Jour. Roy. Hort. Soc. 6: 1-13 (1851)
I. — Gärtner's Observations upon Muling among Plants.
By the Rev. M. J. Berkeley.
(Second Notice see Vol. V., p. 172.)

IN prosecuting experiments on muling, it becomes important to inquire what is the effect of alternating the sexes of the plants under examination, making that species the subject of hybridization which before supplied the male element. The term crossing, which is used by English authors in a very wide sense, is confined to this especial phenomenon by Kölreuter and his followers. Such experiments are by no means universally successful; but the point of interest is that, where impregnation is perfected, the resulting hybrid is in either case precisely the same. Amongst a number of individuals a few slight variations as to size and colour of the flowers or foliage may exist, but only such as occur wherever a multitude of plants are raised immediately from pure species.

This result of cross-impregnation is the more remarkable, since it is directly contrary to what takes place in the animal kingdom, as the well-known crosses between the horse and ass obviously prove. And not only so, but where many individuals are the result of a single act of impregnation, as in the case of the dog and wolf, there is by no means a constant type prevalent through the whole litter. This difference is explicable by the greater unity of type in plants, as regards sexuality, whether absolutely coincident, as in the case of hermaphrodites, or scarcely modified where the sexes are confined to distinct individuals.

There are indeed a few cases amongst plants where the sexes are separate, and the typical form different, as in Lychnis diurna and vespertina; but the mules derived from counter-impregnation present no such striking differences as occur amongst animals, and slight variations of type under like circumstances have been observed in the genus Digitalis. D. lanato-ochroleuca, for instance, has short large blossoms, with the upper lip smaller and its margin waved: D. ochroleuco-lanata has a proportionably longer, thinner, and more cylindrical blossom, with a smaller distinctly three-toothed upper lip, and a sharp point in the centre of the lower lip. Other examples might be brought forward in the same genus.

It is to be observed, however, that many futile attempts at crossing in the genus Digitalis may be expected, though success will most probably repay repeated experiments. It has been lately ascertained that moistening the stigma with the honey secreted by the flowers promotes the fecundation of the ovarium of these plants, and, doubtless, since no impregnation can be effected before the separation of the lobes of the stigma, because the pollen adheres to it in consequence of such treatment until the division takes place.

Though, however, the influence of the sex of the parent is not in general exhibited, as in animals, by any outward signs, yet there is often a great difference in the inward constitution of the plants produced by cross impregnation. It is, for instance, by no means a matter of indifference whether Nicotiana rustica be impregnated with the pollen of N. paniculata, or the contrary. In the former case the mule is more fruitful than in the latter, and other similar instances might be adduced.

In intermediate impregnation, however, it appears to be of no consequence whether the male or female be the intermediate species. Nicotiana chinensis, for instance, produces the same mule with N. rustico-paniculata that it does with N. paniculato-rustica.

The author very justly considers this identity of type in the mules arising from counter-impregnation as a very strong argument against Schleiden's notion of the production of the embryo from the extremity of the pollen tube. Two different kinds of pollen could scarcely be expected, if such were really the case, to produce two perfectly identical plants. And the same inference may be drawn from the fact that in the production of double blossoms, it is a matter of indifference whether the pollen be taken from a double or single variety, provided the flower of the matrix is double.

It is not necessary to go through the facts adduced in the chapter relative to the normal types produced by impregnation between two pure species, as those of most importance have been already mentioned incidentally. Slight changes occur from time to time; but this is nothing more than takes place in raising seedlings from the pure species themselves without any cross whatever. The grand point in all experiments is to secure the parent plants in as pure a state as possible; and where practicable, to select species from the woods and fields, and even these as near the normal form as may be. Where proper pains are taken, exceptional forms are very rare, perhaps one only from the produce of many capsules. Some genera and species seem to have a greater disposition to produce such forms than others, as for instance Dianthus, Digitalis, Lobelia, Passiflora, Nicotiana, and Verbascum, and sometimes this tendency is confined to a single species of a genus. When, however, numbers of such exceptional plants spring from the same seed, as from Verbascum phoeniceo-austriacum, Dianthus arenario-pulchellus, &c., or when, on repeated experiments with the same species, the exceptional type occurs either singly or in numbers, the individuals have as complete a resemblance amongst each other as the normal mules; no variations or transitional forms occur, as is the case where the parents are hybrids: they are for the most part decided types. It is observed too, that neither the male nor female element has an exclusive influence in the formation of such exceptional types, though they exhibit usually a decided predominance either of the character of the male or female parent, the unusual potency of the one or the other having given rise to the abnormal form. Most of them are absolutely sterile, a point which makes the study of them extremely difficult and vexatious; and even where they are fertile, as in Dianthus arenario-pulchellus, there is a tendency, like that so well known amongst animals, and to which the term Atavism has been given, to revert to the ancestral type.

Occasionally, however, especially in plants which have been long cultivated, or amongst species which at the same time have a very intimate elective affinity, and are so closely allied to each other that they seem merely marked varieties, such as Matthiola annua and glabra, Malva sylrestris and Mauritiana, Lychnis diurna and vespertina, Primula elatior and officinalis, differences of form and colour occur amongst the typical hybrids which have no constancy, and winch cannot be obtained with certainty on repeated experiments. These differences are generally such as are manifest only in the living plant, as in the degree of roughness, undulation, viscidity, &c. It is supposed by Mr. Herbert, that variations of this kind arise from the application of a quantity of pollen insufficient to produce a real typical mule, though potent enough to effect slight changes. To this notion, however, the author is opposed, on the ground that experiment seems to show that the completeness or incompleteness of impregnation influences merely the perfection of the fruit and seeds, and that a quantity of pollen insufficient for the impregnation of an ovarium produces only imperfect fruit and seeds, but gives rise to no distinct type; for unless a proper quantity of pollen be applied, the seeds are never perfected, or the embryo is inert. Direct experiments also show that the time of impregnation, whether early or late, be the state of the stigma what it may, has no influence on the produce.

No such deflections from the normal type of the mules have occurred to the author in plants brought immediately from the woods or fields, but only amongst individuals which have been long under cultivation.

It appears from experiment, that where impregnation takes place between two pure species, it is an universal rule, "that the characters of the parents never remain pure and unaltered in the formation of the hybrid." In general every part of the new production is modified, so that it presents a decided difference from either of the parents, though resembling the one more than the other. In no case, however, are anomalous forms generated bearing no resemblance to either. At the same time they are not produced according to mathematical formulae and ratios; their differences are mingled in unequal proportions, and may be arranged under three heads, viz. intermediate, mixed, and decided hybrids, of which mention will be made presently. And it is to be observed, that the change does not always take place in the whole of the plant, as though the principle on which it depends penetrated into the intimate structure of the whole plant; but even in the most decided hybrid types, as Nicotiana paniculoto-vincaeflora, and vincaefloro-Langsdorfii, as well as in the mixed forms, particular organs undergo only a very slight modification, while others are materially changed.

It was remarked, that no instance has occurred in which the characters of the parent remain perfectly unaltered in the hybrid. An apparent exception, however, to this rule requires notice. In 1846 Lychnis flos-cuculi was treated with the pollen of Cucubalus Behen. Amongst many normal plants of the former, a single example occurred, which differed as to habit and blossoms not the least from the female parent, but the leaves in form and in their glaucous appearance, especially the root-leaves, agreed completely with those of the male. Repeated attempts to reproduce this form were unsuccessful. The author, however, does not regard the partial alteration of form and substance of the leaves as the effect of the strange pollen; but believes that all the plants produced arose from an impregnation of the Lychnis with its own pollen after the attempted hybridization with that of the Cucubalus, especially since all of them, whether normal or exceptional, were as fruitful as the mother-plant.

In Nicotiana suaveolenti-Langsdorfii and vincaefloro-Langsdorfii the change is very slight, consisting merely in a trifling alteration of colour in the blossoms, in the violet or blue colour of the anthers, and in a partial separation of the stamens from the tube of the corolla: in other respects the resemblance to the female parent is very close, insomuch that doubts as to their being really hybrids might be entertained, were they not completely barren. Unfortunately it is impossible to reverse the experiment in these cases, as N. Langsdorfii is not fertile except when impregnated with its own pollen.

Various notions have existed, both in the animal and vegetable kingdom, with respect to the degree of influence which the sexes have in the production of hybrids: according to one authority, the male, in animals, giving origin to internal qualities, the female to external; to another, the former to the cellular system, the latter to the nervous, &c. Amongst plants, the difference of opinion is as great hut the truth appears to be, that no general rule can be laid down in Digitalis the influence of the female parent being predominant, in Nicotiana that of the male, and the differences exhibited by individual species are no less decisive against any universal law. And this is no less true as to comparative degrees of fruitfulness. Indeed, the identity of the produce, when the sexes are reversed, is a sufficient proof of its non-existence.

In the occurrence of exceptional types it is probable that the difference arises from some peculiar constitution of the individual ovule, rather than from any different condition of the pollen.

It is curious that the specific difference of nearly allied species appears more distinct in the hybrids to which they give rise than in the pure species. For instance, Lobelia cardinalis, fulgens, and splendens, than which no species can well be more intimately allied, give totally distinct hybrids when united with L. syphilitica; and many other instances might be brought forward while on the contrary Nicotiana magnifolia, macrophylla, marylandica, and petiolata give identical hybrids when impregnated with N. glutinosa. The obvious conclusion is, that the supposed species are in reality mere forms, and that hybrid types depend entirely on the specific distinction of species, and not on any external influences. If such were the case, exceptional types would be of more common occurrence.

One of the most singular effects of hybridization is that which is sometimes produced on the cotyledons of the mule in the first generation. It often happens that in a given genus there is a strong general resemblance in the form of the cotyledons. But this is not always the case. In Dianthus they vary considerably, and there is a corresponding variation in the hybrids, as is also the case with Nicotiana quadrivalvis, in which the cotyledons differ greatly from those of other species.

The general type of a hybrid is preserved throughout the whole life of an individual. Mr. Herbert found this to be the case in hybrids of Camellia. The blossoms, however, do not always remain so constant, especially in hybrid varieties, as the florist knows to his cost.

Different degrees of resemblance to the parent types are exhibited by different hybrids, insomuch that though, as Mr. Herbert says, a well-skilled florist may guess the result of any particular experiment with tolerable accuracy, yet as the forms are not fashioned according to strictly mathematical laws, but after some vital energy which we can only estimate by its effects, we can form no absolutely certain anticipation

It was said above that hybrids may be divided into intermediate, mixed, and decided types. These divisions, however, are far from being strictly definite. Each of these heads may be briefly considered.

* Lychnicucubalus albus (Lychnis vespertina x Cucubalus viscosus)
  Lychnicucubalus ruber (Cucubalus viscosus x Lychnis diurna)

First, then, hybrids occur in which the characters of the parents are so intimately blended, that it is impossible to say to which there is a greater resemblance. Something in such cases must depend upon individual judgment, and the degree of aptitude it tracing accurately differences and points of resemblance. Sometimes the result as to the number of organs, where differences exist in the parents in this respect, is curiously intermediate as, for instance, from the three stigmas of Cucubalus and five of Lychnis, arise the four of Lychnicucubalus.*

The second class is that of mixed hybrids. In these, one part or other of the hybrid approaches the paternal or maternal form, though the characters of the parents never pa altogether pure into the new organism. Melons are a familiar instance; or, to take one more special, in Lychnis vespertino-diurna there is no perfect diurnal sleep, as in L. vespertina, but the petals roll back slightly when the sun shines, or the weather is lot. The hybrid resembles L. vespertina in its smaller leaves, diurna in the vital phenomena vespertina in the larger flowers and straight blunt stigmas, diurna in the pubescence vespertina in the more pyramidal fruit, as also in the size and colour of the seeds.

Thirdly, we have the decided hybrids viz. those in which the resemblance to one of the parents, whether male or female, is so decided, that the agreement is at once perceptible and beyond all doubt. It might be supposed that in this case the predominance of one of the parents might prevent alternate crossing, it being scarcely probable that when the parents were reversed, the result under such circumstances could remain the same. In some instances, indeed, of decided hybrids, alternate crossing does not succeed; but this is far from constant, and in some most decided types the parents may be successfully reversed. Lobelia cardinali-syphilitica is a case in point. There are some species, producing decided hybrids, which, when united with several others successively, always predominate in the resultant types. These species have usually some strong peculiarity about them, but it cannot he asserted priori what species are likely to exercise such an influence, and there are other species, again, which prevail even over these. For instance, Dianthus barbatus, which communicates its type, as regards leaves and general habit, to D. Armeria, prolifer, and Carthusianorum, is in its turn overcome by D. coryophyllus and superbus. These species have been called, but not very happily, generic types. These predominating types show clearly that the ratios of the potency through which the union of two pure species takes place must be unequal, and that there can be no question about equivalence of factors. Even in the intermediate hybrids, where the formative powers are so intimately blended, there is still generally some particular organ which shows the prevalence of the one factor over the other.

We cannot trace the origin and development of the different vegetable forms, from the simple cell to the complete development of the perfect vegetable, through all its phases; much less can we distinguish the connection between the changes wrought by hybridization, and normal vegetable metamorphoses. It is very doubtful, indeed, whether vegetable anatomy, improved as it may be, will ever be in a condition to estimate such vital processes. We must rest content, therefore, with the knowledge of the mere facts of the case. New characters sometimes arise in hybrids entirely distinct from those existent in the parents, so that they might reasonably be taken for distinct species. Hybrid forms of Mirabilis exhibit such changes in a very surprising degree, as does also the genus Rhododendron, and curious instances may be adduced of marked alterations of form and condition in almost every organ.

In few characters is the influence of muling more striking than in the size and colour of blossoms. In many closely allied species, which differ but little in habit or foliage, the colour of the corolla is of great importance. In a wild state it is for the most part constant, and is often indicative of distinct groups or species. In other groups, on the contrary, it is extremely variable, and is notably different at different periods of growth. Where, however, colour is the most constant and distinctive, union is often practicable, and in general the consequence of hybridization is a complete derangement of the laws on which such constancy of hue depends. Neither are the hues resulting from the union necessarily intermediate. Blue and yellow, for instance, do not produce green, as is proved by Verbascum phoeniceum and phlomoides. Gladiolus cardinali-blandus exhibits the less brilliant hue of the male parent rather than the splendour of the mother; and in some cases the tone of colour of one of the parents is exhibited under a more brilliant tint, as in Nicotiana suaveolenti-glutinosa.

Sometimes the change of colour is exhibited in an increase of the number of typical forms arising from any particular union. Geum canadensi-coccineum, for instance, gives a larger number of hybrids with ample orange-coloured flowers, mixed with a small proportion of pale yellow. Mr. Herbert raised from Rhododendron ponticum and Azalea pontica two specimens with yellow scented flowers like those of the Azalea, one with lemon coloured, and one with a chestnut-brown tint intermediate between the purple and yellow of the parents. In reversed impregnation, though the forms are identical, whichever parent be male or female, and the colours generally the same, a change in this respect does occasionally take place. Mr. Herbert's notion, therefore, that the male parent gives the tone to the colouring of the mule is certainly untenable.

In mixed mules, where the pure species is also the female parent of the mule with which it is impregnated, the change of colour is so variable, that scarcely two plants from the same seed exhibit the same colouring. This, consequently, is the most fruitful source of the florist's varieties. In compound bastards, where there are three parents, the mules generally assume the tone of the new male parent.

Not unfrequently flowers of different hues occur in the same plant, as in Mirabilis and Dianthus barbatus. It has been supposed that this arose from the influence of strange pollen ou the blossoms and ovaries but as the blossom is expanded before the access of the pollen, this cannot be the ease. The cause of this variety of colour is at present altogether obscure.

The colours of the capsules and seeds are often altered by hybridization. We have already said that Pisum appears to be the only genus in which the tint of the seeds is immediately affected by impregnation, alterations in general not being apparent till the succeeding generation. The common Maize exhibits, after hybridization with forms possessing differently coloured seeds, not merely differently coloured spikes, but spikes hearing variously coloured seeds. No immediate alteration, however, was effected as in Pisum, and the same may be said of Lychnis diurna and vespertina, of which the former has reddish brown, the latter cinereous seeds. These facts confirm the general law, that the influence of the strange pollen in hybridization makes no alteration in the peculiar form and external peculiarities of the fruit and seeds of this mother plant, but merely produces in the embryo, after germination and in the course of its development, a capability of producing a mixed product from the concurrence of the two factors.

As in the case of flowers, party-coloured or differently coloured fruits exist sometimes on the same stem. Such phenomena do not appear, however, to be the effect of hybridization, but to be ascribable to the tendency of plants to produce varieties. With regard to the fructifying organs, the male are affected more than the female. The number is often increased, but their fertility impaired or wholly destroyed. The stamens, though externally perfect, are often diseased and disposed to fade prematurely; and this not in a few blossoms only, but in all equally, while the anthers are well formed, but for the most part sterile. Sometimes, indeed, they are smaller than in pure species, shrivelled nod discoloured, and contain no perfect granules, but merely an inorganized mass; or the pollen is scanty and white, and no dehiscence takes place. The fertility of the pollen cannot, indeed, always be determined either by external appearance or from direct experiment, as impregnation does not always take place even with pure pollen; and some plants are more easily fecundated with pollen taken from another individual of the same species than with their own. In almost all fertile bastards, the normal pollen grains are mixed with many that are smaller and imperfectly organized; and, as a general rule, the colour is less vivid than in the parent species. Sufficient attention does not appear at present to have been paid to the protrusion of the pollen tubes in hybrids as compared with that in pure species, nor to the contents of the pollen grains.

The number of styles is also frequently increased in the blossoms of hybrids, especially those which open first. It is, however, far more difficult to form any judgment as to their fertility than with respect to the stamens. This may, indeed, be sometimes anticipated from their preternatural elongation, the far rougher surface of the stigma, or the increased time during which the stigma remains moist, or in other cases by its speedy discoloration. The ovules too, though often perfect externally while they are really barren, are frequently shrivelled and abortive; and even in fertile hybrids the number of ovules capable of impregnation appears to be small, whatever quantity of pollen be applied to the stigma.

The greatest change in the fructifying organs takes place in the union of dioecious and hermaphrodite plants, affecting, however, as in animal hybrids, the male organs first, and to a greater degree.

We are obliged to pass without notice several chapters relative to the fruitfulness of hybrids under various points of view, though far from uninteresting. We proceed to consider very briefly a few phenomena exhibited by the impregnation of hybrids.

Those presented by the corolla and female organs are just those which take place on the impregnation of a plant with strange pollen. If the mule is tolerably fruitful, the corolla falls off at the usual time; but if it is only very sparingly fertile or entirely sterile, it remains longer, or the whole blossom falls. The stigma continues moist long after the anthers have lost their pollen, and the whole course, from the perfecting of the stigma to its fading, is longer than in natural fructification.

Even where there are a few perfect pollen grains in the anthers, impregnation does not always take place in mules with their own pollen, probably in consequence of the good grains not being sufficiently numerous to ensure success. Many fertile mules, therefore, require artificial impregnation, and that frequently repeated, to produce fruit and in general typical mules appear to become less fruitful as they recede from their original stock a fact exactly contrary to what is exhibited by varieties. The more fruitful hybrids retain their typical form after many generations, such as Aquilegia atropurpureo-canadensis, &c.; the greater part, however, yield forms deviating more or less from the normal type, the variations being chiefly confined to the flower. Hippeastrum Johnsoni (regio-vittatum) was observed by Mr. Herbert, when fertilized with its own pollen, to have less beautiful and smaller blossoms. Differences, however, of general habit, or of other especial organs, occasionally take place. In cases where impregnation is not effected with their own pollen, mules are often capable of fertilization with the parent pollen, for which in general they exhibit a greater elective affinity than for their own, and the same may sometimes be said of certain pure species. Of the two kinds of pollen, that from the male and that from the female parent, the greater elective affinity appears to reside with that parent to which the hybrid bears the closest resemblance. The impregnation of fertile mules with their own pollen is a very fruitful source of florists' varieties, as well as that of the female parent with the pollen of the mule.

When the hybrid is impregnated by the original male parent, the result is much the same as in simple hybrids self-fertilized, both in respect of the types produced and the degree of fertility. Various forms are raised from one capsule, and the different individuals do not present the same degree of susceptibility for impregnation. Different capsules, too, offer very different results. When these mules are in turn self-impregnated, either naturally or artificially, they are commonly more fruitful than they were after the first impregnation. As might be expected, the seedlings approach nearer to the paternal type: when the original simple mules in their second generation and the paternal hybrids of the second degree exhibit a return to the type of the maternal ancestor, such a return is never perfect, but only partial.

The tendency of varieties to return to the maternal type seems to be a peculiarity general to the vegetable kingdom, especially if left to themselves, free from the trammels of cultivation. This return, however, in the second generation of simple hybrids, or of paternal mules of the second degree, is always effected by fructification, and not by any other mode of propagation. It seems also more easy than the approach to the paternal type, thought in neither case does it take place to a considerable extent, nor does it take place in all genera, and when it does occur the produce is less fertile. Mr. Herbert believed that such deviations from the normal type might arise, when the proper pollen was insufficient for impregnation, from the access of the pollen of some nearly related species; but this contradicts the laws which have been established as to elective affinity.

In very fruitful hybrids no such deviations or different types have been observed; it should seem therefore that the integrity and force of the organs of fructification prevent the occurrence of such deviations. It appears, too, that the paternal element is of greater power to produce variations of form than the maternal.

When hybrids are impregnated a third or fourth time with the pollen of the original male parent, they gradually approximate more and more to the male type, and at last are not distinguishable from it, except perhaps in a less degree of fertility, though this negative sign vanishes sooner or later. There is no certainty as to the number of successive impregnations necessary to produce this complete change. Different species exhibit in this respect very different results. Nicotiana rustico-paniculata, even in the fifth degree, is occasionally completely sterile either as to the stigma or anthers, but especially as regards the latter.

Mules, however, may be also impregnated with the pollen of the mother plant. The maternal type is of course prevalent in such mules, which may account for the greater fertility of maternal than patentai hybrids. A greater number also of different types is formed than in the second generation of simple hybrids, and the second degree of paternal bastards, where they are at most two or three, insomuch that if the colour of the blossoms be taken into the account, they amount in Dianthus chinensi-barbatus chinensis to fifteen. All bear a greater or less resemblance to the maternal type. In the next generation, whether fertilized with their own or with the pollen of the original mother, most of the plants have completely reverted to the mother type.

Thus by the means of hybridization one species is changed gradually into another, though the hybrids themselves through the whole period of their existence preserve their proper type. The simple hybrid reverts to the mother type by repeated impregnation with the maternal pollen, or when the paternal pollen is applied, goes forward to the type of the father the conversion of the mother into the father is, however, seldom synchronous with the contrary change. Nicotiana rustica was changed in this manner by Kölreuter into N. paniculata, and similar changes have been effected by others. The experiments require much care and time, and great caution to avoid error; but there is not the least reason to doubt the truth of the fact, as has been done by some impugners of the sexual theory. It is obvious that to ensure success, species must be chosen winch are pre-eminently disposed to hybridize, and whose mules are fertile, otherwise the experiment will be stopped in some of its stages by the sterility of the hybrids. The predominance of fructiferous power must also be on the female side, as the pollen of the pure species is to be used. The subjoined table shows the number of impregnations requisite to complete the changes: the results of experiment, however, are not always the same.

Aquilegia atropurpurea became A. canadensis in 3 generations
——— canadensis " atropurpurea 4 "
——— ——— " vulgaris 4 "
Dianthus arenarius " caryophyllus 5-6 "
——— ——— " pulchellus 5-6 "
——— ——— " chinensis 5-6 "
——— ——— " superbus 5 "
——— Armeria " deltoides 5-6 "
——— barbatus " Carthusianorum 3-4 "
——— ——— " chinensis 5 "
Geum urbanum " rivale 4 "
Lavatera pseudolbia " thuringiaca 4 "
Lychnis diurna " vesperina 4 "
——— vespertina " diurna 3 "
OEnothera nocturna " villosa 4 "
——— villosa " nocturna 4-5 "
  &c.   &c.    

This change, be it observed, is totally different from those supposed effects of external circumstances in converting one species into another of a very different structure, as oats into rye, rye into Bromus secalinus, Bromus sterilis into Hordeum murinum, Brassica rapa into Thlaspi arvense, or this latter into Camelina sativa and Capsella Bursa pastoris. That all these supposed changes are mere illusions we do not doubt for an instant, and the curious case figured in the 'Gardeners' Chronicle' for 1849 is an instance of one mode in which delusion may have occurred. The effects of grafting are far more interesting, but in reality are little connected with our subject, and we the rather pass them by, as so much has been said oil the subject in Mr. Herbert's valuable memoir in this Journal.

We regret that we have no room fur any notice of the author's remarks oil the Classification of Bastards, though something has already been said oil the subject; and we must also pass over his observations oil varieties and their mules, as also oil the occurrence of wild hybrids. There is, too, a bug chapter of sixty pages on the distinctive marks and peculiarities of hybrids, but it contains little that has not been already noticed.

We conclude with a short abstract of his last chapter, which treats of the practical uses which landowners and floriculturists may derive from the production of mules.

The peculiar tendency of hybrids to luxuriance in their stem and foliage, and the facility with which they are propagated by cuttings, layers, &c., is obviously of great consequence to agriculturists.

Little has been done at present in the hybridizing of cereals, but Mr. Herbert believes that more useful varieties than at present exist of wheat, oats, and barley might be produced by combining the fruitfulness of one variety with the hardiness of another, to both of which might be added the thin skin and consequent superior weight of a third. Knight's wrinkled peas are a proof of what may be done by hybridizing, and it is probable that much might be effected in beet, cabbages, carrots, celery, &c., by especial attention to this point.

Amongst woody plants also there are instances of peculiarly luxuriant growth, such as Lycium barbato-afrum. Varieties therefore might be produced, of much more rapid growth, which for some purposes might have their value, though the quality of the timberr would probably suffer.

Another peculiarity of hybrids is their precocity of which advantage may be taken where early fruit is desirable, or where the summers are not long enough to ripen the later fruit.

A very importaut quality of hybrids is also their posvcr in very many cases of enduring a greater degree of cold than the pure species from which they are derived, aud hence the acclimatisation of many useful plants by means of hybrid forms or varieties may be effected. The hybrids, fur insttuiee, of NIcotiana are far less susceptible of frost than their pure parents, a circumstance of very great importance if the cultivation of tobacco were to be materially extended.

The great fruitfulness of many hybrid varieties is also a material point as regards their useful qualities, especially in orchards a,ol vitievards, and where ornament, effect, or what the Germans call aesthetic botany in its various branches is concerned, hybrids supply ao endless subject of experiment.

And lastly, the longer duration of many hybrids and their more persistent larger blossoms make them especial objects of favour and delight.

The great difficulty in the way of experiment is the frequent want of fertility in the seeds of hybrids, and their tendency to wear ont, wherever there is a possibility of impregnation from neigidiourimig varieties.

An appendix is subjoined to the work, containing an account of tof the manipulations of which the author made use in his experiments, and a list extending to above forty pages of all the species which have been submitted to experiment.