Allowing all reasonable latitude to the conception of species, there can be no doubt that M. Godron's conclusion upon this head is untenable. M. Naudin's positive evidence suffices to prove the contrary, although his fertile hybrids belong to but four Natural Orders, and their fertility was found to be of every degree. With regard, indeed, to the absolute sterility or otherwise of hybrids, we had sufficient evidence to settle this point some years ago.

From M. Naudin's long experience in connection with the hybridizing of plants, especially species of his favourite group, the Gourds and Cucumbers (Cucurbitaceae), his opinion upon such matters is entitled to great weight. Some of the minor results of his experiments are of considerable interest, so far as they go, as are also the re-statement and discussion of conclusions which he had previously published; we therefore feel it due to our readers to supply them with a short review of that portion of his memoir which has reached us.

M. Godron's memoir, though containing many details of interest to botanists engaged in similar pursuits, is unsatisfactory in several respects, and we concur in the judgment of the Commission appointed by the Academy to report upon the essays, that it should be classed "au deuxième rang, en laissant même une distance notable entre celui-ci et le No. 1 (M. Naudin's)."

The absolute sterility of certain hybrids, M. Naudin believes to be due not only to the atrophy of the anthers, or of the pollen which they contain, but also to some defect in the ovules, the nature of which defect is not clearly made out, nor is any attempt made to explain it. He finds confirmation of this view in the fact that there are hybrids in which a portion of the ovules is capable of being fertilised, while another portion is incapable of fertilisation.

M. Naudin acknowledges, however, with his predecessors, that the sterilizing action of hybridisation tells primarily upon the pollen, and he has not found any hybrids with only a very few grains of well-developed pollen in their anthers, the ovules of which were, at the same time, incapable of being fertilised either by their own or by strange pollen.

The atrophied condition of organs occasioned by hybridising is exhibited not only in the pollen or anthers, but also in the entire flower. For example, all hybrids, of which the common Thorn-apple is one of the parents, invariably lose their flowers from the lower They fall without even expanding. A similar tendency

forks.

is apparent in hybrids of the genera Mirabilis, Nicotiana, and Luffa.

With regard to the general law, allowed by M. Naudin, that the readiness with which species may be hybridised, is generally in relation, or proportion, to the apparent affinity of the parents, some curious exceptions are pointed out. Thus the three" species" of Gourd affording esculent fruit, which resemble each other so nearly that "most botanists are unable to distinguish them," each wholly refuses to be crossed by the pollen of the other two. But this exemplary conduct on the part of these Gourds finds a marked contrast in this respect in the case of their ally, the Melon, which is readily fertilised by Cucumis trigonus, a very different-looking plant.

Hybrids of the first generation, that is, resulting from a first cross between two species, as a rule present no more (or but slightly more) variation amongst themselves than do the individuals resulting from a legitimate fertilisation. And, further, M. Naudin states, that so far as his experience goes, in reciprocal crosses (that is, when A. is fertilised by B. and B. also by A.), the resulting hybrids do not differ more widely from each other than do those of a simple cross. He by no means denies that cases do occur in which the difference may be very marked between the hybrids of reciprocal crosses, but he says such cases are rare and exceptional.

As to the prepotent influence upon the hybrid of the male or female parent, M. Naudin tells us that he has not found anything to confirm the view, held by some as a rule, that certain organs, or sets of organs, in the hybrid, bear the special impress of the male, and others, of the female parent. The inequalities of resemblance to either parent, which are often sufficiently obvious, belong, he thinks, as previously indicated in a Memoir of the author's, published in the Annales (sér. iv. ix. 257), to the marked preponderance which certain species appear always to exercise, whether they take part as male or female parents of the hybrid offspring. Thus hybrids between Petunia violacea and P.nyctaginifiora much more closely resemble the former than the latter species, whichever species be the male parent. M. Godron refers, in his Memoir, to the same tendency in hybrids between these species. We observe, by the way, in a recent number of the "Revue Horticole," (1863, p. 333), a report by M. Carrière upon some experiments tried by M. Quêtier of Meaux, which go far to confirm the view here called in question, and remind us of the importance of numerous further observations, with this matter specially in view.

M. Quêtier has been hybridizing Papaver bracteatum and P. orientale with double-flowered P. Rhaas, and one of these species with the very distinct P. somniferum. M. Carrière states that the general result of the experiments shows that the male parent specially confers the external or physical characters, whilst the female reproduces more particularly "le tempérament," or organic character. All the hybrids were perennial, as were the fe:nale parents; the male parent being annual in each experiment. They also uniformly retained the form and colour of the corolla of the female. M. Godron, in a footnote to his Memoir, says, "C'est en vain que, pendant plusieurs "années j'ai tenté de féconder les unes par les autres diverses espèces "de Pavots." He goes on to say that he has failed even to cross the White Persian Poppy (P. somniferum) with the pollen of P. setigerum, D.C. He adds, that, although these two plants are considered by most botanists of the present day, as races of one species, "les "anciens botanistes les avaient parfaitement distinguées, et je me "suis rangé à leur avis."

This note of M. Godron's seems to show what amount of value ought to be attached to negative evidence, derived from a few isolated cases, in the present inquiry.

But whatever the relation of the first hybrid generation to its parent forms, and however constant in character its members may be amongst themselves, in the second and subsequent generations modifications are introduced, which terminate in an absolute return to one or other of the parent types. This is one of M. Naudin's most important generalisations, though not now broached by him for the first time. The hypothesis which he advances in explanation of the phenomenon deserves further examination. In the great majority of cases, and perhaps, according to M. Naudin, in all, some indication of a return to a parent type is apparent in the second generation, which, as compared with the first generation, is often characterized by extreme variety of individual forms, some partaking of the male, others of the female character, and that in all possible degrees. The return to the parent may be either abrupt or gradual, and the hybrids may incline either to the same or to both parents. Some hybrids return completely to the parent form even in the second generation, and once returned remain stable, so far as is known; others return wholly or in part, and after some generations retrograde to the hybrid character, or even actually assume the character of the other parent. All this bizarrerie M. Naudin explains in this way.

An hybrid is an individual in which we find two different elements, as it were intermixed, but not truly combined. When the intermixture of the parent elements is most intimate, every organ, perhaps, indeed, every individual cell, equally partakes of the compound hybrid nature. When the mixture is less intimate, entire organs, or sets of organs, present all the characteristics of a single parent. It is owing to this tendency of the specific essences to disengage themselves from their union, that, according to M. Naudin, some naturalists have been led to hold that hybrids tend to resemble the female parent in their leafage, the male in their flowers, or conversely; as in the cases above referred to. This tendency to segregation of the specific elements, M. Naudin inclines to believe increases with the age of the individual hybrid, attaining its most marked degree in the development of the essential reproductive organs. In support of this view, he cites the case of Cytisus Adami as an instance of this disjunction occurring in the inflorescence, of Citrus and Datura Stramonio-laevis in the fruit, and of Mirabilis and Linaria in the colour of the corolla. Upon these facts M. Naudin bases the hypothesis that it is in the pollen and the ovules, and especially in the former, that this tendency to disunion of the specific elements is most energetic. "Cette hypothèse admise," he says, " et j'avoue qu'elle me

parait extrêmement probable, tous les changements qui surviennent "dans les hybrides de deuxième génération et de générations plus "avancés s'expliquent pour ainsi dire d'eux-mêmes; ils seraient, au "contraire, inexplicables si on ne l'admettait pas."

"Let us suppose," he goes on, "that in the first generation of the hybrid Linaria purpurea, this disjunction takes place both in the anther and contents of the Ovary in such a way that some of the pollen-grains belong wholly to the male parent, others wholly to the female parent, while in others again, the disunion of the parent elements has not commenced, or has but just commenced. Suppose, further, that the ovules in like manner exhibit disunion to the same extent in the direction of both parents. What will be the consequence when the pollen-tubes descend into the ovary to fertilize the ovules? If the tube given off by a pollen grain which has returned to the male parent encounter an ovule similarly segregated, the consequence will be, a perfectly legitimate fecundation, resulting in an offspring wholly reverted to the species of the male parent. If a similar combination take place between a pollen-grain and ovule both returned to the female parent,

Prof. A. Braun's figures have familiarized us with these freaks in the floral organs, but surely in this hybrid the disjunction originates more usually in the vegetative, than in the reproductive organs. Prof. Braun says of Cytisus Adami, "The formation of sprouts, and indeed the formation of twig-buds, is consequently the turning point here, with which the recurrence of the hybrid into the parent species occurs." He then goes on to say that this reversion sometimes takes place in the inflorescence.-Bot. and Physiol. Memoirs (Ray Soc.), 1853, p. 318, tab. v.

the product will, in like manner, exhibit complete reversion to the species of this parent. If, on the other hand, the ovule and pollen-grain be disjoined in opposite directions, a true hybrid fertilization will take place, like that which gave origin to the hybrid itself, and thus, a second time, a form intermediate between the two specific parent types will result."

But the amount of reversion in the pollen and ovules is of every degree, and as to the combinations which take place between them, 'le hasard seul dirige;' we find, therefore, an infinity of intermediate forms usually resulting, both in the case of the hybrid Linaria, and Petunia, after the second generation. The retrogression of a hybrid in course of return (or apparently wholly returned), to one of its parent species, is explained by assuming that the opposite specific element, not wholly eliminated, had retained sufficient energy to decide the reversion of a few pollen granules or ovules in its favour. The consequence would be, that the hybrids resulting from the combination of pollen or ovules thus affected, would exhibit retrogression to the characters of the first hybrid generation: or, if both pollen-grain and ovule chanced to have reverted in the same direction, the apparent anomaly of a sudden change of species would present.

The return of hybrids to a parent type, though sometimes 'brusque,' is often by insensible gradations, brought about only through a long series of generations. In the case of Luffa acutangulo-cylindrica, even in the third generation, but a single individualwas found, amongst some 40, which had entirely assumed the external appearance of L. cylindrica. Hybrids of Nicotiana persica and N. Langsdorffii may require a dozen generations, or even more, to enable them to complete their reversion. M. Naudin points out that, in the latter case, the hybrids do not present any external sign of the segregation of the two 'specific essences,' which appear to be thoroughly intermingled throughout the entire plant.

The chapter on Reversion of Hybrids to the parent Types, concludes with the following passage:-" En résumé, les hybrides fertiles et se fécondant eux-mêmes reviennent tôt ou tard aux types spécifiques dont ils dérivent, et ce retour se fait soit par le dégagement des deux essences réunies, soit par l'extinction graduelle de l'une des deux. Dans ce dernier cas la postérité hybride revient tout entière et exclusivement à une seule des deux espèces productrices."

M. Regel and others hold that certain fertile hybrids may acquire fixity in their hybrid characters, and thus pass as constant varieties, or, indeed, as veritable species. In this way they explain the multitudinous forms amongst the Willows, Roses, Brambles, and the like.