccineum varieties; the latter has also produced the rose d'amour; so that altogether four varieties have originated by bud variation from Rollisson's Unique. Mr. Salter remarks that these four varieties "may now be considered as fixed, although they occasionally produce flowers of the original colour. This year coccineum has pushed flowers of three different colours, red, rose, and lilac, upon the same truss, and upon other trusses are flowers half red and half lilac." Besides these four varieties, two other scarlet Uniques are known to exist, both of which occasionally produce lilac flowers identical with Rollisson's Unique (11/29. Ibid 1861 page 945.); but one at least of these did not arise through bud-variation, but is believed to be a seedling from Rollisson's Unique. (11/30. W. Paul 'Gardener's Chronicle' 1861 page 968.) There are, also, in the trade (11/31. Ibid page 945.) two other slightly different varieties, of unknown origin, of Rollisson's Unique: so that altogether we have a curiously complex case of variation both by buds and seeds. (11/32. For other cases of bud-variation in this same variety see 'Gardener's Chronicle' 1861 pages 578, 600, 925. For other distinct cases of bud-variation in the genus Pelargonium see 'Cottage Gardener' 1860 page 194.) Here is a still more complex case: M. Rafarin states that a pale rose-coloured variety produced a branch bearing deep red flowers. "Cuttings were taken from this 'sport,' from which 20 plants were raised, which flowered in 1867, when it was found that scarcely two were alike." Some resembled the parent-form, some resembled the sport, some bore both kinds of flowers; and even some of the petals on the same flower were rose-coloured and others red. (11/33. Dr. Maxwell Masters 'Pop. Science Review' July 1872 page 254.) An English wild plant, the Geranium pratense, when cultivated in a garden, has been seen to produce on the same plant both blue and white, and striped blue and white flowers. (11/34. Rev. W.T. Bree in Loudon's 'Gardener's Mag.' volume 8 1832 page 93.)

CHRYSANTHEMUM.

This plant frequently sports, both by its lateral branches and occasionally by suckers. A seedling raised by Mr. Salter has produced by bud-variation six distinct sorts, five different in colour and one in foliage, all of which are now fixed. (11/35. 'The Chrysanthemum: its History and Culture' by J. Salter 1865 page 41 etc.) A variety called cedo nulli bears small yellow flowers, but habitually produces branches with white flowers; and a specimen was exhibited, which Prof. T. Dyer saw, before the Horticultural Society. The varieties which were first introduced from China were so excessively variable, "that it was extremely difficult to tell which was the original colour of the variety, and which was the sport." The same plant would produce one year only buff-coloured, and next year only rose- coloured flowers; and then would change again, or produce at the same time flowers of both colours. These fluctuating varieties are now all lost, and, when a branch sports into a new variety, it can generally be propagated and kept true; but, as Mr. Salter remarks, "every sport should be thoroughly tested in different soils before it can be really considered as fixed, as many have been known to run back when planted in rich compost; but when sufficient care and time are expended in proving, there will exist little danger of subsequent disappointment." Mr. Salter informs me that with all the varieties the commonest kind of bud-variation is the production of yellow flowers, and, as this is the primordial colour, these cases may be attributed to reversion. Mr. Salter has given me a list of seven differently coloured chrysanthemums, which have all produced branches with yellow flowers; but three of them have also sported into other colours. With any change of colour in the flower, the foliage generally changes in a corresponding manner in lightness or darkness.

Another Compositous plant, namely, Centauria cyanus, when cultivated in a garden, not unfrequently produces on the same root flowers of four different colours, viz., blue, white, dark-purple, and parti-coloured. (11/36. Bree in Loudon's 'Gardener's Mag.' volume 8 1832 page 93.) The flowers of Anthemis also vary on the same plant. (11/37. Bronn 'Geschichte der Natur' b. 2 s. 123.)

ROSES.

Many varieties of the Rose are known or are believed to have originated by bud-variation. (11/38. T. Rivers 'Rose Amateur's Guide' 1837 page 4.) The common double moss-rose was imported into England from Italy about the year 1735. (11/39. Mr. Shailer quoted in 'Gardener's Chronicle' 1848 page 759.) Its origin is unknown, but from analogy it probably arose from the Provence rose (R. centifolia) by bud-variation; for the branches of the common moss- rose have several times been known to produce Provence roses, wholly or partially destitute of moss: I have seen one such instance, and several others have been recorded. (11/40. 'Transact. Hort. Soc.' volume 4 1822 page 137; 'Gardener's Chronicle' 1842 page 422.) Mr. Rivers also informs me that he raised two or three roses of the Provence class from seed of the old single moss-rose (11/41. See also Loudon's 'Arboretum' volume 2 page 780.); and this latter kind was produced in 1807 by bud-variation from the common moss-rose. The white moss-rose was also produced in 1788 by an offset from the common red moss-rose: it was at first pale blush-coloured, but became white by continued budding. On cutting down the shoots which had produced this white moss-rose, two weak shoots were thrown up, and buds from these yielded the beautiful striped moss-rose. The common moss-rose has yielded by bud-variation, besides the old single red moss-rose, the old scarlet semi-double moss-rose, and the sage-leaf moss-rose, which "has a delicate shell-like form, and is of a beautiful blush colour; it is now (1852) nearly extinct." (11/42. All these statements on the origin of the several varieties of the moss-rose are given on the authority of Mr. Shailer, who, together with his father, was concerned in their original propagation. See 'Gard. Chronicle' 1852 page 759.) A white moss-rose has been seen to bear a flower half white and half pink. (11/43. 'Gardener's Chronicle' 1845 page 564.) Although several moss-roses have thus certainly arisen by bud-variation, the greater number probably owe their origin to seed of moss-roses. For Mr. Rivers informs me that his seedlings from the old single moss-rose almost always produced moss-roses; and the old single moss-rose was, as we have seen, the product by bud-variation of the double moss-rose originally imported from Italy. That the original moss-rose was the product of bud-variation is probable, from the facts above given and from the de Meaux moss-rose (also a variety of R. centifolia) (11/44. 'Transact. Hort. Soc.' volume 2 page 242.) having appeared as a sporting branch on the common rose de Meaux. Prof. Caspary has carefully described (11/45. 'Shriften der Phys. Oekon. Gesell. zu Konigsberg' February 3, 1865 s. 4. See also Dr. Caspary's paper in 'Transactions of the Hort. Congress of Amsterdam' 1865.) the case of a six-year-old white moss-rose, which sent up several suckers, one of which was thorny, and produced red flowers, destitute of moss, exactly like those of the Provence rose (R. centifolia): another shoot bore both kinds of flowers, and in addition longitudinally striped flowers. As this white moss-rose had been grafted on the Provence rose, Prof. Caspary attributes the above changes to the influence of the stock; but from the facts already given, and from others to be given, bud- variation, with reversion, is probably a sufficient explanation.

Many other instances could be added of roses varying by buds. The white Provence rose apparently originated in this way. (11/46. 'Gardener's Chronicle' 1852 page 759.) M. Carriere states (page 36) that he himself knows of five varieties thus produced by the Baronne Prevost. The double and highly-coloured Belladonna rose has produced by suckers both semi- double and almost single white roses (11/47. 'Transact. Hort. Soc.' volume 2 page 242.); whilst suckers from one of these semi-double white roses reverted to perfectly characterised Belladonnas. In St. Domingo, varieties of the China rose propagated by cuttings often revert after a year or two into the old China rose. (11/48. Sir R. Schomburgk 'Proc. Linn. Soc. Bot.' volume 2 page 132.) Many cases have been recorded of roses suddenly becoming striped or changing their character by segments: some plants of the Comtesse de Chabrillant, which is properly rose-coloured, were exhibited in 1862 (11/49. 'Gardener's Chronicle' 1862 page 619.), with crimson flakes on a rose ground. I have seen the Beauty of Billiard with a quarter and with half the flower almost white. 'The Austrian bramble (R. lutea) not rarely (11/50. Hopkirk 'Flora Anomala' 167.) produces branches with pure yellow flowers; and Prof. Henslow has seen exactly half the flower of a pure yellow, and I have seen narrow yellow streaks on a single petal, of which the rest was of the usual copper colour.

The following cases are highly remarkable. Mr. Rivers, as I am informed by him, possessed a new French rose with delicate smooth shoots, pale glaucous-green leaves, and semi-double pale flesh-coloured flowers striped with dark red; and on branches thus characterised there suddenly appeared in more than one instance, the famous old rose called the Baronne Prevost, with its stout thorny shoots, and immense, uniformly and richly coloured double flowers; so that in this case the shoots, leaves, and flowers, all at once changed their character by bud-variation. According to M. Verlot (11/51. 'Sur La Production et la Fixation des Varietes' 1865 page 4.), a variety called Rosa cannabifolia, which has peculiarly shaped leaflets, and differs from every member of the family in the leaves being opposite instead of alternate, suddenly appeared on a plant of R. alba in the gardens of the Luxembourg. Lastly, "a running shoot" was observed by Mr. H. Curtis (11/52. 'Journal of Horticulture' March 1865 page 233.) on the old Aimee Vibert Noisette, and he budded it on Celine; thus a climbing Aimee Vibert was first produced and afterwards propagated.

DIANTHUS.

It is quite common with the Sweet William (D. barbatus) to see differently coloured flowers on the same root; and I have observed on the same truss four differently coloured and shaded flowers. Carnations and pinks (D. caryophyllus, etc.) occasionally vary by layers; and some kinds are so little certain in character that they are called by floriculturists "catch- flowers." (11/53. 'Gardener's Chronicle' 1843 page 135.) Mr. Dickson has ably discussed the "running" of particoloured or striped carnations, and says it cannot be accounted for by the compost in which they are grown: "layers from the same clean flower would come part of them clean and part foul, even when subjected to precisely the same treatment; and frequently one flower alone appears influenced by the taint, the remainder coming perfectly clean." (11/54. Ibid 1842 page 55.) This running of the parti- coloured flowers apparently is a case of reversion by buds to the original uniform tint of the species.

I will briefly mention some other cases of bud-variation to show how many plants belonging to many orders have varied in their flowers; and many others might be added. I have seen on a snap-dragon (Antirrhinum majus) white, pink, and striped flowers on the same plant, and branches with striped flowers on a red-coloured variety. On a double stock (Matthiola incana) I have seen a branch bearing single flowers; and on a dingy-purple double variety of the wall-flower (Cheiranthus cheiri), a branch which had reverted to the ordinary copper colour. On other branches of the same plant, some flowers were exactly divided across the middle, one half being purple and the other coppery; but some of the smaller petals towards the centre of these same flowers were purple longitudinally streaked with coppery colour, or coppery streaked with purple. A Cyclamen (11/55. 'Gardener's Chronicle' 1867 page 235.) has been observed to bear white and pink flowers of two forms, the one resembling the Persicum strain, and the other the Coum strain. Oenothera biennis has been seen (11/56. Gartner 'Bastarderzeugung' s. 305.) bearing flowers of three different colours. The hybrid Gladiolus colvilii occasionally bears uniformly coloured flowers, and one case is recorded (11/57. Mr. D. Beaton in 'Cottage Gardener' 1860 page 250.) of all the flowers on a plant thus changing colour. A Fuchsia has been seen (11/58. 'Gardener's Chronicle' 1850 page 536.) bearing two kinds of flowers. Mirabilis jalapa is eminently sportive, sometimes bearing on the same root pure red, yellow, and white flowers, and others striped with various combinations of these three colours. (11/59. Braun 'Ray Soc. Bot. Mem.' 1853 page 315; Hopkirk 'Flora Anomala' page 164; Lecoq 'Geograph. Bot. de l'Europe' tome 3 1854 page 405; and 'De la Fecondation' 1862 page 303.) The plants of the Mirabilis, which bear such extraordinarily variable flowers in most, probably in all, cases, owe their origin, as shown by Prof. Lecoq, to crosses between differently coloured varieties.

LEAVES AND SHOOTS.

Changes, through bud-variation, in fruits and flowers have hitherto been treated of; incidentally some remarkable modifications in the leaves and shoots of the rose and Paritium, and in a lesser degree in the foliage of the Pelargonium and Chrysanthemum, have been noticed. I will now add a few more cases of variation in leaf-buds. Verlot (11/60. 'Des Varietes' 1865 page 5.) states that on Aralia trifoliata, which properly has leaves with three leaflets, branches frequently appear bearing simple leaves of various forms; these can be propagated by buds or by grafting, and have given rise, as he states, to several nominal species.

With respect to trees, the history of but few of the many varieties with curious or ornamental foliage is known; but several probably have originated by bud-variation. Here is one case:—An old ash-tree (Fraxinus excelsior) in the grounds of Necton, as Mr. Mason states, "for many years has had one bough of a totally different character to the rest of the tree, or of any other ash-tree which I have seen; being short-jointed and densely covered with foliage." It was ascertained that this variety could be propagated by grafts. (11/61. W. Mason in 'Gardener's Chronicle' 1843 page 878.) The varieties of some trees with cut leaves, as the oak-leaved laburnum, the parsley-leaved vine, and especially the fern-leaved beech, are apt to revert by buds to the common forms. (11/62. Alex. Braun 'Ray Soc. Bot. Mem.' 1853 page 315; 'Gardener's Chron.' 1841 page 329.) The fern-like leaves of the beech sometimes revert only partially, and the branches display here and there sprouts bearing common leaves, fern-like, and variously shaped leaves. Such cases differ but little from the so- called heterophyllus varieties, in which the tree habitually bears leaves of various forms; but it is probable that most heterophyllous trees have originated as seedlings. There is a sub-variety of the weeping willow with leaves rolled up into a spiral coil; and Mr. Masters states that a tree of this kind kept true in his garden for twenty-five years, and then threw out a single upright shoot bearing flat leaves. (11/63. Dr. M.T. Masters 'Royal Institution Lecture' March 16, 1860.)

I have often noticed single twigs and branches on beech and other trees with their leaves fully expanded before those on the other branches had opened; and as there was nothing in their exposure or character to account for this difference, I presume that they had appeared as bud-variations, like the early and late fruit-maturing varieties of the peach and nectarine.

Cryptogamic plants are liable to bud-variation, for fronds on the same fern often display remarkable deviations of structure. Spores, which are of the nature of buds, taken from such abnormal fronds, reproduce, with remarkable fidelity, the same variety, after passing through the sexual stage. (11/64. See Mr. W.K. Bridgeman's curious paper in 'Annals and Mag. of Nat. Hist.' December 1861; also Mr. J. Scott 'Bot. Soc. Edinburgh' June 12, 1862.)

With respect to colour, leaves often become by bud-variation zoned, blotched, or spotted with white, yellow, and red; and this occasionally occurs even with plants in a state of nature. Variegation, however, appears still more frequently in plants produced from seed; even the cotyledons or seed-leaves being thus affected. (11/65. 'Journal of Horticulture' 1861 page 336; Verlot 'Des Varietes' page 76.) There have been endless disputes whether variegation should be considered as a disease. In a future chapter we shall see that it is much influenced, both in the case of seedlings and of mature plants, by the nature of the soil. Plants which have become variegated as seedlings, generally transmit their character by seed to a large proportion of their progeny; and Mr. Salter has given me a list of eight genera in which this occurred. (11/66. See also Verlot 'Des Varietes' page 74.) Sir F. Pollock has given me more precise information: he sowed seed from a variegated plant of Ballota nigra which was found growing wild, and thirty per cent of the seedlings were variegated; seed from these latter being sown, sixty per cent came up variegated. When branches become variegated by bud-variation, and the variety is attempted to be propagated by seed, the seedlings are rarely variegated: Mr. Salter found this to be the case with plants belonging to eleven genera, in which the greater number of the seedlings proved to be green-leaved; yet a few were slightly variegated, or were quite white, but none were worth keeping. Variegated plants, whether originally produced from seeds or buds, can generally be propagated by budding, grafting, etc.; but all are apt to revert by bud- variation to their ordinary foliage. This tendency, however, differs much in the varieties of even the same species; for instance, the golden-striped variety of Euonymus japonicus "is very liable to run back to the green- leaved, while the silver-striped variety hardly ever changes." (11/67. 'Gardener's Chron.' 1844 page 86.) I have seen a variety of the holly, with its leaves having a central yellow patch, which had everywhere partially reverted to the ordinary foliage, so that on the same small branch there were many twigs of both kinds. In the pelargonium, and in some other plants, variegation is generally accompanied by some degree of dwarfing, as is well exemplified in the "Dandy" pelargonium. When such dwarf varieties sport back by buds or suckers to the ordinary foliage, the dwarfed stature still remains. (11/68. Ibid 1861 page 963.) It is remarkable that plants propagated from branches which have reverted from variegated to plain leaves (11/69. Ibid 1861 page 433; 'Cottage Gardener' 1860 page 2.) do not always (or never, as one observer asserts) perfectly resemble the original plain-leaved plant from which the variegated branch arose: it seems that a plant, in passing by bud-variation from plain leaves to variegated, and back again from variegated to plain, is generally in some degree affected so as to assume a slightly different aspect.

BUD-VARIATION BY SUCKERS, TUBERS, AND BULBS.

All the cases hitherto given of bud-variation in fruits, flowers, leaves, and shoots, have been confined to buds on the stems or branches, with the exception of a few cases incidentally noticed of varying suckers in the rose, pelargonium, and chrysanthemum. I will now give a few instances of variation in subterranean buds, that is, by suckers, tubers, and bulbs; not that there is any essential difference between buds above and beneath the ground. Mr. Salter informs me that two variegated varieties of Phlox originated as suckers; but I should not have thought these worth mentioning, had not Mr. Salter found, after repeated trials, that he could not propagate them by "root-joints," whereas, the variegated Tussilago farfara can thus be safely propagated (11/70. M. Lemoine quoted in 'Gardener's Chronicle' 1867 page 74 has lately observed that the Symphytum with variegated leaves cannot be propagated by division of the roots. He also found that out of 500 plants of a Phlox with striped flowers, which had been propagated by root-division, only seven or eight produced striped flowers. See also on striped Pelargoniums 'Gardener's Chronicle' 1867 page 1000.); but this latter plant may have originated as a variegated seedling, which would account for its greater fixedness of character. The Barberry (Berberis vulgaris) offers an analogous case; there is a well-known variety with seedless fruit, which can be propagated by cuttings or layers; but suckers always revert to the common form, which produces fruit containing seeds. (11/71. Anderson 'Recreations in Agriculture' volume 5 page 152.) My father repeatedly tried this experiment, and always with the same result. I may here mention that maize and wheat sometimes produce new varieties from the stock or root, as does the sugar-cane. (11/72. For wheat see 'Improvement of the Cereals' by P. Shirreff 1873 page 47. For maize and sugar-cane Carriere ibid pages 40, 42. With respect to the sugar-cane Mr. J. Caldwell of Mauritius says ('Gardener's Chronicle' 1874 page 316) the Ribbon cane has here "sported into a perfectly green cane and a perfectly red cane from the same head. I verified this myself, and saw at least 200 instances in the same plantation, and the fact has completely upset all our preconceived ideas of the difference of colour being permanent. The conversion of a striped cane into a green cane was not uncommon, but the change into a red cane universally disbelieved, and that both events should occur in the same plant incredible. I find, however, in Fleischman's 'Report on Sugar Cultivation in Louisiana for 1848 by the American Patent Office, the circumstance is mentioned, but he says he never saw it himself.")

Turning now to tubers: in the common Potato (Solanum tuberosum) a single bud or eye sometimes varies and produces a new variety; or, occasionally, and this is a much more remarkable circumstance, all the eyes in a tuber vary in the same manner and at the same time, so that the whole tuber assumes a new character. For instance, a single eye in a tuber of the old FORTY-FOLD POTATO, which is a purple variety, was observed (11/73. 'Gardener's Chronicle' 1857 page 662.) to become white; this eye was cut out and planted separately, and the kind has since been largely propagated. KEMP'S POTATO is properly white, but a plant in Lancashire produced two tubers which were red, and two which were white; the red kind was propagated in the usual manner by eyes, and kept true to its new colour, and, being found a more productive variety, soon became widely known under the name of TAYLOR'S FORTY-FOLD. (11/74. 'Gard. Chronicle' 1841 page 814.) The old FORTY-FOLD POTATO, as already stated, is a purple variety; but a plant long cultivated on the same ground produced, not, as in the case above given, a single white eye, but a whole white tuber, which has since been propagated and keeps true. (11/75. Ibid 1857 page 613.) Several cases have been recorded of large portions of whole rows of potatoes slightly changing their character. (11/76. Ibid 1857 page 679. See also Philips 'Hist. of Vegetables' volume 2 page 91 for other and similar accounts.)

Dahlias propagated by tubers under the hot climate of St. Domingo vary much; Sir R. Schomburgk gives the case of the "Butterfly variety," which the second year produced on the same plant "double and single flowers; here white petals edged with maroon; there of a uniform deep maroon." (11/77. 'Journal of Proc. Linn. Soc.' volume 2 Botany page 132.) Mr. Bree also mentions a plant "which bore two different kinds of self-coloured flowers, as well as a third kind which partook of both colours beautifully intermixed." (11/78. Loudon's 'Gardener's Mag.' volume 8 1832 page 94.) Another case is described of a dahlia with purple flowers which bore a white flower streaked with purple. (11/79. 'Gard. Chronicle' 1850 page 536; and 1842 page 729.)

Considering how long and extensively many Bulbous plants have been cultivated, and how numerous are the varieties produced from seed, these plants have not perhaps varied so much by offsets,—that is, by the production of new bulbs,—as might have been expected. With the Hyacinth, however, several instances have been given by M. Carriere. A case also has been recorded of a blue variety which for three successive years gave offsets producing white flowers with a red centre. (11/80. 'Des Jacinthes' etc. Amsterdam 1768 page 122.) Another hyacinth bore (11/81. 'Gardener's Chronicle' 1845 page 212.) on the same truss a perfectly pink and a perfectly blue flower. I have seen a bulb producing at the same time one stalk or truss with fine blue flowers, another with fine red flowers, and a third with blue flowers on one side and red on the other; several of the flowers being also longitudinally striped red and blue.

Mr. John Scott informs me that in 1862 Imatophyllum miniatum, in the Botanic Gardens of Edinburgh, threw up a sucker which differed from the normal form, in the leaves being two-ranked instead of four-ranked. The leaves were also smaller, with the upper surface raised instead of being channelled.

In the propagation of TULIPS, seedlings are raised, called selfs or breeders, which, "consist of one plain colour on a white or yellow bottom. These, being cultivated on a dry and rather poor soil, become broken or variegated and produce new varieties. The time that elapses before they break varies from one to twenty years or more, and sometimes this change never takes place." (11/82. Loudon's 'Encyclopaedia of Gardening' page 1024.) The broken or variegated colours which give value to all tulips are due to bud-variation; for although the Bybloemens and some other kinds have been raised from several distinct breeders, yet all the Baguets are said to have come from a single breeder or seedling. This bud-variation, in accordance with the views of MM. Vilmorin and Verlot (11/83. 'Production des Varietes' 1865 p. 63.) is probably an attempt to revert to that uniform colour which is natural to the species. A tulip, however, which has already become broken, when treated with too strong manure, is liable to flush or lose by a second act of reversion its variegated colours. Some kinds, as Imperatrix Florum, are much more liable than others to flushing; and Mr. Dickson maintains (11/84. 'Gardener's Chronicle' 1841 page 782; 1842 page 55.) that this can no more be accounted for than the variation of any other plant. He believes that English growers, from care in choosing seed from broken flowers instead of from plain flowers, have to a certain extent diminished the tendency in flowers already broken to flushing or secondary reversion. Iris xiphium, according to M. Carriere (page 65), behaves in nearly the same manner, as do so many tulips.

During two consecutive years all the early flowers in a bed of Tigridia conchiflora (11/85. 'Gardener's Chron.' 1849 page 565.) resembled those of the old T. pavonia; but the later flowers assumed their proper colour of fine yellow, spotted with crimson. An apparently authentic account has been published (11/86. 'Transact. Lin. Soc.' volume 2 page 354.) of two forms of Hemerocallis, which have been universally considered as distinct species, changing into each other; for the roots of the large-flowered tawny H. fulva, being divided and planted in a different soil and place, produced the small-flowered H. flava, as well as some intermediate forms. It is doubtful whether such cases as these latter, as well as the "flushing" of broken tulips and the "running" of particoloured carnations,—that is, their more or less complete return to a uniform tint,—ought to be classed under bud-variation, or ought to be retained for the chapter in which I treat of the direct action of the conditions of life on organic beings. These cases, however, have this much in bud-variation, that the change is effected through buds and not through seminal reproduction. But, on the other hand, there is this difference—that in ordinary cases of bud- variation, one bud alone changes, whilst in the foregoing cases all the buds on the same plant were modified together. With the potato, we have seen an intermediate case, for all the eyes in one tuber simultaneously changed their character.

I will conclude with a few allied cases, which may be ranked either under bud-variation, or under the direct action of the conditions of life. When the common Hepatica is transplanted from its native woods, the flowers change colour, even during the first year. (11/87. Godron 'De l'Espece' tome 2 page 84.) It is notorious that the improved varieties of the Heartsease (Viola tricolor), when transplanted, often produce flowers widely different in size, form, and colour: for instance, I transplanted a large uniformly-coloured dark purple variety, whilst in full flower, and it then produced much smaller, more elongated flowers, with the lower petals yellow; these were succeeded by flowers marked with large purple spots, and ultimately, towards the end of the same summer, by the original large dark purple flowers. The slight changes which some fruit-trees undergo from being grafted and regrafted on various stocks (11/88. M. Carriere has lately described in the 'Revue Horticole' December 1, 1866 page 457, an extraordinary case. He twice inserted grafts of the Aria vestita on thorn- trees (epines) growing in pots; and the grafts, as they grew, produced shoots with bark, buds, leaves, petioles, petals, and flower-stalks, all widely different from those of the Aria. The grafted shoots were also much hardier, and flowered earlier, than those on the ungrafted Aria.) were considered by Andrew Knight (11/89. 'Transact. Hort. Soc.' volume 2 page 160.) as closely allied to "sporting branches," or bud-variations. Again, we have the case of young fruit-trees changing their character as they grow old; seedling pears, for instance, lose with age their spines and improve in the flavour of their fruit. Weeping birch-trees, when grafted on the common variety, do not acquire a perfect pendulous habit until they grow old: on the other hand, I shall hereafter give the case of some weeping ashes which slowly and gradually assumed an upright habit of growth. All such changes, dependent on age, may be compared with the changes, alluded to in the last chapter, which many trees naturally undergo; as in the case of the Deodar and Cedar of Lebanon, which are unlike in youth, whilst they closely resemble each other in old age; and as with certain oaks, and with some varieties of the lime and hawthorn. (11/90. For the cases of oaks see Alph. De Candolle in 'Bibl. Univers.' Geneva November 1862; for limes etc. Loudon's 'Gard. Mag.' volume 11 1835 page 503.)]

GRAFT-HYBRIDS.

Before giving a summary on Bud-variation I will discuss some singular and anomalous cases, which are more or less closely related to this same subject. I will begin with the famous case of Adam's laburnum or Cytisus adami, a form or hybrid intermediate between two very distinct species, namely, C. laburnum and purpureus, the common and purple laburnum; but as this tree has often been described, I will be as brief as I can.

[Throughout Europe, in different soils and under different climates, branches on this tree have repeatedly and suddenly reverted to the two parent species in their flowers and leaves. To behold mingled on the same tree tufts of dingy-red, bright yellow, and purple flowers, borne on branches having widely different leaves and manner of growth, is a surprising sight. The same raceme sometimes bears two kinds of flowers; and I have seen a single flower exactly divided into halves, one side being bright yellow and the other purple; so that one half of the standard-petal was yellow and of larger size, and the other half purple and smaller. In another flower the whole corolla was bright yellow, but exactly half the calyx was purple. In another, one of the dingy-red wing-petals had a narrow bright yellow stripe on it; and lastly, in another flower, one of the stamens, which had become slightly foliaceous, was half yellow and half purple; so that the tendency to segregation of character or reversion affects even single parts and organs. (11/91. For analogous facts see Braun 'Rejuvenescence' in 'Ray Soc. Bot. Mem.' 1853 page 320; and 'Gardener's Chronicle' 1842 page 397; also Braun in 'Sitzungsberichte der Ges. naturforschender Freunde' June 1873 page 63.) The most remarkable fact about this tree is that in its intermediate state, even when growing near both parent-species, it is quite sterile; but when the flowers become pure yellow or pure purple they yield seed. I believe that the pods from the yellow flowers yield a full complement of seed; they certainly yield a larger number. Two seedlings raised by Mr. Herbert from such seed (11/92. 'Journal of Hort. Soc.' volume 2 1847 page 100.) exhibited a purple tinge on the stalks of their flowers; but several seedlings raised by myself resembled in every character the common laburnum, with the exception that some of them had remarkably long racemes: these seedlings were perfectly fertile. That such purity of character and fertility should be suddenly reacquired from so hybridised and sterile a form is an astonishing phenomenon. The branches with purple flowers appear at first sight exactly to resemble those of C. purpureus; but on careful comparison I found that they differed from the pure species in the shoots being thicker, the leaves a little broader, and the flowers slightly shorter, with the corolla and calyx less brightly purple: the basal part of the standard-petal also plainly showed a trace of the yellow stain. So that the flowers, at least in this instance, had not perfectly recovered their true character; and in accordance with this, they were not perfectly fertile, for many of the pods contained no seed, some produced one, and very few contained as many as two seeds; whilst numerous pods on a tree of the pure C. purpureus in my garden contained three, four, and five fine seeds. The pollen, moreover, was very imperfect, a multitude of grains being small and shrivelled; and this is a singular fact; for, as we shall immediately see, the pollen-grains in the dingy-red and sterile flowers on the parent-tree, were, in external appearance, in a much better state, and included very few shrivelled grains. Although the pollen of the reverted purple flowers was in so poor a condition, the ovules were well formed, and the seeds, when mature, germinated freely with me. Mr. Herbert raised plants from seeds of the reverted purple flowers, and they differed a VERY LITTLE from the usual state of C. purpureus. Some which I raised in the same manner did not differ at all, either in the character of their flowers or of the whole bush, from the pure C. purpureus. Prof. Caspary has examined the ovules of the dingy-red and sterile flowers in several plants of C. adami on the Continent (11/93. See 'Transact. of Hort. Congress of Amsterdam' 1865; but I owe most of the following information to Prof. Caspary's letters.) and finds them generally monstrous. In three plants examined by me in England, the ovules were likewise monstrous, the nucleus varying much in shape, and projecting irregularly beyond the proper coats. The pollen grains, on the other hand, judging from their external appearance, were remarkably good, and readily protruded their tubes. By repeatedly counting, under the microscope, the proportional number of bad grains, Prof. Caspary ascertained that only 2.5 per cent were bad, which is a less proportion than in the pollen of three pure species of Cytisus in their cultivated state, viz., C. purpureus, laburnum, and alpinus. Although the pollen of C. adami is thus in appearance good, it does not follow, according to M. Naudin's observation (11/94. 'Nouvelles Archives du Museum' tome 1 page 143.) on Mirabilis, that it would be functionally effective. The fact of the ovules of C. adami being monstrous, and the pollen apparently sound, is all the more remarkable, because it is opposed to what usually occurs not only with most hybrids (11/95. See on this head Naudin ibid page 141.), but with two hybrids in the same genus, namely in C. purpureo-elongatus, and C. alpino-laburnum. In both these hybrids, the ovules, as observed by Prof. Caspary and myself, were well-formed, whilst many of the pollen-grains were ill-formed; in the latter hybrid 20.3 per cent, and in the former no less than 84.8 per cent of the grains were ascertained by Prof. Caspary to be bad. This unusual condition of the male and female reproductive elements in C. adami has been used by Prof. Caspary as an argument against this plant being considered as an ordinary hybrid produced from seed; but we should remember that with hybrids the ovules have not been examined nearly so frequently as the pollen, and they may be much oftener imperfect than is generally supposed. Dr. E. Bornet, of Antibes, informs me (through Mr. J. Traherne Moggridge) that with hybrid Cisti the ovarium is frequently deformed, the ovules being in some cases quite absent, and in other cases incapable of fertilisation.

Several theories have been propounded to account for the origin of C. adami, and for the transformations which it undergoes. The whole case has been attributed by some authors to bud-variation; but considering the wide difference between C. laburnum and purpureus, both of which are natural species, and considering the sterility of the intermediate form, this view may be summarily rejected. We shall presently see that, with hybrid plants, two embryos differing in their characters may be developed within the same seed and cohere; and it has been supposed that C. adami thus originated. Many botanists maintain that C. adami is a hybrid produced in the common way by seed, and that it has reverted by buds to its two parent-forms. Negative results are not of much value; but Reisseck, Caspary, and myself, tried in vain to cross C. laburnum and purpureus; when I fertilised the former with pollen of the latter, I had the nearest approach to success, for pods were formed, but in sixteen days after the withering of the flowers, they fell off. Nevertheless, the belief that C. adami is a spontaneously produced hybrid between these two species is supported by the fact that such hybrids have arisen in this genus. In a bed of seedlings from C. elongatus, which grew near to C. purpureus, and was probably fertilised by it through the agency of insects (for these, as I know by experiment, play an important part in the fertilisation of the laburnum), the sterile hybrid C. purpureo-elongatus appeared. (11/96. Braun in 'Bot. Mem. Ray. Soc.' 1853 page 23.) Thus, also, Waterer's laburnum, the C. alpino-laburnum (11/97. This hybrid has never been described. It is exactly intermediate in foliage, time of flowering, dark striae at the base of the standard petal, hairiness of the ovarium, and in almost every other character, between C. laburnum and alpinus; but it approaches the former species more nearly in colour, and exceeds it in the length of the racemes. We have before seen that 20.3 per cent of its pollen-grains are ill-formed and worthless. My plant, though growing not above thirty or forty yards from both parent-species, during some seasons yielded no good seeds; but in 1866 it was unusually fertile, and its long racemes produced from one to occasionally even four pods. Many of the pods contained no good seeds, but generally they contained a single apparently good seed, sometimes two, and in one case three seeds. Some of these seeds germinated, and I raised two trees from them; one resembles the present form; the other has a remarkable dwarf character with small leaves, but has not yet flowered.) spontaneously appeared, as I am informed by Mr. Waterer, in a bed of seedlings.