Trans. Bot. Soc. Edinburgh 8:113-131 (1866)

I. New Researches on Hybridity in Plants.
By M. CH. NAUDIN.
Translated from the Annales des Sciences Naturelles, by GEORGE MAY LOWE, Esq.

(1.) On the Sterility and Fecundity of Hybrids.

A century ago, Koelreuter demonstrated by proofs which no other observer has ever surpassed in exactitude, and which still retain all their value, the fact of the sterility of hybrids being absolute in some cases, but only partial in others. These two facts, since so frequently confirmed, cannot now be disputed. In a former paper I gave some examples which serve to illustrate them.

We have seen Nicotiana-californico-rustica, N. glutinoso-macrophylla, N. glutinoso-angustifolio-macrophylla, Digitalis luteo-purpurea and Ribes Gordonianum, sterile both by the stamens and ovary-the former being totally destitute of pollen well formed, and the latter incapable of impregnation by the pollen of the parent plants. But as the pistil does not in every case present any appreciable deformity, it is natural to seek in the ovule itself the true cause of this inaptitude to receive impregnation.

It has been fully proved by many cases of hybridity, in which, in the same ovary one portion of the ovules resists impregnation, whilst the other becomes converted into embryonic seeds capable of germinating—that this defectiveness exists in the ovule, and not in the more exterior parts of the pistil.

We have seen this in the three hybrid generations of Luffa acutangulo-cylindrica, also in Luffa amaro-cylindrica, Cucumis Meloni-trigonus, Nicotiana rustico-paniculata, and paniculato-rustica, &c. Cucumis myriocarpo-Figarei is a not less convincing proof, since among 100 fruits which were developed and ripened under the influence of pollen derived from the maternal species, 19 at least were destitute of seeds, and each fruit, among the small number which contained any, only yielded one seed. I might mention, in support of this fact, the example of Mirabilis longifloro-Jalapa, though in this case the ovary is uniovular. The stigmas of this hybrid were all equally developed, and in this respect not inferior to those of the parent. species; yet eleven attempts to impregnate it with the pollen of Mirabilis longiflora were made without effect, and even ten were necessary with that of M. Jalapa to determine the increase of a single ovule. In the Luffa hybrids just mentioned, and also in the case of Cucumis Meloni-trigonus, however poor the pollen might have been which was employed to fertilise their ovaries, it is beyond doubt that the number of good grains deposited on their stigmas far exceeded that of the ovules which were developed into seeds.

This, it is true, is only hypothetical, but it is extremely probable. It remains to be confirmed by the anatomical examination of the ovule, and it would be very interesting to discover in what part the defectiveness exists; but this is a peculiar kind of research, very difficult, very minute, often uncertain in its results, and which one cannot enter upon without being well accustomed to it, and provided with excellent instruments, two things in which I am deficient.

I therefore contented myself with verifying experimentally the fecundity or the sterility of the ovaries, which was more expeditious, and probably more conclusive; but it is not less a subject to be recommended to professed micrographers.

That the sterilising action of hybridisation exerts much more force on the pollen than on the ovules is a most indubitable fact, and one well known to all hybridologists. This need not surprise us, since the pollen is, of all parts of the plant, the most elaborated, the most animalised, if such an expression can be used. Frequent chemical analyses prove that it is in these granules that the phosphorised and azotised materials are more accumulated than elsewhere, and thus it may be conjectured that it is this high organisation which is injured in hybrids, where the whole vegetation suffers from the disturbance which results from the intermixture of two specific essences created to live separately. The hybrids of which I have given an account present several examples. We have seen that Mirabilis longifloro-Jalapa yields pollen unfit for fertilisation, whether it be applied upon the stigmas of the hybrid, or upon those of its two parents, whilst in twenty-one attempts to impregnate it with the pollen of these last (M. longiflora and M. Jalapa), there was only one which took effect, and enlarged the ovary. This result is quite in accordance with those which M. Lecoq ("Revue Horticole," 1853, pp. 185 et 207) announced that he obtained from the same hybrid, the pollen of which he always found useless, but he was able to fertilise it by that of M. Jalapa. The difference in the strength of the pollen and the ovules becomes still more manifest in Nicotiana glauco-angustifolia (and it would undoubtedly have been the case with N. glauco-macrophylla if the experiment had been made on it), where the whole pollen mass is defective and inert, whilst the ovary becomes filled with seeds, when it is fertilised with the pollen of N. Tabacum and N. macrophylla.

All the hybrids I have observed, containing well-developed grains of pollen in their anthers, have been fertile, often to a high degrees by their ovaries. I have never seen, and I do not believe it possible to mention a single instance in which, the ovary being sterile, the stamens have been fertile, even in the least degree.

The deleterious influence which hybridisation exercises upon the fertilising apparatus shows itself in different forms.

The most common, or at least the most remarkable case, is the direct atrophy of the pollen in the anthers, more rarely the atrophy of the anthers themselves; but we have also seen it act on the entire flowers. It is so among all the hybrids produced by the agency of Datura Stramonium, the flowers in the lowest branches invariably fall without opening; also among all the individuals of Luffa acutangulo-cylindrica of the first generation,—all the primary male flowers perish entirely, and also some flowers which begin to open when the plants are more than full grown, and have lost part of their vigour. The same phenomenon is observed in Mirabilis longifloro-Jalapa, which loses three-fourths of its buds, in Nicotiana rustico-paniculata and paniculato-rustica of three consecutive generations, &c. In fine, another mode of sterilisation is that effected by the changing of monoecious male flowers into female, as we have seen in Luffa hybrids of the third generation.

I have every reason to believe, although I cannot positively affirm so, that the specimen of Cucumis Figarei, so remarkably large, and peculiar by the nearly total absence of male flowers, which I experimented on in 1856, and which yielded the results I have mentioned, owed both its great size and almost female unisexuality to hybridisation.

(2.) On the Difference of the Fertility of Hybrids.

Hybrids are self-fertile in all cases in which their anthers contain well-organised pollen; but if the quantity is very small, it is well not to leave the impregnation to chance, but to aid artificially in fertilising the hybrid with its own pollen. I have done this in Luffa acutangulo-cylindrica of the first generation, which has but few male flowers and a small quantity of good pollen.

In the majority of cases microscopic inspection sufficiently shows the character of the pollen; the difference in form, size, and transparency distinguishes the good and bad; and it is easy to judge, at least approximatively, of the relative quantity. Yet there are some cases, though not very common, where this examination of the pollen is not sufficient to determine whether it is active or inert; for it may happen that it has all the appearance of good pollen without having its qualities. Such was that of Mirabilis longifloro-Jalapa, whose grains, although unequal, were not deformed, and appeared full of fovilla, notwithstanding their inefficacy upon the stigmas of the two parent plants, as well as upon those of the hybrid. Perhaps the employment of chemical reagents would better determine their impotency.

There are various degrees of fertility in hybrids by means of pollen. We have seen Luffa acutangulo-cylindrica of the first generation extremely low in this respect, but in the third remarkably productive. It is the same, and nearly to the same degree, in Luffa amaro-cylindrica, Nicotiana rustico-paniculata, and paniculato-rustica, and in a great number of the toad-flax hybrids (Linaria purpureo-violacea) of the second, third, fourth, and fifth generations.

A greater richness of pollen is seen in Primula officinali-grandiflora of the first, and especially second, generation, and in Cucumis Meloni-trigonus, &c. In fine, there are some hybrids where the pollen is little inferior, if at all, in perfection to the most legitimate species. This is the case in Coccinia Schimpero-indica, Datum meteloido-Metel, D. Stramonlo-Metel, D. Stramonio-laevis, Nicotiana angustifolio-macrophylla, N.. texano-rustica, N.. persico-Langsdorffii, Petunia violaceo-nyctaginiflora, &c.; and the same in many of the toad-flax hybrids of the third and fourth generations, already very close to Linaria vulgaris.

In a word, as I said at the commencement of this article, hybrids are found of all degrees of fertility, from the extreme case where the ovary only is fertile to that where all the pollen is as perfect as that of the best-established species.

(3.) Is the Aptitude of Species to cross each other, and the Fertility of the
Hybrids which result, proportional to the apparent Affinity of the Species?

In general this is the case; but there are exceptions, and we have stated some. There are, indeed, some species, closely allied in exterior organisation and physiognomy, which are less disposed to mutual crossing than other species which are far distant in their outward appearance. Thus we have seen three species of eatable gourds, so closely resembling each other that most botanists fail to distinguish them, resist all attempts to cross them; whilst the melon and Cucumis trigonus, so very different from one another, easily give origin to very fertile hybrids, though the pollen is a little defective. Such is the case with Nicotiana glauca, which, although very distant from N. angustifolia and macrophylla, yet gives hybrids with them, having very fertile ovaries; whilst N. glutinosa, more difficult to cross with them, although belonging to the same section of the genus, only gives one sterile hybrid both by the pollen and ovary. I might also mention the crossing of D. Stramonium and D. ceratocaula, two species strangers to each other, from which there results a fertile hybrid, although attended by that peculiar kind of partial sterility which consists in the loss of the first flowers.

These exceptions, for which it is probably impossible to assign a cause, do not prevent the affinity of species, as revealed by the exterior organisation, from indicating generally the degree of aptitude to cross, and do not prevent us from forming a conjecture to a certain extent as to the fertility of the hybrids. We have seen the proof in Datura Meteloido-Metel, D. Stramonio-Tatula and Tatulo-Stramonium, D. Stramonio-laevis, Nicotiana texano-rustica and rustico-texana, N. angustifolio-macrophylla, &c. which hybrids, with the marked exception of those of D. Stramonium, have perfect fertility. The aptitude of species for mutual impregnation, and the degree of fertility of the hybrids which result, are therefore the true signs of their special affinity as regards generation; and in the great majority of cases this affinity is indicated by the exterior organisation—in other words, by the physiognomy of the species.

(4.) On the Physiognomy of Hybrids.

To give a just idea of the aspect which hybrids present, it is essential to distinguish between the first generation and those which follow.

I have always found in those hybrids which I have obtained myself, and whose origin has been well known to me, a great uniformity of aspect between individuals of the first generation, no matter how numerous, provided they proceed from the same crossing. This we have seen in Petunia violaceo-nyctaginiflora, Datura Tatulo-Stramonium, and D. Stramonio-Tatula, D. Meteloido-Metel, D. Stramonio-laevis, Nicotiana texano-rustica, and N. rustico-texana, N. persico-Langsdorffii, &c.

I do not mean to say that all the individuals of the same crossing are absolutely counterparts of one another; there are sometimes slight variations between them, but not sufficient to alter the general uniformity in a sensible degree, and it does not appear to me that these differences are any greater than those which are frequently seen between the seeds of legitimate species of the same production. In short, it may be said that hybrids which proceed from the same crossing, resemble each other, in the first generation, as much as, or nearly as much as those which proceed from the same legitimate species.

Must it be admitted, as M. Klotzsch maintains, that mutual hybrids (those which proceed from the two possible crossings between the two species) are markedly different from each other; for example, the hybrid obtained from the species A fertilised by the species B, differs sensibly from that which is obtained from the species B fertilised by the species A? I cannot deny this in an absolute manner; it would be necessary to see the hybrid which induced M. Klotzsch to make this statement; but I can assert, that all the mutual hybrids which I have obtained, as well between allied species as between distant ones, resembled one another as much as if they proceeded from the same crossing. I have already pointed this out when speaking of Datura Stramonio-Tatula and Tatulo-Stramonium, Nicotiana paniculato-rustica and rustico-paniculata, N. angustifolio-macrophylla and macrophyllo-angustifoiia, N. texano-rustica, and rustico-texana, N. persico-Langsdorffii, &c.; without doubt it may not be always so, but if the fact is true, it must be rare, and considered more as the exception than the rule.

All hybridologists are agreed that hybrids (and it always applies to hybrids of the first generation), are mixed forms, intermediate between two parent species. And this is really what does take place in the great majority of cases; but it by no means follows that these intermediate forms are always at an equal distance between the two species. On the contrary, it is often observed that they are frequently much nearer one than the other. Besides, we may conceive, that the appreciation of these relations is always a little vague, and that it is the idea which determines it. We may also remark that hybrids resemble sometimes one of the two species in one character, whilst they resemble the other in another character. This is very true, and we have seen an example in Mirabilis longifloro-Jalapa, which is distinctly more like M. longiflora in the organs of vegetation, and M. Jalapa in the flowers. But I think it is wrong to refer this distribution to the part which the species have played as father or mother in the crossing whence the hybrid has arisen. At least I have not seen anything which confirms this opinion.

M. Regel asserts (Die Pflanze und ihr Leben, &c., p. 404, et suiv.), that when the hybrid proceeds from species of different genera, their flowers bear the essential characters of those of the father; but we have seen in the Datura ceratocaulo-Stramonium, proceeding from two nearly generically different species, the flowers were absolutely like those of the mother (D. Stramonium); in Nicotiana glauco-angustifolia, and glauco-macrophylla obtained from very different species, they were remarkably more like those of the mother than those of the father; whilst in N. californico-rustica and glutinoso-macrophylla they were very distinctly intermediate between the parent species.

The rule laid down by M. Regel seems to me therefore very hazardous, or at least founded upon insufficient data. For my own part, I believe that these inequalities in resemblance, sometimes very great between the hybrid and its parents, are maintained chiefly by the marked preponderance which many species exercise in their crossings, whatever may be the part which they act (whether as male or female).

This we have seen in the hybrids of Petunia violacea and P. nyctaginiflora which have a greater resemblance to the first than the second; in Luffa acutangulo-cylindrica of which the forms are far more like Luffa cylindrica than the conjoined species; and especially in Datura ceratocaulo-Stramonium and D. Stramonio-laevis, of which all the individuals are incomparably nearer D. Stramonium than the other species, although in one case D. Stramonium fulfils the function of the male, and in the other that of the female.

Commencing with the second generation, the physiognomy of hybrids is modified in a most remarkable manner. Very often the perfect uniformity of the first generation is succeeded by an extreme medley of forms, the one approaching the specific type of the father, the other that of the mother-sometimes returning suddenly and entirely into the one or the other. At other times this recurrence towards the generating types is performed by degrees and slowly, and sometimes the whole collection of hybrids is seen to incline to the same side.

I think it is now placed beyond dispute that this dissolution of hybrid forms commences, in the great majority of eases (it may be in all) in the second generation.

(5.) On the return of Hybrids to the specific forms of the producing species.
What is the cause which determines this return?

In every hybrid which I have examined, the second generation presented changes of aspect, and a manifest tendency to return to the forms of the producing species, and that under such conditions that it was impossible for the pollen of those species to have concurred in bringing them back. We have seen striking examples in Primula officinali-grandifiora, in all the hybrids of Datura Stramonium, D. Meteloido-Metel, the mutual hybrids of Nicotiana angustifolia and macrophylla, N. persica, and Langsdorffii, Petunia violacea and nyctaginiflora, in Luffa acutangulo-cylindrica, and still more in Linaria purpureo-vulgaris. Among many of these hybrids, from the second generation, a complete return to one or other, or even both, of the two parent species has been seen, and approaching them in different degrees; among many also we have observed forms continuing intermediate, whilst simultaneously other specimens of the very same production have effected the return of which I am about to speak. Further, we have stated in some cues (Linaria purpureo-vulgaris) that, in the third and fourth generation, true retrogression towards the hybrid form takes place; and sometimes even we have seen individuals of a plant to all appearance wholly returned to one of the two species, which seemed to revert almost entirely into the opposite species.

All these facts are naturally explained by the disjunction of the two specific essences in the pollen and ovules of the hybrid. A hybrid is an individual in which two different essences are found united, each having its particular mode of vegetation and finality, which are mutually opposed, and are constantly striving to disengage themselves from one another. Are these two essences intimately blended? Do they reciprocally penetrate every part, so that each particle of the hybrid plant, however minute or divided, contains equal portions of both?

It may be so in the embryo and first stages of the development of the hybrid; but it seems to me more probable that this last, at least in the adult state, is an aggregation of particles, both homogeneous and unspecific when taken separately, but distributed more or less equally between the two species, and mixed in different proportions in the organs of the plant. The hybrid, according to this hypothesis, would be a living mosaic, the discordant elements of which, so long as they remained mixed, would be undistinguishable to the eye; but if, in consequence of their affinities, the elements of the same species approached each other and agglomerated themselves in small masses, parts and sometimes entire organs, would then be visible, as we have seen in Cytisus Adami, and the bizarre group of the orange and citron hybrids, &c. It is this tendency of two specific essences to disengage themselves from their combination, which has induced some hybridologists to say, that hybrids resemble the mother by their leaves and the father by their flowers.

Although the facts may not be sufficiently numerous to conclude with certainty, it seems that the tendency of species to separate, or, so to speak, to localise themselves in various parts of the hybrid, increases with the age of the plant, and is more and more pronounced as the vegetation approaches its term. These disjunctions become more manifest in the highest organisms of hybrids, about the reproductive organs; in Cytisus Adami disjunction shows itself in the flowering branches; in the orange anomalies and Datura Stramonio-laevis in the fruit itself. In Mirabilis longifloro-Jalapa and Linaria purpurea the corolla manifests the phenomenon of disjunction, by the separation of the colour peculiar to the producing species. These facts authorise the idea that the pollen and ovules, but especially the former, are precisely the parts of the plant where disjunction goes on with most energy; and what adds a greater degree of probability to this hypothesis is, that they are at the same time very elaborate and minute organs—a double reason for rendering the localisation of the two essences more perfect. This hypothesis being admitted, and I confess it seems to me extremely probable, all the changes which supervene in hybrids of the second and more advanced generations would explain themselves, as it were; but if, on the contrary, it be not admitted, they would be perfectly inexplicable.

Let us suppose in the Toadflax hybrid of the first generation, that disjunction takes place both in the anther and contents of the ovary; that some grains of pollen entirely belong to the paternal species, others to that of the mother; that in others disjunction has not, or, at least, only just commenced. Again, let us suppose that the ovules are, to the same degree, separated both in the direction of the male and female parent; what will result when the pollen tubes descend into the ovary to fertilise the ovules? If the tube of a pollen grain, which has returned to the male parent, meet an ovule separated in the same direction, a perfectly legitimate fecundation will be produced, from which will result a plant entirely returned to the paternal species. A similar combination effected between a pollen grain and ovule, both returned to the female parent, the product will return in the same manner to the species of this last; if, on the contrary, combination is effected between an ovule and pollen grain, separated in opposite directions, they will perform a true crossed fecundation, like that which gave origin to the hybrid itself; and there will again result a form intermediate between two specific types. The fertilisation of an ovule non-separated, by a pollen grain separated in either direction, would give a quadroon hybrid; and since disjunctions, as much in the pollen as in the ovules, can take place in all degrees, every sort of possible combination will result as chance may direct. We have seen these multitudinous forms produced in the Toadflax hybrids, and Petunia from the second generation. the second generation, notwithstanding appearances, still retained some essence of the yellow-flowered Linaria, and this strange particle was sufficient to bring some pollen grains and ovules back either to a mixed state, or altogether to Linaria vulgaris.

Similar actions are produced, though less marked, in the descent of hybrids of the second generation, which seem entirely returned to the type of L. vulgaris, and even to a certain extent in that of Datura Stramonio-laevis, where some individuals return to laevis, preserving up to the third generation the accessory characters which belong to that form of hybrids. All these facts show us that the separation of specific forms allied in hybrids, is not always completed so rapidly as one might be led to suppose, judging from physiognomy and external appearance.

The return of hybrids to the forms of the parent species is not always so sudden as that which we have observed in the Primroses, Petunias, Linaria purpureo-vulgaris, D. Meteloido-Metel, &c.; it is frequently completed by insensibly minute gradations continued through long series of generations. We have seen, for example, in Luffa acutangulo-cylindrica, even in the third generation, that among forty individuals only one was found which had wholly reassumed the external appearance of L. cylindrica.

Hybrids of Nicotiana persica and Langsdorffii, modify themselves slowly, and ten or even more generations may be insufficient to bring them back entirely to the specific forms.

It is remarkable in the latter case, that the hybrids do not present any appreciable mark of disjunction of the two specific essences, which appear intimately blended together in every part of the plant. Nevertheless the traits of one of the two species sensibly disappear from generation to generation, as if extinguished by degrees; but it not unfrequently happens that this extinction takes place with such rapidity as to be completed in the second generation.

En résumé, hybrids fertile and self-fertile return sooner or later to the specific types from which they were derived, and this return is effected either by the separation of the two mixed essences, or by the gradual extinction of one of the two. In the latter ease the hybrid posterity returns entirely and exclusively to one only of the two producing species.

(6.) Are there any exceptions to the law of return of hybrids to the parent forms?
Do certain hybrids become fixed and give rise to new species?

I have not been long enough engaged in the study of hybrids to have formed any settled opinion on this question. Many botanists of good authority believe that some hybrids, if not all, can become fixed and pass to the state of constant varieties, that is to say, true species, intermediate between those of their parents; this is in particular the opinion of M. Regel, who regards it as probable that in the group of willows, roses, and many other genera rich in nearly allied forms, the nomenclature of which is very embarrassing to the botanist, there originally existed but a small number of species (two or three), the fertile crossings of which have given rise to equally fertile hybrids, which, in their turn, crossing between themselves and their parents, have produced, age after age, those multitudes of forms which exist at the present day.

Such may be the case, but it is without proof, and the hypothesis is entirely gratuitous. In my opinion the fact may be explained otherwise in a much more natural and probable manner, viz., by the inherent property of all organisms (at least vegetable) to modify themselves to a certain extent according to the influence of the surrounding medium, in other words, by the innate tendency of what we call species to subdivide into secondary species. How can it be admitted, for example, that roses, which are disseminated over the whole extent of the Old World, from Ireland to Kamschatka, from the Atlas and Himalayas to the glacial ocean, which cover all North America, which are often isolated in narrow spaces and different localities, can have met each other to give rise to hybrid forms?

It would be hardly possible to conceive such a fact. Have roses never been subjected to experiment to ascertain how far they can mutually hybridise, and if their hybrids would be fertile or not? I can affirm this, that I have never obtained a hybrid which manifested the least tendency to form a specific stock.

At present I only know a single instance which might serve as a basis for the hypothesis of fixation of hybrids. Still this fact is doubtful; it is that of AEgilops, closely allied to wheat, which was cultivated at the Museum about ten years, during which the successive generations did not produce any appreciable modification.

It remains to be proved whether the AEgilops cultivated at the Museum (AE. speltaeformis, Jord.) is really a hybrid, and that it does not modify itself during a long series of generations: it would be an exception; but this very general rule would not be weakened, at least so long as the fact remained isolated.

(7.) Is there any precise limit between Hybrids and Crosses?

Most hybridologists insist on making a distinction between hybrids and crosses, and nothing could be more easy to understand; the hybrid results from the crossing of two distinct species, two true species, as M. Regel says—the crosses from that of two races or varieties.

Theoretically, nothing is more clear; in practice, nothing is more difficult than the application of these two words.

For example, ought the produce obtained by the crossing of the Cantaloup Melon and Netted Melon, that of the Netted Melon and Dudaim, that of Dudaim and Cucumis Pencherianus, or even that of Datura Stramonium and of Datura Tatula, &c., to be called hybrids or crosses? This question gives rise to another, that of the distinction of species, races, and varieties, an everlasting subject of dispute among naturalists, which too often ends in a war of words unworthy of the science; to settle which, it is necessary to turn to the examination of what is understood by the term Species, Race, and Variety.

(8.) What, therefore, is a Species, Race, and Variety?

Let us start at the very origin of the notion of species, and not lose sight of the fact that all our ideas arise from the contrast of things.

The man blind from birth has no idea of darkness, because being deprived of the sensation of light he does not perceive the difference between the two; even one possessed of sight would have no idea of the light which surrounds him if the whole world was luminous and that to the same degree. The notion of species does not escape the common law; it is more complex, and is formed from more elements, as we shall attempt to elucidate.

If there existed in nature but one vegetable form, wheat for instance, always and everywhere alike, without any variation in the innumerable individuals which represent it, we might arrive at the idea of an individual and vegetable, but not species; wheat and vegetable would be confounded in one's mind as one and the same thing.

Let us suppose also that nature had created an indeterminate number of different organisms, and each of them represented on the earth by only a single individual, incapable of multiplying itself, but indestructible and imperishable; even here we could not arrive at the conception of a species, for each type of organisation would be isolated, and have no resembling individual.

To have a species it is necessary, therefore, 1st, To have a plurality of similar individuals, that is to say, a group, a collection; 2d, That this group or collection of individuals contrast in some degree with other groups of individuals likewise resembling each other, and yet able to approach one another in some common points which render them comparable.

It follows that the idea of species is connected with that of kind or genus (I mean genus taken in a philosophical sense); that the one fact always supposes the other; that, in a word, they are inseparable and unable to exist apart. And as, in the organic world, individuals have a transitory existence, reproducing themselves by generation, it is necessary, 3dly, In order that species may have consistence and duration—that the resemblance of individuals forming a specific collection shall continue in successive series of generations.

Thus a plurality of similar individuals forming a group, and the contrast of groups among themselves by certain characters common to different groups; and, lastly, The perpetuation of resemblances between the individuals of the same group constitute the elements of species. Species contain nothing more or less.

It is not, therefore, an ideal type, as certain abstract-loving naturalists have suggested; it is essentially a collection of similar individuals. The abstract ideal type of a common organisation is only, as it were, a tie, which in our mind collects similar individuals in the same bundle, and sums up the contrasts (or differences) which separate their group from every other.

It is necessary, then, to return to the pure and simple definition of Cuvier,—viz., A species is a collection of individuals descended from one another, or from common parents, and from these which resemble them as much as they resemble themselves.

Let us remark, in passing, that in thus defining species, Cuvier did not take races and varieties into consideration.

Everywhere where there is a group of similar individuals, contrasting in some measure with other groups, and preserving through a series of generations the physiognomy and organisation common to all the individuals,—there is a species.

It is by their contrast that species are distinguished from one another, and it is by comparison that their contrasts appear. Contrasts may be more or less great according to the objects compared. If they are very great and well marked, all the world acknowlege the specific distinction of the compared forms; if they are very weak, almost inappreciable, opinions are divided; one party separating the feebly contrasting forms into distinct specific groups, the other collecting them into one, and applying to them in the mean time the qualifications of races or varieties.

These collections and separations are purely optional, and they can have no other rule than scientific or economic advantage; in order to determine them it is necessary to be endowed with a certain tact which is ordinarily acquired by experience.

In short, there is no qualitative difference between species, races, and varieties; it is idle to seek one. These three things are formed from one, and the terms which pretend to distinguish them only indicate degrees of contrast between compared forms.

It must be understood that here the question is not concerning simple individual variations, non-transmissible by way of generation, but only forms common to an indefinite number of individuals, and transmitting themselves faithfully and indefinitely by generation.

Contrasts between compared forms are of all degrees, from the strongest to the weakest, which simply means that following the comparisons which are established between groups of similar individuals, species are found of all degrees of strength and weakness; and if it was attempted to express these degrees in so many words, the whole vocabulary would be insufficient.

The delineation of species is therefore as I said before entirely optional; it makes them larger or smaller, according to the importance which is given to the resemblances and difference of various groups of individuals taken with respect to each other, and these appreciations vary according to men, times, and phases of science. How many modifications have certain great species of Linnaeus and Jussieu undergone during fifty years!

The division of old species, their pulverisation, if I may use such a term, seems to have now reached its extreme limits, and many botanists are led by this tendency to complicate the descriptive part of the science in such a way as to threaten to involve the whole life of a man in its minutiae. Notwithstanding this, if those who have inaugurated these scientific refinements have not committed error by taking individual alterations, non-transmissible and not forming a group, that is to say simple variations, for forms common to an indefinite number of individuals, very constant and very faithfully transmissible in every consecutive generation, there is reason to believe that they have proceeded logically. The whole question is to know if it be advantageous to science to distinguish and enrol in its catalogues, these feebly contrasting species; but it is essentially necessary to be assured that the characters which are assigned them are really specific—that is to say, common to an unlimited number of individuals, and always faithfully reproduced in every generation.

But it is more than probable that in a multitude of cases (in the genus Rubus, for example) purely individual variations without persistence, have been taken for common characters, constant and transmissible.

Does it then follow that the terms race and variety ought to be banished from the science? Certainly not, for they are convenient to designate weak species that ought not to be enrolled among the official species, but it is proper to give them their true signification, which is absolutely the same as that of species properly so-called, and to see in forms designated by these terms some unity of a weak kind, which might be neglected without inconvenience to the science.

(9.) Can Artificial Hybridisation furnish a mark to determine what it is proper to distinguish as Species?

I have not the least doubt but that there are some cases where it would be of a slight assistance, and again a greater number where it would not be practicable. Here are some examples of its practical utility.

I have stated before, in speaking of the three species of eatable gourds, that they but slightly differ in outward appearance, and even by their intimate characters, for most botanists cannot clearly distinguish them; Linnaeus himself confounded them in one. But these three plants refuse to give hybrids by mutual crossing; they are then three self-governed species perfectly distinct.

M. Dunal, in his Monograph of Solanaceae, combines into one species Datura Stramonium and D. Tatula, considering them as simple varieties of the same species. But the produce of their crossing does not vegetate altogether like these two forms; it grows much larger and flowers less, inasmuch as it loses its flower-buds in the seven or eight first branches. This disturbance caused in the vegetation of the mixed produce, is an indubitable sign of a difference in the autonomy of the two parent forms; therefore these forms ought to be held as distinct species.

Datura Metel and Meteloides are at least as nearly allied to one another as the two preceding; but, from the second generation, their hybrids cease to resemble them, and a certain number of individuals return to one or other of the two parent forms. Let us therefore conclude that these forms are specific, that they each have their autonomy and deserve, notwithstanding their affinity, to be distinguished from one another.

Nicotiana macrophylla and N. angustifolia combined in the "Prodromus" of De Candolle with N. Tabacum, give hybrids which, after the second generation, manifest a very appreciable commencement of return towards the producing forms. These last have therefore also their manner of growth proper to each of them. Why do we not admit them as distinct in our botanical catalogues?

But when the forms are so closely allied to one another that they are with difficulty distinguished, their hybrids must differ still less from one another than they differ between themselves. The data furnished by hybridisation, therefore, here lose their value; but then it becomes a matter of indifference, whether to separate the two forms as distinct species, or to combine them, by the title of simple varieties, under a common specific denomination.

It follows from all we have said, that the application of the terms hybrid and cross is determined by the rank which may be assigned to the individuals from the crossing of which the mixed forms requiring to be named have been produced—that is to say, it is entirely left to the judgment and tact of the nomenclator.