The Complete Work of Charles Darwin Online

RECORD: Darwin, C. R. 1857. Natural selection, Chapter 9.— Hybridism. CUL-DAR12.(1-145). (John van Wyhe ed., 2002-. The Complete Work of Charles Darwin Online, http://darwin-online.org.uk/)

REVISION HISTORY: First transcribed by R. C. Stauffer in Natural selection, F1583. Text prepared and edited by John van Wyhe. RN1

NOTE: See record in the Darwin Online manuscript catalogue, enter its Identifier here. Reproduced with permission of Cambridge University Press, the Syndics of Cambridge University Library and William Huxley Darwin.

The transcription from R. C. Stauffer's book has been placed alongside the images with minor additions. There are a few discrepancies between the edited text and the original manuscript. For more detail about Darwin's chapter, see of course Stauffer, Natural selection.

Chapter 9.— Hybridism

This important subject concerns us under the following five heads. Firstly: are species invariably sterile when crossed, & are the resultant hybrids likewise sterile? Undoubtedly they are very generally infertile in some degree. But besides the extreme difficulty of deciding in some cases what forms to rank as species & what as varieties, we shall see that there is so insensible a gradation from utter sterility to perfect fertility that it is most difficult to draw any distinct line of demarcation between the two;—more especially as other quite independent causes often simultaneously tend to give some degree of infertility. In some very few cases it is, I think impossible to withstand the evidence that forms which are universally admitted to be good species are quite fertile together & produce quite fertile offspring.

Secondly: are those forms which from their known descent or other reasons must, in accordance with common usage, be called varieties, invariably quite fertile together & produce quite fertile offspring?

This question may be answered by an almost universal affirmative; even in the case of varieties differing in an extreme degree from each other. But we shall see from a few experiments, carefully conducted by hostile witnesses, that the fertility of varieties when crossed can hardly be considered as absolutely universal. Nevertheless the extreme rarity of any, even the slightest degree of infertility between the most distinct varieties more especially in the animal kingdom is one of the greatest difficulties opposed to the theory of species being only strongly marked & constant varieties; a difficulty far more grave in my opinion than the sterility of crossed species.

Thirdly: do the several laws governing the degree & kind of infertility in the first cross & in the hybrid offspring, when these latter are paired inter se, or with one of their pure parents or with distinct species, indicate that species were created with this tendency to sterility in order to keep them distinct; or does the sterility seem to be an incidental consequence of other differences in their organisation? I think the numerous facts, which we shall give, clearly point to this latter alternative.—

Fourthly: can the sterility of one species when fertilised by another & of their hybrid offspring be in any degree explained; so that the view of their sterility being only an incidental consequence on other differences be, at least partially, supported? I think that some little light can be thrown on this subject by the analogy of what often takes place, when organic beings are placed out of their natural conditions of existence.

Lastly: independently of the question of fertility, do the offspring of two species & of two admitted varieties, when crossed, follow the same laws in their variability, in their resemblances to their parents, & in other such points? I believe it can be shown that they do.—

Sterility of species when crossed & of their hybrid offspring.— The sterility of two pure species, when first crossed & that of their hybrid offspring has not always been kept sufficiently distinct. It does not seem a priori improbable that there should be difficulties in the union of two distinct species; we might imagine, for instance, that in plants the pollen tube

3 insert

I may premise that the whole subject is extremely complicated & that it is scarcely possible to make any universal proposition on any one head. On many points it seems to make great difference whether the forms experimentised on, have been long cultivated or domesticated.

of one did not grow sufficiently long or in the right direction to reach the ovule of another species &c; though in truth the obstacle is probably always of a more recondite nature. But when the germ has been fertilised, & a healthy, long-lived hybrid is produced, it seems a far more wonder-ful fact that it should remain throughout its life utterly sterile. It is generally supposed that species have been created with this quality of being sterile one with another in order to prevent the many varied forms in nature becoming blended in extricable confusion. And this at first seems extremely probable; for no doubt if species did blend together, much of the perfect adaptation, —that division of labour—by which each species is excellently fitted for its own particular line of life would be lost; & consequently a lesser amount of life be supported in any given area. It is, also, generally supposed that the hybrids themselves have been rendered sterile in order that when formed (& undoubtedly they are occasionally formed in a state of nature) they should not perpetuate themselves; but on the view

of sterility having been impressed on species by direct creative action, it seems rather strange that it should not have been impressed with sufficient strength to prevent the production of a hybrid in any case.

If it could be proved or rendered highly probable that sterility in the first cross or in the hybrid offspring was a specially created endowment, it would be to us a fatal difficulty. By our theory this sterility, whether or not we can throw light on its origin, must be looked at as an incidental concomitant; like, for instance, the greater or lesser facility with which one kind of tree can be budded on another. This must be so, for sterility cannot have been produced, at least in the case of the hybrids themselves, by natural selection, as sterility obviously could not be favourable to them. In the case of sterility between species & species, in as much as this is favourable to them by keeping their characters pure & unmixed, it is just

possible that the tendency might have been acquired through natural selection; but I know of no fact leading to this conclusion; whereas I do know of facts leading to the view that it is an incidental concomitant of other differences.

The important service rendered by sterility in keeping the forms in nature distinct, perhaps, leads us to overrate its importance as a criterion of species. To explain what I mean: different species of trees graft on each other with different degrees of facility, & though trees in forests occasionally become naturally grafted together, no one would look at this difference in facility, as an endowed quality to prevent the more distinct kinds from becoming inarched in a forest. Yet if it could be shown that invariably different species of trees could not be grafted together or grafted with difficulty, whereas all varieties could invariably be grafted with perfect facility, this quality, though quite unimportant to the plant, would be nearly or quite as good a criterion & as valid an objection to our

theory, as the sterility of species when crossed.

I will first treat of Plants & will subsequently make only a few comparative remarks on animals; for Hybridism has been attended to with infinitely more care amongst plants than with animals. Kölreuter & C.F. v Gartner almost devoted their lives to this subject; & the care, the conscientious accuracy & the astonishing amount of labour exhibited by them is admirable. Next comes the Hon. & Rev. W. Herbert (Dean of Manchester), who experimentised during even a longer period, but who never kept or published such systematic records; but had one advantage in

having large means at his disposal & in being one of the most skilful of horticulturists. Besides these three great authorities, we have Andrew Knight, Sageret, Lecoq & Wiegmann & many others.

I may premise that I have used the term mongrel for the off-spring of two reputed varieties, & that of hybrid for the offspring of two reputed species.

By "reciprocal crosses" I mean the union of species A the father & species B the mother, & on the other hand of B as father with A as mother

By "reduction", I mean the process by which the off spring of A & B, whether species or varieties, is brought, by repeated crosses in successive generations with either A or B, nearer & nearer to that form.

Kölreuter, whose admirable labours have been confirmed by every subsequent observer, concludes that all species whatever, when crossed, are in some degree sterile; but then he cuts the knot, for when in 10 cases he finds two reputed species quite fertile together he assumes that they are varieties. Unfortunately it is

(a) text

Hybrids are designated by the names of the parent species combined by a hyphen; & the first name in the mother; thus Dianthus armeria-deltoides, means a hybrid form D. armeria fertilised by the pollen of D. deltoides.

B text

for instance Dianthus armeria-deltoides & D. deltoides-armeria are hybrids from reciprocal crosses.––

not now possible always to know what plants, he really experimentised on;1 but it is probable that several of the ten would be considered by the best authorities as truly only varieties.

Gaertner, after his truly vast experience, comes emphatically to the same conclusion, namely that two distinct species are never perfectly fertile together: he even disputes the entire fertility2 of Kölreuter's ten cases & will not admit that they are varieties; but as from his table it appears that he has tried only three of them, I do not see what right he has to come to this conclusion.

The laborious plan followed in every instance by Gaertner to measure the fertility of species when crossed with other species (& likewise of their hybrid offspring) was to take the average number of seeds in both pure parents growing naturally (& this is not quite so difficult1 as might have been anticipated as I have found by trying in a few cases), & then to take the maximum number of seeds ever produced by the crossed species. Gaertner took the maximum in order to eliminate the acknowledged ill effects2 of the pollen not being always applied at exactly the right time or not often enough at successive periods,

1 Bastardz. &c. p. 207, 211.

2 Bastardz. &c. p. 212. See also Gartner's Beitrage zur Kenntniss der Befruchtung 1844. p. 332 p. 365. at p 600 there is case in Tropaeolum showing the good of successive applications of pollen.—In some cases, as was also found by Kölreuter, culture in pots tends to increase the fertility of crossed plants & Hybrids.—To give one single instance of the ill effects of artificial fertilisation, taken by hazard from Gaertner Bastard. p. 385: Lychnis vespertina naturally yields 210-230 seeds, but fertilised with own pollen artificially it yielded as maximum only 192 seeds.

~~Ch. 9 Sterilitly of Hybrids~~

~~Gaertner, for when labours one man~~

9 insert

1 . Thus the plants now corresponding with the Hibiscus manihot & vitifolius of Linnaeus, (Syst. 2 Edit) which Kölreuter experimentised on & found quite fertile together, Dr. Hooker tells me appear to be very distinct forms & have even been ranked in distinct genera. It seems, also, difficult to make out what is meant by Sida crista minor & major: Dritte Fortsetzung p. 114, 118.—1766.—

2 Versuche & Beobachtungen Ueber die Bastarderzeugung 1849. p. 414 & 579 et passim. The three which he has tried are Datura stramonium & tatula: D. laevis & stramonium: & Malva sylvestris & mauritiana.

1 Bastardz. &c. p. 207, 211.

and of the plant being cultivated in a pot & placed not in a greenhouse but in a chamber, & lastly of the early castration of the anthers. He admits3 that in order to get the proper maximum, many flowers in successive years should be experimentised on. Hence it is much to be regretted that he did not take for his standards of comparison the same species artificially fertilised with their own pollen & treated in every way like the crossed species. But I sup-pose the labour would have daunted the almost dauntless Gaertner. To test the ill effects of the processes, I have gone through the Table, & have picked out all the cases,4 in which Gaertner actually did artificially fertilise plants, 20 in number, with their own pollen, or with that of another plant, universally admitted, & even by Gaertner himself, to be a mere variety; & these latter are 13 in number. Thus we have altogether 33 cases; & out of them 16 are marked as having had less than full fertility & 17 as producing the full number of seed. Hence the necessary treatment lessens the fertility of every other plant, when artificially self-fertilised.

10 insert

3 Bastardz. p. 210, 214.

4 I have made the case as favourable as possible to Gaertner by not counting those cases of artificial self fertilisation in which as in the Leguminosae there is great difficulty in the operation, as I have myself found. Nor have I counted some cases in which he utterly failed, as this would indicate that there was some fundamental error in the operation; thus I have not included amongst the varieties, Primula veris acaulis & elatior or Anagallis coerulea & arvensis, as these, more-over, are not considered as varieties by Gaertner. I may add in respect to the Leguminosae, that Gaertner crossed 32 flowers of the common Pea with the pollen of undoubted varieties & did not in one single instance obtain full fertility, nor did he with the Kidney Bean.—I have repeatedly tried to cross the varieties of the Sweet Pea, & have always failed, except [ ]. Andrew Knight succeeded with the common Pea, as I have also succeeded.

Now admitting that the number of species, which are quite fertile when crossed is extremely small, the effects of the treatment alone would reduce the number by half. Moreover Gaertner himself admits that to get the proper maximum in crossed plants, many flowers should be experimentised on during successive years; & this has been done in comparatively few cases.

Although these considerations seem to me to throw some doubt on the universality of Gaertner's statement that species when crossed are never equally fertile with the pure species; they do not in the least make me doubt the high generality of his conclusion; for he experimentised on many hundred plants, & he asserts that he never once1 got the full & normal number of seeds. But the case already given in our fourth Chapter of the very numerous experiments made during four years by Gaertner on no less than 170 flowers of Primula veris & acaulis & on Anagallis arvensis & coerulea, in which genera there is no apparent difficulty in effecting a cross, nor have other experimentisers found any difficulty, must give rise to serious misgivings; for Gaertner only twice succeeded in getting any good but scanty

11 insert

1 There are some contradictions between the text & Table, which I cannot reconcile: thus in table it would seem that he once got the full number of seed from Lychnis diurna & L. vespertina .So Matthiola annua & glabra & reciprocally are marked in table as fully fertile; but the contrary is stated in text. p. 102, 197. There are, moreover, some similar contradiction in regard to the fertility of some hybrids in the genera, Malva, Lychnis, Lobelia, & Verbascum.

seed from Primula, & none from Anagallis. As I cannot doubt, at least in the case of the primrose & cowslip, that they are only varieties; & as Gaertner failed either wholly or nearly so in crossing them, one may well ask in how many cases he may have failed in a lesser degree?

I will give two or three examples of the results obtained by Gaertner by counting the seeds. In the several species of Dianthus the normal number in a capsule varies from 80 to 120; whereas in the many species which he cross-fertilised he obtained only from 2 to 54.2 For this genus & for Verbascum & Lychnis he gives the following decimal table;3

	Proportional numbers ofseeds.
Dianthus barbatus naturally fertilised with } own pollen	1.0000
Crossed with pollen of
D. superbus	0.8111
iaponicus	0.6666
armeria	0.5333
chinensis	0.2600
collinus	0.2333
deltoides	0.2222
carthusianorum	0.1111
virgineus	0.0111
&c &c
diutinus	0.0033
Lychnis diurna nat. fert.	Proport. no. of seeds
with own pollen }	1.0000
crossed with pollen of
L. vespertina	0.7777
Cucubalus viscosus	0.2222
L. flos cuculi	0.0021
Silene noctiflora	0.0011
Verbascum Lychnitis nat. fert.	Proport. no. of seeds
own pollen	1.0000
crossed with pollen of
V. phoeniceum	0.8061
— nigrum	0.6336
— blattaria	0.6224
—thapsiforme	0.4081
— austriacum	0.3877
—macranthum	0.2653
— thapsus	0.2142
— pyramidatum	0.0306

2 Bastardz. s. 195.

12v

3 Bastardz. s. 216, 219.––

12bis

In those very few cases in which Kolreuter obtained fertile hybrid plants, he holds that their parents should be considered as varieties; & as with his first crosses Gaertner disputes their entire (perfect) fertility. That hybrid plants are very generally in some degree sterile, & that they stubbornly retain their sterility, I cannot in the least doubt; & I will give a few of the cases, which have most vividly impressed this conclusion on my mind. That hybrid plants are universally sterile, I cannot admit, from facts presently to be given. Indeed if the first cross between two species be ever quite fertile as I believe it to be; nothing is known to make one suppose that its offspring would be sterile.

Even in hybrids when crossed daring successive generations by the pollen of either pure parent, although the progeny in each generation gradually assumes the characters of the pure parent & acquires fertility, yet perfect fertility is the element last acquired.1 A hybrid plant may thus come perfectly to resemble in external appearance one of its pure ancestors & yet be

1 Gaertner Bastardz. s. 450, 459.

12bis v

I may here remark that even in the most sterile hybrids, the pistil, ovary & even ovules appear to the eye perfect, but the ovules will not form an embryo; so it is with the stamens, but the pollen is manifestly imperfect as may be seen by everyone who has ever examined a hybrid. With hybrid animals in like manner the spermatozoa are imperfect, though the microscopical structures of the testis, even in so sterile an animal as the common mule, present no imperfection: the ovules also are to all appearance perfect in the female common mule.1

1 M. Coste is the authority for ovules, see Colin Traite de Phys. Comp. 1856. Tom 2. p. 530. For microscopical structure of testis & state of spermatozoa see, Lallemand in Annal. des Sc Nat 2 series. Tom 15 p. 52. p. 298 &c. For state of ovules &c. in plants, see Gaertner Bastardzeugung s. 262.394

utterly sterile!2 For example, the hybrid called by Gaertner, Nicotiana paniculato-rustica, which means that the gr-gr-grandmother was pure N. paniculata, all nearer relations having been pure N. rustica, (or in the language of breeders having only 1/16 of blood of N. panicu-[la]lata) differed in no respect from N. rustica, except in producing less seed.3 Kölreuter4 found the very same thing in the successive crosses between these same two species; but made reciprocally, so that in the fourth generation, the plant could not be distinguished from N. paniculata, but its pollen was not so good, especially in the autumn. Again Kölreuter5 found that the hybrid Mirabilis jalapa-longifiora fertilised by the pure M. longiflora, produced plants more sterile than their hybrid mother. These plants (which were 3/4 M. longiflora & 1/4 M. jalapa) produced with their own pollen seven seedlings, of which some were quite sterile. But three of the seven produced altogether 15 plants which were very sterile. But one of the fifteen produced nine seedlings; the seeds of these nine seedlings seemed nearly worthless. Here then we have a high degree of

2 Gaertner Bastardz. 449, 460.

3 Bastardz. s. 447.

4 Dritte Fortsetzung s. 47.

5 Compare Nova Acta Petropol. for 1795, p. 324. and 1797 p. 373, 375.

sterility continued down to the gr-gr-grandchildren (self-fertilised in each generation) of a hybrid which was fertilised by one of its own pure parents.

In the cases just given, the first hybrid had been fertilised, either in one or in all the succeeding generations by the pollen of one of the two parent species. In hybrids fertilised from the first by their own pollen, Gaertner repeatedly states that he has never known the fertility to increase in the successive generations, even in the case of the most fertile hybrids; but he has often known it to decrease; so that in a late generation the hybrid could not be fertilised even by the pollen of either pure ancestral species.1 In the successive generations of self-fertilised hybrids, occasionally a seedling is produced extremely like one of its pure ancestral species; but such seedlings are not more, generally less, fertile than the first hybrid.2 Gaertner gives a full account of the successive generations of Dianthus armeria-deltoides: this hybrid yielded seed for ten generations; having sown

1 Bastardz. s. 418-421.

2 Bastardz s. 439.

itself in his garden for the first six or eight; at each generation it yielded less & less seed, & at the tenth its fertility was quite lost.3

I will abstract two analogous cases from Kölreuter: Two hybrid plants of Mirabilis jalapa-longiflora, self-fertilised produced 16 seedlings (grandchildren of the two pure species), most of which were very sterile; but one produced nine seedlings. Of these nine, four were slightly fertile & altogether yielded ten plants, which were excessively sterile, only one having produced anything, namely three seedlings. These three were the gr-gr-gr-grandchildren of the two pure species.4

Kölreuter found the cross between Mirabilis jalapa & M. dichotoma, nearly as fertile as the pure species so that he says he should have doubted whether the parents ought to have been considered as distinct species, had it not been for the portentous stature of the hybrids. One of the hybrids thus raised, & self-fertilised produced 28 seedlings, of which 14 were more fertile than their hybrid parent & some of them even more fertile than their pure grandmother, M. jalapa; but the remaining 14 were considerably less fertile, Kölreuter then took eight of the most fertile of these 28 hybrids,

3 Bastardz. s. 553.

4 No record is given of the fertility of these three last plants. Compare Nova Acta Petropol. 1795. p. 332, & 1797. p. 373, 381, 392, & 403.

& raised from them 34 seedlings: of the 34, (which were gr-grandchildren of the two pure species) only one produced an abundance of seed, & nine were excessively sterile. So that we have seen in some of the hybrids of the second generation a marked increase of fertility (in opposition to Gaertner's statement), but found in all except one of the third generation a high degree of sterility.5

I have given only a few examples, but it is impossible to study the work of Kolreuter & Gaertner, without coming to the con-viction that the fertility of hybrids, when self-fertilised during successive generations, rarely, perhaps never, increases; on the contrary it generally decreases. But this latter fact, I think, is perhaps partly due to an independent cause. Gaertner repeatedly states1 that hybrids, even the less fertile kinds, if artificially fertilised with pollen of their own hybrid sort for some generations, sometimes decidedly improve in fertility. This is a very surprising fact, considering that, as we have lately seen, the artificial process of fertilisation lessens the fertility of about half the pure species experimentised on,—those flowers

5 Nova Acta, 1793. p. 394; 1795 p. 316; 1797, p383-389.

1 Bastardz. s. 418, 421, 554.

which presented any peculiar difficulty for the operation having been excluded from the enume-ration. But I think the increased fertility from artificial fertilisation may be explained in the case of hybrids, by the undoubted good which always follows from a cross with another individual of the same kind, as shown in our third chapter. When a plant is artificially fertilised it is castrated at an early period, & the pollen from another individual, or at least another flower must necessarily be used during each successive generation. On the other hand when a hybrid is allowed to seed spontaneously, it will have to be isolated in a green house or chamber, in order to prevent accidental crosses from either pure parent or allied hybrids, which the experi-mentiser will generally possess; & hence the visits of insects will be checked or quite prevented & the pollen will not then be carried from flower to flower or plant to plant of the hybrid; or if the hybrid is grown in a garden,2 there will seldom be, owing to the trouble of making hybrids & their sterility, a large bed of the same kind. Consequently the hybrid will generally be fertilised by its own individual pollen, & far from receiving the benefit of a cross in each generation, it will suffer from the undoubted

17 insert

2 It must be owned that Dianthus armeria-deltoides, before alluded to, was grown in a garden for six or eight generations, but it is not said whether there were many plants of this hybrid.—

ill-effects of breeding in & in. As in hybrids we already start with <a strong tendency to) sterility, "I think the close interbreeding of carefully guarded hybrids will account in part for their increasing sterility; but not wholly, for the increase in some cases is too rapid, being observed even in the second generation.3 As, however, Gaertner found the fertility of even the less fertile hybrids was actually improved by the process of artificial fertilisation, which we know is so often injurious & can hardly fail to be injurious in some degree, I can hardly think that the fertility of any hybrid has been fairly tested for successive generations until a large bed of it has been left growing in the open air, freely exposed to the visits of insects & the other means by which nature habitually crosses the individuals of the same species.

Let us now hear the results arrived at by W. Herbert, the third greatest Hybridiser who ever lived. He agrees generally in the closest details with Kölreuter "Gaertner, with one important exception, namely that he attributes much more fertility

3 Gaertner Bastardz. s. 421.

in many cases both to the first cross between species & to their hybrid offspring. He says1 "it is certainly not correct as a general law, though some have stated it, that the number of seeds in one pericarp is smaller in hybrid than in the case of natural impregnation; it is true in some cases, & the reverse occurs in others."This difference may be partly due to Herbert having accidentally experimentised on more favourable groups of plants; to his having in many cases raised at some time whole beds of hybrids; partly, perhaps, to his not having so closely observed the slight shades in sterility; but chiefly I am inclined to think to his great skill as a horticulturist & to his having ample means in numerous green & hot-houses: for it is certain2 that hybrids are more sensitive in their fertility to their conditions of life than are pure species.

I will now give some of Herberts principal facts:

he asserts,3 that the hybrids from "the yellow Linaria genistifolia & the purple L. purpurea, & from Pentstemon angustifolium & pulchellum are both perfectly fertile, sowing themselves about the garden". Again the hybrid from Lobelia siphylitica & fulgens "reproduces itself abundantly".

1 Amaryllidaceae p. 354.

2 Gaertner Bastardz. s. 10, 32, 384.

3 Amaryllidaceae p. 345.

19 insert

Petunia came into flower. Gaertner fried this cross, but with him the fertility did not come up to that of the parent species.5

Of the species of Gladiolus

5 Bastardz. s. 388, 719. The reciprocal crosses (s. 177) between these species are not of equal fertility.

Gaertner tried the cross between these same two Linaria & absolutely failed: he seems, also, to have found the hybrid between the two Lobelias much less4 states that the hybrids between Petunia nyctanigenaeflora & phoenicea are "not only fertile but seed much more freely than either parent":—<Here we see that Herbert has apparently tested the fertility of the hybrid by Gaertner's plan, namely by actual comparison of the number of seeds produced &> in this instance there could have been no error by fertilisation through one of pure parent species, for the hybrid was forced & set its seed before any other

Petunia came into flower. Gaertner tried this cross, but with him the fertility did not come up to that of the parent species. Of the species of Gladiolus,

Herbert6 remarks that there, can scarcely be two more dissimilar than G. cardinalis & tristis; "yet the produce of these intermixed is fertile, & where the third species G. blandus has been, also, admitted into the (compound) union, it is fertile in the extreme, incomparably more so than the pure G. cardinalis ["]: this, I may add makes the case very singular, as complicated crosses of three or more species are usually very sterile.

In Hippeastrum, Herbert says1 that the species when crossed produce "offspring invariably fertile". In Crinum2 Herbert had

4 Amaryll. p 379.

5 Bastardz. s. 388, 719. The reciprocal crosses (s. 177) between these species are not of equal fertility.

6 Journal of the Horticult. Soc. vol 2. 1847 p. 88.

1 Amaryll. p. 345.

when ferti[li]sed by the pollen of some other hybrid Hippeastrum yielded much more seed than with its own pollen. He was thus led to try an analogous experiment on a pure species, namely on a bulb of Hippeastrum aulicum, lately imported from the Organ Mountains of Brazil; this bulb4

4 [The manuscript is sheared off at this point. The corresponding passage in Variation, II, 139[...]]

21A

vegetated freely."Herbert adds "this is a strange truth, "the more remarkable from the difficulty of obtaining cross-bred seed at all in the genera which are most nearly related to Hippeastrum, namely Habranthus & Zephyranthes ".—

21 insert

Here it is impossible that the species self-impregnated in its wild state could have been more fertile, even if we assume that it is normally as fertile. This last case leads on to perhaps the most extraordinary fact recorded on hybridism, namely those cases in which a plant is less fertile with its own pollen, than with the pollen of a distinct species; though its own pollen is proved to be good by fertilising other species. Herbert was led to make | the following experiment from having observed during several years (in a letter to me in 1839 he says he had then made the observations during five seasons) that every hybrid Hippeastrum3

3 Amaryllidaceae p 371. Journal of Hort. Soc. vol 2. 1847 p. 19.

21 insert v

(a) This leads a

(Used)

Gaertner has observed5 analogous facts occasionally occurring in Lobelia: thus in two instances, the ovaries of L. fulgens could not be fertilised by their own pollen, though they set seed with the pollen of L. syphilitica & of L. cardinalis; & yet the pollen of those flowers of L. fulgens was good, for it fertilised L. syphilitica. So it likewise was with Verbascum nigrum.

5 Bastardz. s.357.

Kölreuter moreover1 described long ago a similar case in Verbascum phoeniceum, which was fertilised by 4 other species, but yielded no seed to its own apparently good pollen. | In Passiflora2 also, it has been found that the plants could be much more easily fertilised by the pollen of a distinct species, than with its own.—

In these several curious cases, more especially in those which are only occasionally occurrent, we must suppose that the plants are in an abnormal state; though they are not to the eye in any way imperfect; & both pollen & ovaries are quite capable of performing their proper function when exposed to the action of a distinct species. We may attribute this result in part to the good always resulting from a cross. But it is a most singular fact that the self-fertilising power of a plant should ever be actually inferior to its perfect capacity simultaneously for hybridising & being hybridised.

1 2 Fortzet. p. 10. & 3 Fort. p. 40.

23v

2 H. Lecoq De la Fecond: et L'Hybrid: 1845. p. 70. Likewise M. Mowbray in Transactions of Horticultural Soc. vol 7. p. 95.—Bosse has made same observation, Gaertners Bastardz. s. 64.

Lastly we must allude to the numerous & complicated crosses, though their history is very imperfectly known, now carried on for many years by horticulturists, amongst the species of Azalea, Rhododendron, Calceolaria, Fuchsia, Rosa, Petunia & Pelargonium. In this latter genus, according to | Herbert,3 the first great step "was the production of the plant called, ignescens by the intermixture of the group to which betulinum, citriodorum &c belong, with a tuberous rooted scarlet one. The fertility of that plant set wide the doors to innovation but the stream is confined within certain limits."| Very many of the beautiful varieties of Pelargonium, are extremely sterile; but this seems often quite independent of their hybrid origin; some varieties having become barren & some having come fertile after a few years culture.4 The species & varieties of Calceolaria have been crossed, as Herbert remarks ad infinitum: he states5 that even the amongst the highly cultivated & generally sterile Races is of hybrid origin.1

Everyone has seen the splendid results of the most complicated crosses between the several species of Azalea & Rhododendron2 : Mr. Gowen who raised some of the early crosses at Highclere assures me that some of them yielded numerous self-sown seedlings.

3 Journal. Hort. Soc. vol 2. p. 95.

4 D. Beaton in Cottage Gardener 1856 p. 44, 55, 61, 94, 109.

5 Journal Hort. Soc. vol 2. p. 86.

2 Herbert, Amaryll. p. 359. do in Hort. Journal vol 2. p. 86.

hybrid from C. integrifolia, a woody shrub, & C. plantaginea, as humble & herbaceous as a plantain, though at first sterile, during the second year "reproduced itself as perfectly as if it were a natural species from the mountains of Chile". The other great American genus of Fuchsia has likewise been crossed in the most complicated manner. | Yet there is no difficulty in getting abundance of seed from several of the varieties & so it is with Petunia.—One of the best seeders amongst the highly cultivated & generally sterile Races is of hybrid origin.1

These facts, though many of them are not known with scientific precision, are important because it might have been inferred from Kölreuter's & Gaertners experiments that the successive generations of

1 Cottage Gardener 1856. p 206—Mr. Appleby on the Petunia—P. Phoenicea crossed with P. violacea have produced all the pink & purple vars—crossed with P. nyctaginiflora have produced the white vars.

2 Herbert, Amaryll. p. 359. do in Hort. Journal vol 2. p. 86.

25 insert

[deleted text]

25 insert v

I applied to Mr. C. Noble of Bagshot, so well known for the numerous splendid Rhododendrons raised by him, on the degree of fertility of his hybrids, & he has given me the names of several, the offspring of R. arboreum & maximum, & of altaclerense (itself a hybrid from Pontico-catawbiense fertilised by arboreum) & catawbiense, which he says he is sure produces as many seeds as any pure species. He adds that the kind raised in great numbers as stocks for grafting, is a hybrid from R. Ponticum & Catawbiense, & that this "seeds as freely as it is possible to imagine."

25 insert 2

Cottage Gardener 1856. p 206—Mr. Appleby on the Petunia—P. Phoenicea crossed with P. violacea have produced all the pink & purple vars—crossed with P. nyctaginiflora have produced the white vars.

both simple hybrids & hybrids reduced one or two steps towards either parent form, invariably became more & more barren, but with what is known of the history of these several genera of highly cultivated plants, it is scarcely possible to believe in this conclusion. A steady & quickly increasing degree of sterility would have struck nurserymen & horticulturists. These facts moreover, strengthen my previous remark, that the only fair way of testing, as nature would test, the fertility of hybrids in successive generations, is to have numbers of the same kind growing in the open air, & allowed freely to cross.

Reviewing all these facts on the fertility of the first cross between two species & of their hybrid offspring, the precise observations of Kölreuter & still more of Gaertner demonstrate that in the great majority of cases no union whatever can be effected between two species; & that when effected their fertility & that of their offspring is very generally impaired to a serious degree. On the other hand, the fact that when Kölreuter found what he considered

perfect fertility, he at once ranked the two forms as varieties;—that in Gaertner's experiments the fertility not rarely approached pretty closely to that of the pure species, & that the necessary artificial fertilisation is shown to be in about half the cases decidedly injurious;—that Gaertner failed in some cases in which Herbert succeeded; that he failed almost entirely in crossing primrose & cowslips & entirely in crossing the blue & red Anagallis,—that Herbert & others in some cases found undoubtedly distinct species when crossed not only fertile, but actually more fertile than the pure species, every ovule in the pericarp in one instance setting;—that in some few cases plants have hybridised other species, & been hybridised by them, far more readily than they could be self-fertilised;—and lastly attaching some little importance to the unscientific experiments tried on so large a scale by Florists; it seems to me impossible to admit, that species when crossed & their offspring are invariably sterile even in a slight degree, or that the sterility invariably increases in successive generations.

On the difficulty in distinguishing species from varieties by the test of fertility.—Forms known to have descended from a common parent are universally admitted to be varieties; but this can seldom be told except with cultivated plants, & in other cases, in order to decide whether to rank a plant as a species or a variety, we must rely on the opinions of the best & most cautious Botanists, who, however may of course be easily mistaken. If we followed Kölreuter's simple rule & called all plants, which were quite fertile together, varieties, it might be thought that we should at least arrive at a decided result; but this is not so, for we have seen that the two most laborious & careful experimentisers who ever lived, often come to a diametrically opposite conclusion on this head: and this alone almost suffices to show that, practically, fertility will not serve to distinguish varieties from species.

We will now briefly consider those cases, in which Gaertner found considerable fertility in the first cross & in the hybrids, but in which on counting the seeds, he ascertained their maximum was less than the average yielded by the pure species growing freely under the most favourable conditions. By comparing, in the case of closely allied & more or less doubtful species,

29a

(Small type) References in text

Matthiola annua, glabra & incana: Gaertner experimentised on the two first species of Stocks. In the table at end of column (Bastardz. p. 706) their union is marked as yielding the full & normal number of seeds, but in the text (p. 102 &c) they are expressly said not to be perfectly fertile. Mat. glabra fertilised by annua yields more seed than the reciprocal cross. The hybrids from this cross are said not to vary (p. 168) like the off-spring of crossed varieties; but yet (p. 247) some slight variation is

the evidence from fertility with that which can be derived from any other source, & with the opinions of the best Botanists, we shall see in how curious & instructive a manner, the evidence is almost equally doubtful, & graduates away on both sides.)

admitted: Gaertner attributes the greatest importance to variation in the first offspring as determining whether forms are to be ranked as species or varieties; but I cannot think, and we shall hereafter have to discuss this point, that this is of so much importance. With respect to the hybrid Mat. annuo-glabra (p. 388) its fertility is said to approach very nearly but not to equal, that of its pure parents. Kölreuter (Dritte Fortsetsung p. 116) crossed Mat. annua & incana & reciprocally; he obtained perfectly fertile capsules, & the hybrids raised from them, are expressly stated to be as fertile as the pure species ever are: Gaertner doubts this, but he does not appear to have tried these two forms. Now Robert Brown, <a host in himself,> & a few other Botanists as Spach consider Mat. annua, glabra as only varieties of incana, whereas most Botanists have treated them as species. As far as fertility serves for a test, we see the two greatest authorities divided. I should trust most to Gaertner on the point of fertility; & everyone I suppose on Botanical grounds would prefer leaning on the opinion of Robert Brown.

Datura stramonium & tatula: These species when united reciprocally are not according to Gaertner (p. 197) equally fertile: the hybrid off-spring do not vary (p. 168); Gaertner (p. 385) asserts that the hybrid D. stramonio-tatula gave at most 220-280 good seeds, whereas the two pure species give respectively 800 & 600 seeds. Gaertner (p. 273) lays great stress on the fact that these two species when crossed with D. quercifolia, yield very different hybrids, which Gaertner does not believe is the case when two varieties of one species are crossed with another species: but this conclusion

seems grounded exclusively on some experiments with the varieties of Nicotiana, hereafter to be discussed, & which seem to me to be contradicted by some experiments made by Kölreuter on the same genus.—

Kölreuter (Zweite Fortsetsung p. 125) tried reciprocally Datura stramonium & tatula & found the hybrids thus produced as fertile as the pure species: he also got hybrids from D. stramonium & inermis or laevis (Acta Acad. St Petersburg 1781. Part ii. p. 304) & these he calls "foecundissimae," yielding 400-500 or even more seeds in each capsule; so that he concludes all these species are varieties. Now Asa Gray in his Manual (2 Edit [p. 341]) (See in his large Flora whether he enters in details) considers D. stramonium & tatula as only varieties; & the late Dr. Bromflield an excellent observer states (Phytologist Vol. 3. p 597) that he traced in the U. States every grade between D. stramonium & the purple D. tatula. Are we then to throw over such excellent observers as Asa Gray & Bromfield & believing Kölreuter was mistaken, follow Gaertner & Linnaeus & deduce that these two forms are true species. Or shall we believe that the lessened fertility of Gaertners hybrids, careful & conscientious as he seems to have been, was due to the treatment of the plant; in an analogous manner to the fact recorded by Kölreuter (Acta Acad 1781 p 303) that the hybrid Datura <inermi> laeviferox when planted in the open air yielded from 120-130 seeds in each capsule, but in pots only from 60-70?

31v

Gaertner, also, crossed D. laevis with stramonium, & the union (p. 687) yielded less than the normal number seeds. On the other hand,

I confess my opinion of Gaertner's circumspection is so high that, though I do not believe that Kölreuters hybrid Daturae were as little fertile as Gaertner's yet I do believe, considering that with Gaertner, the hybrid D. stramino-tatula yielded only about 1/3 of the seed of the pure species, that there really is some lessened fertility; nevertheless it may be questioned, as we shall see from facts presently to be given, whether these forms should be considered as anything but varieties.

Lychnis diurna & vespertina: Gaertner made very numerous experiments on these plants during several years. In the table at the end it seems that he occasionally got the normal number of seeds from L. diurna fertilised by L. vespertina; & from its resultant hybrid; as well as from the reciprocal cross & its resultant hybrid. But in the text (p. 218) he arranges the species of this genus according to their sexual affinity, as deduced from counting their seeds, whence it would appear that L. diurna fertilised with the pollen of L. vespertina yields as a maximum only 777/1000 of the average number of seed of L. diurna; & secondly that L. vespertina fertilised by the pollen of L diurna, yielded 810/1000 of the average seed of L. vespertina; Hence, also, we here see that the reciprocal crosses are not of equal fertility. With respect to the fertility of the hybrids, L. diurnovespertina (p. 385) yielded the maximum of 125; the pure L. diurna yielding 150-180, & L. vespertina through artificial impregnation with its own pollen giving 192 seeds, but when naturally fertilised, from 210-230 seed. On the other hand,

the reciprocal hybrid L. vespertino-diurna, artificially impregnated with its own pollen, gave (Kenntniss der Befruchtung p. 598) a maximum of 234 seed, which is a greater number than occurs in either pure parent.

Gaertner gives without any details, one marvellous statement (Bastarderzeugung p. 515) namely that the hybrid resulting from the union of the two reciprocal hybrids Lychnis diurno-vespertina & L. & L. vespertino-diurna is absolutely sterile! I have only to add that Gaertner remarks (Bastarderzeugung p. 577) that the hybrids from Lychnis diurna & vespertina in their variability are analogous, to the products of two varieties crossed. Hence we see that the general evidence, adduced by Gaertner is in favour of these two forms being species, though it seems from the Table at the end of the book that the fertility of the first cross & of the hybrids occasionally mounts up to the full normal number; & one case is given in detail of hybrids being fertile in excess.—Now if we look to the opinion of Botanists we find Linnaeus, Sir James Smith, (Hooker) & Prof. Henslow consider them, as mere varieties, whereas

33v

C. C. Sprengel gives facts showing they they are species.—

[Not transcribed by Stauffer. JvW]

33 insert

(B) Gaertner further states (p. 68) that in L. diurna after self-fertilisation, the pistil & petals turn colour in half an hour, but when dusted with the pollen of L. vespertina, they do not change their colour till from one to one & a half hour—facts which show in L. diurna a quicker fertilisation by its own pollen than by that L. vespertina.

most botanists think them distinct. I have cultivated L. diurna for three generations & could not observe in its variations any approach to L. vespertina; & as far as I can judge from the various published statements the two ought to be considered distinct. (C. C. Sprengel has some remarks on this subject in his Geheimniss der Natur s. 260: Gaertner discusses the subject at length: Tausch shows in the Flora 1833 B 1. s. 225, that L. diurna sometimes produces a white coloured variety.)

Gaertner crossed several times in several ways, Cucubalus alpinus, C. Behen latifolius, C. Behen angustifolius, C. Italicus, C. pilosus & C. littoralist & never obtained the full & normal complement of seed, though the fertility seems to have been generally only one degree under it. This was the case even with the two admitted varieties of C. Behen.

In all the several Floras which I have consulted, Dianthus glaucus is considered a variety of D. deltoides: Kölreuter (Dritte Fortsetsung s. 94) sowed seeds of the former & ocasionally raised plants closely resembling D. deltoides: further he (Nova Acta Petro: 1785 s. 284) crossed these two

34 insert

Yet, as I am informed on the high authority of Mr. Bentham all these forms are in his opinion & in that of many Botanists, only varieties of Silene inflata. All such cases of course remain subject to the doubt whether Gaertner experimentised on properly named plants.

forms, & raised many plants "in summo gradu foecundae". Nevertheless Gaertner (Bastardz. s. 539, 414) thinks Kolreuter is in error, & concludes that these are true species.

I need not here do more than to recall to mind that Gaertner after the most persevering effort concluded that, Primula veris, acaulis, elatior, (the primrose, cowslip & oxlip) are good & distinct species, from being highly infertile one with another; & that Anagallis arvensis & coerulae, tried on 19 flowers were absolutely sterile! I must believe that these experiments failed from causes analogous with those which prevented his entire success in his crosses of the garden varieties of the common Pea & Kidney Bean.—

These facts suffice, I think, to show that when forms in nature approach each other so closely that Botanists are divided whether or not to rank them as species, their fertility when crossed, & that of their hybrid offspring, approaches so closely to the normal value, that it is most difficult to decide the point by the test of fertility. Anyhow we see two observers, the most experienced of any, having the same theoretical opinions on the independent creation of species & who grudged no labour to arrive at the truth, often coming to directly opposite conclusions.

On the infertility of varieties when crossed.—I remarked at the beginning of this chapter, that the perfect fertility, even in most cases the decidedly increased fertility, of the most distinct varieties when crossed, was the gravest difficulty in our present subject. It is notorious, for instance, that the several varieties of cabbage, though so widely distinct in general appearance, are perfectly fertile together. So it is with our different breeds of cattle, dogs & poultry; but we shall, as yet, confine our attention to plants. I shall now show, taking the best evidence which can possibly be obtained,—though I admit that the evidence here is not un-impeachable, any more than in regard to the sterility of undoubted species when crossed,—that some forms which are generally or universally recognised as varieties, are not perfectly fertile together. These cases, though necessarily few in number, as the crossing of varieties has very seldom been carefully attended to, are, I think, of great importance in our present subject.

Gaertner fertilised 13 flowers, on different plants,

36V[1]v

Kenntniss der Befruchtung p. 137; and Bastarderzeugung p. 92 & p. 181.

36V[2]v

Gaertner made most numerous experiments on many species of Verbascum, & with nearly all the species, he tried both the white & yellow varieties of V. lychnitis & blattaria; & he asserts most distinctly in two of his works3

36V[3]v

Zea minor semine luteo: Bastardz. s. 87, 169. From the table at the end, it, moreover, appears that altogether this experiment was tried on 22 flowers.—

that the white-flowering varieties of Verbascum crossed with the white-flowering species bear more seed than when yellow-flowering varieties are crossed with white flowering species. of this genus. So again similarly-coloured varieties of the same species are more fertile together I than when differently coloured varieties of the same species are crossed. That these really are varieties, no one has doubted; & Gaertner actually raised4 one variety from seed of the other. The serial arrangement of the species & varieties according to their sexual affinity or number of seeds yielded, was ascertained by experiments on no less than nine species repeatedly crossed by both the yellow & white varieties of the above two species.5 In one instance alone, Gaertner enters6 into minute details on this head:

4 Bastardz. s. 307.

6 Bastardz. s. 216.

38 insert

5 Some errors have crept into the Table; for the degree of fertility assigned to each species & variety does not always perfectly accord with their serial arrangement.

I am enabled to give one other & slightly different case & founded on much better evidence than the last. Kölreuter describes minutely five varieties of

40a

infertility between the varieties of the same species, as tested by crosses with an extremely distinct species, seems to me a particularly interesting case.

In the varieties of the Maize, Verbascum & Gourd, it will have been observed that the fertility seems to have been slightly lessened only in the first cross, & not in the mongrel offspring: had, indeed, the yellow & red-seeded Maize, for instance,

40 insert

It may be noticed that Gaertners strong wish to draw a strong line of demarcation between species & varieties, must have made him unwilling to admit his own curious discovery of this sexual affinity in Verbascum of varieties to each other & of varieties to distinct species, with flowers of the same colour.> *(1)

in addition produced barren offspring, they would have been unanimously ranked as distinct species. In Nicotiana, however, certain varieties of one species, when crossed with a very distinct species, did produce hybrids more sterile, than when another variety of the very same species was used in the same cross as either father or mother. Although the infertility of hybrids is in itself, I think, a much more remarkable fact than the infertility of the first cross, yet this latter fact has been universally acknowledged as a test of equal value for discovering the essence of a species; indeed in some respect it seems of higher value, as more directly tending to keep the forms in nature distinct. We shall, also, presently see that in crosses of undoubted species there is by no means a uniform relation between the difficulty of the first cross & the sterility of the hybrid offspring.

To our short list of varieties in some degree infertile together, or having different degrees of fertility when, as in Verbascum & Nicotiana, crossed with other species,

those several Botanists who believe that the several forms before specified in the genera, Matthiola, Datura, Lychnis, Cucubalus (Silene) & Dianthus, are not species, but varieties, will have to add them to the list.

On the other hand some may say that not only the forms just alluded to, but that the reputed varieties of the Maize, Verbascum, Gourd & tobacco are true species. But even Gaertner with his strong predisposition to call the finest forms species, did not venture to do this: in the case of Verbascum no botanist considers the yellow & white flowered forms of V. blattaria & lychnitis as species: Gaertner's statement that he raised one variety from the other would, also, have to be dis-believed. Moreover on this view, it must be admitted in the case of the maize & tobacco that the hybrids raised from crossing these supposed species are perfectly fertile.

With much more apparent probability, others may say that the difficulty

in crossing the varieties in the four genera must have been entirely caused by some want of skill or by the injurious effects of the necessary manipulation. But in the Zea & Cucurbita no manipulation was requisite, as the sexes stand in separate flowers. And if in these several cases, the lessened fertility has been thus caused; & if we add to these cases, those of the primrose, cowslip, & oxlip, of the anagallis & all leguminous plants, it cannot possibly be any longer pretended that we have evidence worth anything on the infertility of a vast number of related forms, which are universally acknowledged to be distinct species, but which when crossed are in some degree fertile together.

As for myself I believe in the very general infertility of even closely related species when crossed; & further I am forced to believe that the forms generally called varieties, even those which have originated under culture, are occasionally in some slight degree infertile together. Believing this, & as it has, I think, been shown in the first section of this Chapter that some few undoubtedly distinct species are perfectly fertile together, I conclude, that not only the test of fertility practically fails, as we have seen by comparing the results arrived at by Kölreuter & Gaertner; but that theoretically the forms called species cannot in all cases be distinguished from those called varieties either by their fertility when first crossed; or by the fertility of their offspring.

Laws & circumstances governing the infertility of crossed <species> plants & their hybrid offspring —.

My object in treating this subject at some little length, is to see how far the facts indicate that the infertility is a quality, especially created, in accordance with the common view, to prevent species mingling in nature.

When the results obtained by the almost innumerable experiments tried on various plants are compared, we find a perfect gradation from absolute sterility to perfect fertility,—even to fertility, according to Herbert, beyond the natural degree. In plants belonging to different Families, the pollen of one when placed on the stigma of another, has no more effect than so much inorganic dust: this, also, is not rarely the case1 with species even of the same genus. The first evidence of some sexual affinity between two plants, is a little shorter persistence of the corolla or calyx in a flower when dusted by the pollen of another species, than when simply castrated.1 Even the pollen itself behaves differently when laid on the stigma of an allied plant, to what it does on one not at all sexually allied to it; & so does the stigma itself.2

1 Gaertner. Bastardz. s. 96.

2 R. Brown Linn. Transact. vol. 16 p. 708 Gaertner Bastardz. s. 9, 19, 110.

Gaertner, also, describes a curious gradation3 in the quicker & quicker withering & change of colour in the corolla, & in the more & more perfect development of the pericarp & external parts of the seeds, in accordance with the closer sexual affinity of the crossed plants. We then come to plants, in which after trials prolonged over many years, one or two out of thousands apparently perfect seeds, will germinate4. From this low degree of fertility, in which a single seed is occasionally found to germinate, a perfect series can be most easily shown, in the increasing proportional number of good seeds, up to nearly perfect fertility: but facts enough have already been incidentally given on this head.

Hybrid seedlings from plants with very little sexual affinity are sometimes weakly & tender, & cannot be raised with the greatest care:5 but generally hybrids, as we have seen in our third chapter, are more hardy, vigorous, precocious & of larger stature than their pure parent species. Herbert, I may add | has clearly shown6 that some kinds of Narcissus now cultivated in our gardens must have been formed by hybridism between two & three centuries ago, & have been propagated ever since by | offsets./

3 Bastardz. s. 68, 102.

4 Gaertner Bastardz. s. 8, 101, 138.

45 insert

5 Herbert has given striking cases in Horticultural Journal vol 2. p. 11. also Amaryllidaceae p. 360; See also Gaertner's Bastarderz. p. 520 p. 548; & Kölreuter in Nova Acta Petrop. 1794 p 391, and 1795. p 325.

As in the first cross, so in the hybrid offspring a perfect gradation from sterility to fertility can be shown. But even in the most sterile hybrids, the pollen of either pure parent7 will generally cause the flower to endure longer than it otherwise would have done: so that even in these hybrids, the sterility can hardly be considered absolute, though such hybrids never have yielded, & probably never could yield a single seed, which could germinate.

The degree of facility in effecting a cross, & the fertility of hybrids are both much influenced by the more or less favourable conditions to which they are exposed.8 Besides this extreme susceptibility to external circumstances, it is most clearly proved by Kölreuter & Gaertner that the degree of innate fertility in the same hybrids is excessively variable. Hybrids raised one year from the same parents will be far more fertile than others raised another year.1

7 Gaertner.. Bastardz. s. 412, 533.

8 Gaertner Bastardz. s. 10, 32, 384.

1 Gaertner Bastardz. s. 385, 391. Thus Aquilegia vulgaris-canadensis sowed itself freely one year; but similar hybrids in a subsequent year would produce scarcely any good seed.

Both simple hybrids in the first & succeeding generations, & hybrids in course of reduction to either parent form2, when raised in these several cases from the very same capsule, will differ extremely in fertility3. So strongly marked is this variability in the fertility of hybrids from the same parents & under the same conditions that Gaertner has remarked4 that this quality is to a large extent contingent on the individual, as well as on its parentage. Again, in regard to the first cross between two pure species, sometimes individuals are found5 which obstinately refuse to cross; & he concludes that many individuals must always be tried, before the sexual affinity of two forms can be determined.

The fact that no one has crossed plants belonging to distinct Families, & that in bigeneric crosses, the genera have generally been closely allied6 , & the cases before given of the high or very nearly perfect fertility of the crosses between species which are so closely

2 Kölreuter 2 Forts. s. 98: Gaertner Bastardz. s. 461.

3 Kölreuter 1 Fortsetz. s. 14. Gaertner Bastardz. s. 366, 554.

4 Bastardz. s. 143, 406

5 Gaertner Bastardz. s. 165.

47 insert

6 The following are some cases of bigeneric crosses: Rhododendron & Azalea.—Rhododendron & Rhodora, a remarkable cross.—Lychnis & Silene (Gaertner)—Hymenocallis & Ismene (Herbert), Gloxinia & Sinningia,.—Brunsvigia & Vallota (D. Beaton) not considered good genera.—Cereus, & Epiphyllum not considered good genera by many Botanists.—Cereus (Gaertner Bastard. 179) with Echino-cactus & Melocactus, genera usually admitted.—

related that botanists have doubted whether to call them species or varieties, all show that commonly there is a pretty close parallelism between systematic affinity, & the fertility both of the first cross & of the hybrid offspring. But this parallelism is very far from being invariable or uniform.7 Every single experimentiser has been struck with surprise at the numerous cases of most closely allied species8 which cannot be made to unite or which produce utterly sterile offspring, & on the other hand at the very great dissimilarity of some forms which unite most easily. In the same Family, the species of one genus, for instance Dianthus, will cross very easily & yield unusually fertile hybrids, whereas in another genus, as in Silene, the most persevering efforts of Kölreuter & Gaertner9 failed in producing hybrids between the very closely allied species. Even within the limits of the same genus, for instance Nicotiana, in which the many species have been more largely

7 Gaertner Bastardz. 121, 168, 408. Gaertner says, (s. 194) the parallelism is less strict with the fertility of the hybrid than with the facility of making the first cross.

8 Gaertner Bastardz. s. 174, 164 gives a whole catalogue of such forms.

9 Bastardz. s. 140, 195, 197.

crossed than in almost any other genus, Gaertner found one species, N. acuminata, which is nearly related to the other species, yet absolutely failed to fertilise or be ferti[li]sed by no less than eight other species!1 Analogous facts were observed by Kölreuter in Digitalis. Herbert2 remarks that Crinum Capense, Zeylanicum & scabrum are very similar in their general appearance & yet produce excessively sterile hybrids; whereas one of them namely C. Capense yields when crossed with C. pedunculatum fertile offspring; yet the latter is as unlike C. Capense "as perhaps any two species of any known genus" & indeed has been put by some authors into a distinct genus. So again Herbert3 asks how it comes that all the forms of Hippeastrum are excessively fertile together, whilst, in a closely allied genus Habranthus (or rather in the opinion of most botanists a mere section of the same genus) every attempt to cross the species has entirely failed. Numerous similar cases4 could be added. Although we may predicate that forms very remote in our systematic classification certainly will not unite; yet

1 Bastardz. s. 147.

2 Amaryllidaceae. p 343.

3 Hort. Journal. vol. 2. p. 10.

49V

Thus in same Family, the species of one genus, for instance Dianthus, will cross very easily & yield unusually fertile hybrids, whereas in another genus, as in Silene, the most persevering efforts of Kölreuter & Gaertner9 failed in producing hybrids between the very closely allied species.

49Vv

4 Lindley in his Theory of Horticulture p.336 [330], gives the currant & gooseberry, apple & pear, as cases of close species which will not unite. Herbert in Hort. Journal vol 2. p 82 insists on the impossibility of crossing the very closely allied species of crocus & iris [crinum], & contrasts this with the facility of crossing extremely unlike species of Narcissus. So again (ib. p. 95) in the genus Pelargonium, between certain species, there is a "secret insuperable bar"—

assuredly systematic affinity does not unlock the law regulating the fertility of the first cross & of the hybrid offspring. We shall, also, find this conclusion strongly corroborated when we come to compare reciprocal crosses.

Gaertner has shown5 that external differences even when strongly marked, in the form & size of the flower, fruit, seed, pistil, pollen & cotyledons, do not always prevent a union between two species. So it is with remarkable differences in general habit, for a shrubby & herbaceous calceolaria have crossed, as have two species of Erythrina, one of which bears its flowers directly from the root, & the other is almost arborescent & blossoms from the exits of its leafy branches6 . In the Cactae we have still more surprising unions with fertile offspring, as far as external shape is concerned; namely between the prickly & angular Cereus speciosissimus, to whip-like C. flagelliformi & the unarmed C. (Epiphyllum) phyllanthoides: several species, also, of Cereus have been crossed with Melocactus & Echinocactus.

Herbert believed that the difficulty in effecting a cross

5 Bastardz. s. 180, 183, 275.

6 Herbert. Amaryllidaceae p. 364, & Hort. Journal vol 2. p. 102.

depends on some constitutional difference between the species; if by this, differences not externally visible are only meant, it certainly is true. But if we take a more common acceptation of the word & imply such differences, as the duration of life, period of flowering, adaptation to climate, this view cannot be considered as generally holding good, though it apparently does so in some cases. Thus the tender Indian Rhododendron arboreum has been crossed with the hardy R. Ponticus, & even with R. Dauricum1 , which flourishes under the intensely cold climate of Eastern Siberia. In the crosses between Rhododendron & Azalea we see evergreen & deciduous bushes united. And Gaertner2 has shown that annual, biennial & perennial species can be united. A marsh & woodland species of Crinum, have produced a hybrid, as I was told by Herbert.

A remarkable discovery made by Kölreuter, shows I think even better than the above special cases, that neither systematic affinity, or resemblance in general habit or in constitution will account for the capacity of some species to unite & for the incapacity of others. I allude to reciprocal crosses. There are very many

1 Herbert Amaryll. p. 359.

2 Bastardz. s. 143.

cases, in which species A can be easily fertilised by the pollen of B, whilst B absolutely resists or receives with great difficulty the pollen of A. Thus Kölreuter found that Mirabilis jalappa fertilised by the pollen of M. longiflora produced a good many hybrids, which self-fertilised yielded seeds "numero non adeo exigeo" [sic]; whereas the reciprocal cross (i.e. M. longiflora fertilised by M. jalappa), was tried during fourteeen years more than 200 times, & yet utterly failed.3 I will give only two striking cases from Gaertner:4 Nicotiana Langsdorfii will fertilise four other species, but cannot be fertilised by them: the common & Canadian Columbines naturally yield nearly the same number of seed; but A. vulgaris fertilised by A. Canadensis gives us a maximum 151 seed, whereas the reciprocal cross yielded as a maximum only 29 seeds. In sea-weeds, Mr. Thuret5 has shown that Fucus serratus could quite easily be fertilised by F. vesiculosus, whereas he never once could effect, after repeated trials, the reciprocal cross.

3 Nova Acta Petrop. 1793, p. 391. Kölreuter gives other nearly as striking cases in Lycium & Linum, in Acta Acad. 1778. p 219 & Nova Acta 1783. p. 339.

4 Bastardz. s. 147, 195, 199: see his general remarks on this head s. 176.

5 Annal. des Scienc. Nat. 4 series Tom 2 & 3.

Gaertner found that this unequal fertility in reciprocal crosses was extremely common in all intermediate & lesser degrees: it could even be detected between species, very closely related to each other, as in Matthiola annua & glabra, Datura stramonium & tatula, which some botanists consider to be only varieties.1 It is also an important fact that the hybrids raised from reciprocal crosses themselves have not equal, occasionally even very unequal fertility2. These facts are the more remarkable as the hybrid off-spring of reciprocal crosses are generally so like as to [be] un-distinguishable in appearance;3

Now it seems impossible in these numerous cases to say that species A absolutely refuses to be fertilised, or is fertilised with great difficulty by the pollen of B, on account of any systematic or, in the common sense, constitutional difference, whilst the very same individual B can be easily fertilised by the pollen of this species or individual A. We are driven to look to

1 Gaertner Bastardz. s. 177, 197

2 Gaertner Bastardz. s. 407.

53 insert

; yet in their inner most nature they cannot be identically the same, for Gaertner4 found that reciprocal hybrids were reduced at different rates, when repeatedly crossed with one of the parent-species.

4 Bastardzeugung s. 459, 465.

54a

I may give one other statement by Gaertner in regard to the hybrids from reciprocal crosses, which it requires almost more than my faith in his accuracy to credit:—namely that a hybrid, even a very fertile hybrid, when crossed with a reciprocal hybrid, is quite sterile. Thus the hybrid Lychnis diurno-vespertina & the hybrid L. vespertino-diurna which differ from each other only in their reversed parentage are both very fairly though unequally fertile, but when these two hybrids are crossed, they produce no offspring. He gives five other instances, & appends an et cetera to his list; & concludes by saying "we have found these hybrids absolutely sterile ".1

1 Gaertner Bastardz. s. 515.

some difference in the sexual relations of the two species, which Gaertner calls their (Wahlverwandtschaft) elective, & which I have called their sexual affinity: by this expression we must include the relation of the pollen to the stigma, as well as to the ovule, & of the stigma to the pollen; & possibly even the relation of the hybrid embryo to the seed, formed by the pure mother-plant, as long as it is nourished by it.

If sterility had been ordained simply to prevent the confusion of specific forms, it might have been expected, I think, that there would have been an uniform relation between the difficulty in effecting the first cross between any two species, and the sterility of their hybrid offspring. Such relation does hold good to a large extent: if two plants can be crossed very easily, more especially if they can be reciprocally crossed very easily, their hybrid offspring are generally pretty fertile; & conversely if they cannot be easily crossed reciprocally the hybrids are generally very sterile. But there are strong & curious exceptions to all such rules.2 The hybrids, moreover, as we have just seen from the very same two species, when crossed reciprocally, often differ, even considerably in fertility. As a general rule hybrids when self-fertilised, & even when fertilised by the pollen of either pure parent species, yield far less seed than did their parent in the first cross by which they were formed:3 but some few hybrids, as from between the species of Datura & Dianthus yield more seed than does the first cross. On the other hand, there

2 Gaertner Bastardz. s. 200, 406, 407.

3 Gaertner Bastardz. s. 13, 425.

are many most striking cases of species which can be united with facility, whose offspring are excessively or even absolutely sterile: thus Nicotiana suaveolens fertilised by N. glutinosa yielded no less than 256 good seed in one fruit: but the hybrids raised from these seeds were absolutely sterile4 . So again the closely related species of Verbascum unite so easily that this not rarely happens without any artificial aid, yet the hybrids raised from these species are excessively sterile.5 Within the same genus Dianthus some species unite very easily, but produce hybrids most sterile; whereas other species can be united with the utmost difficulty, but produce hybrids very fairly fertile6.

Hence we must conclude that the fertility of the first cross & of the resultant hybrids, certainly in several cases follows widely different laws.

4 Gaertner Bastardz. s. 194, 405.

5 Kölreuter Dritte Fortsetz. s. 37, 42, 4: Gaertner Bastardz. s. 580, 591.

6 Kolreuter 2 Forts. s 108. 3 Forts. s. 108.

I will now give a few other of the best ascertained facts, to show on what special, curious & complex laws, the fertility of first crosses & of hybrids depends. Hybrids always yield more seed to the pollen of either parent-species, than to their own; & the pollen even of a third & quite distinct species is sometimes more effective than their own.1 During the reduction of a hybrid to either parent-form, which I may add requires more or fewer generations according to the species & even according to the individuals experimentised on, the fertility is extremely variable, but gradually increases (with a few exceptions) as the hybrids assume the character of either pure parent; yet it is sometimes seriously impaired, after the hybrid perfectly resembles the pure parent form. A hybrid reduced by the use of the pollen of the mother-species acquires fertility in the successive generations quicker, than when the same hybrid has been reduced by the pollen of the father-species.2 In all cases the male sexual organs suffers first; that is the pollen sooner suffers & during reduction is more slowly reperfected, than the capacity of the ovule for fertilisation.3

1 Gaertner Bastardz. s 425-427.

2 Gaertner Bastardz. s 419-455.

3 Gaertner Bastardz. s. 350, 355, 435. Kölreuter has made the same observation on the greater liability of the male organs than the female organs of hybrids to suffer.

Hybrids in the first generation generally all resemble each other, but occasionally single seedlings are produced differing considerably from the rest, & these are called by Gaertner "exceptional types"; they closely semble either the father or mother species, & are almost always quite sterile; even when the other hybrids from the same capsule have considerable fertility.4 These "exceptional types", also, sometimes appear in the succeeding generations of hybrids, whether fertilised by their own pollen or once by that of either pure parent; & as in the first generation, these exceptional types have diminished or even quite destroyed fertility.5

Again hybrids are usually nearly intermediate in appearance between their parents; but some species regularly produce what Gaertner calls "decided types", which take much more after one parent-species than the other. This is caused by the prepotency of one species over the other.

4 Gaertner Bastardz. s. 244. In some hybrids which I raised between the common Carnation & Spanish pink, one plant was extraordinarily like the pure Spanish pink, but it was not more fertile than the other quite sterile hybrids.

5 Gaertner Bastardz. s. 439, 442. Kölreuter gives, also, strong instances of this law.

These "decided types" are with some few exceptions sterile. This stands in close relation with the rule that when species cannot be easily crossed reciprocally their offspring are sterile; for it is found that those species which yielded "decided types" cannot be reciprocally united: thus if the hybrid offspring of A & B be nearly intermediate in appearance, then A can be fertilised by B, & B by A; but if the offspring takes decidedly after either parent, then a reciprocal cross between them can seldom be effected 1

In the foregoing cases, we see that close resemblance in hybrids to either pure parent, if it appears exceptionally in only a few of the hybrids, or if it appears in a very "decided" manner in all the hybrids, is connected with lessened fertility;—a fact which assuredly would never have been anticipated. It stands, moreover, in direct opposition to what usually takes place in the gradual & regular reduction of hybrids by the application of the pollen of either pure parent-species in each successive generation; for in this case as the hybrids gradually approach the pure parent-form, they acquire

1 Gaertner. Bastardz. s. 221, 286.

fertility. (We see, also, in the foregoing cases how little necessary relation there is between fertility & external resemblances).

Several cases are known of species which will not unite with each other, but will both unite with a third & distinct species.2

A species, when crossed with several other species of the genus, may have a very strong power of transmitting its likeness to all its hybrid offspring: a species may, also, have a remarkable power of fertilising the other species of the same genus; but these two powers are quite distinct & by no means necessarily go together.3

In nearly all cases in which three or four species have been united, Gaertner found the hybrids to be excessively variable & extremely sterile. But

2 Gaertner Bastardz. s. 202, gives examples in Nicotiana & Dianthus.

3 Gaertner Bastardz. s. 289.

61 & 62

this sterility in complex crosses is not invariable, as Herbert & others have shown4 in the genera Gladiolus, Crinum & Rhododendron even when four or more species have been united.

Taking a general review of the facts now given on the infertility both of first crosses & of hybrids,—we see a most insensible gradation from absolute sterility to high or perfect fertility, —we see the fertility not only eminently susceptible to external conditions, but independently of conditions innately variable in an extreme degree, so as sometimes to depend to a large extent merely on the individual selected;—we see that the infertility does not closely follow mere external, or systematic, or constitutional (in common sense) differences; we see this very plainly in reciprocal crosses in which there is a very general, & sometimes an enormous, difference in the result, solely owing to one of the two species having been used as father or mother; nor is this difference confined to the first cross, but affects the fertility of the hybrid offspring;—again we

62 insert

In some of the hybrid Rhododen-drons, raised by Messrs. Standish & Noble' no less than six species have been blended together by successive crosses into a single hybrid, namely Rhododendron campanulatum, maximum, Ponticum, purpureum, Catawbiense & arboreum,—species coming from the most distant quarters of the world, & having the most different climates; most of these complex Rhododendron hybrids seem to be very fairly fertile; & some of them extremely fertile, as has been previously mentioned.

see that the fertility of the first cross & of the hybrid offspring by no means always runs parallel;—we see several other curious facts, the pollen of the mother-species giving fertility during the reduction of a hybrid, sooner than that of the father-species;—the male sex failing easier than the female;—the extreme sterility of exceptional types i.e. of hybrids which suddenly assume the appearance of either pure parent, & of decided types or those which have not intermediate structure but regularly take after either parent;—& other such odd cases.—Now do these several laws & facts,—which it should be observed include all the known principal facts in hybridism, look as if they had been specially ordained for the simple purpose of keeping specific forms in nature distinct? I think that their complexity & singularity give a decided negative to this question. The several laws & facts seem to me to be incidental on other & unknown differences in the sexual organs & products of the two species which are crossed. And differences in the sexual organs & products

will stand in some relation, but by no means necessarily in a close & uniform relation, to systematic differences, which imply the sum of all the differences of all kinds.

I may illustrate what I mean by these laws of infertility being incidental on the sexual differences of species, by the action of an artificial poison, which, from what we know would certainly be in some degree different in widely different plants, but would probably be nearly alike on members of the same genus; & this action might be called incidental, for as the poison did not exist naturally the species would not have been created, or modified by natural selection, so as to have different powers of resisting it. The action moreover, would be incidental in differences in the absorbent & nutritive systems, wholly inappreciable by us. But another illustration already alluded is so apposite, that it is worth giving in rather fuller detail: I refer to grafting & budding.

The capacity for grafting, <like that for hybridism,> is limited but less closely than of hybridism, by systematic affinity. The ash, Lilac olive, Phyllirea, Chionanthus & Fontanesia, belong to the same order, & though so very different in appearance, can be grafted together; but the Jasmine, as Von Martius has remarked, cannot be grafted on them, "which confirms the propriety of separating these two orders"1 . Many species of Pyrus can be grafted together; & the Pear can be grafted on three other genera, namely the white-thorn (Crataegus) Quince & Cotoneaster; but it would be an error to suppose that all the species in the same Family can be grafted together, or grafted with equal facility; the apple though so closely allied to the pear takes on it with much difficulty.

Although a small bush can sometimes be grafted on a tree, as Cytisus purpureus on C. alpinus; although plants of the most widely different external shape can be grafted together as various Cacteae; though a plant from a warm climate will take on one from a cold, as with species of Rhododendron,—a shrubby plant on an herbaceous one, as the tree Paeony on the common, a deciduous tree on an evergreen, as the common cherry which

1 Lindley The Vegetable Kingdom 3 Edit p. 616. I have myself had the Lilac grafted on the common Ash, & reciprocally the ash on Lilac.

will even fruit on the Portugal laurel; yet it has been observed2 that trees of very different size, or which grow at very different rates or have wood of very different hardness or saps of different nature or flowing at different periods, cannot be grafted together or take with much difficulty. But who can explain why the pear takes so infinitely easier on the quince, classed in a distinct genus; than on the apple a species of the same genus; or why the pear takes easier on the apple, than the apple on the pear;3

By these latter facts we are reminded of the unequal reciprocity of Hybrids.

Thouin names1 three species of Robinia which when grafted on other species are generally quite barren or yield but very few seeds whilst ungrafted they often seed pretty copiously; on the other hand some species of Sorbus when grafted on

2 N. C. Seringe Flore des Jardins p. 250: Loudon in Encyclop of Gardening p. 650 asserts that the Acer platanoides on account of its milky sap will not receive other maples.

3 Loudon's Gardener's Mag. vol. 1, p. 200. Diet has made the same observation as quoted by Gaertner, Bastardz. s. 632. For the Currant & Gooseberry case, see [Godsall,] Gardener's Chronicle 1857. p. 757.

1 Annales du Museum. Tom XVI. p. 214.

66 insert

; or again why the common Gooseberry cannot after repeated trials be grafted on the currant, though so closely allied systematically & apparently in constitution; whereas, on the other hand, the currant will take on the gooseberry.

distinct species yield twice as much fruit, as do seedlings on their own root. We can, I apprehend no more account for this latter fact, than for some of Herbert's crosses yielding more seed than the same species naturally self-fertilised; nor can we account for the barrenness of the Robinias any more than for the barrenness of many hybrids.

Although we may probably account for Peaches succeeding best on plum stock by the hardiness of the plum's roots in our climate; & for the pear often succeeding best on the quince by its over luxuriance being checked; & for certain apples flourishing best when grafted on the Paradise variety, owing to the fibrous root of the latter not penetrating so deeply into poor soil; and for late & early Varieties of the Walnut2 not taking kindly on each other owing to their sap flowing at different periods. Yet who can explain why one variety of the Pear succeeds far better than another variety on the Quince; & why, as it is positively asserted3 some varieties will not succeed at all on the quince.

2 Loudon Encyclop. of Gardening p. 650.

3 Loudon's Gardener's Mag vol. 3. p. 380. All horticulturists have remarked on some degree of diversity in the different varieties of the Pear. see Sagerets Pom. Phys. p. 65 p. 222. A. Knight in Hort. Transact. vol 2. p. 203.

Why do certain varieties of the Apricot, & of the Peach, prefer certain varieties of the plum?4 Why will not the Golden variety of the common Lime take on its own species, but freely on the distinct American species?5 Sageret, moreover, gives reasons for believing that individual stocks have a repugnance to receive the grafts of certain varieties,6 in the same way as we have seen the individual plants resist being hybridised: and the rare cases of varieties crossing with different degrees of facility with each other & with distinct species, have been paralleled by the foregoing not rare case of varieties grafting with different degrees of facility.

In drawing this parallel I am very far from wishing to make it appear that grafting & crossing are allied processes: many species will graft together with the utmost facility which cannot be crossed: the mere cohesion of the cells in grafting, & the intimate fusion of the two cells in sexual union must be fundamentally different.1 But I think facts enough have been given to show that the capacity both for grafting & for crossing is limited, but by no means wholly,

4 Sageret Porn Phy. p. 321, 346.

5 Loudon's Gardener's Mag. vol. 6 p. 317.

6 Pom. Phys. p. 222.

1 Gaertner in his Bastardz. s. 606-633 has an excellent discussion on this subject.

governed by systematic affinity; & that in many cases we can assign no cause why certain forms will not graft or will not cross. We have seen, also, that varieties are not exempted from differences in their capacity for grafting & crossing. I presume that no one would say that the capacity for grafting with its somewhat complex & obscure laws, is a specially created endowment; I presume that all will admit the capacity is incidental on differences in the constitution, more especially of the vegetative tissues, in different species. So do I conclude that the capacity for crossing is incidental on differences, more especially in the sexual relation (taken in its largest sense) of the species subjected to experiment.

I may add that in the Robinias & in some other cases which might have been added, the barreness is incidental on the grafting, as the power of being grafted is incidental on other contingencies. The lessened fertility, also, from close interbreeding must be looked at as incidental on wholly unknown laws, for it does not seem probable in the case of plants & the lower animals that this should be a special endowment. So again the very frequent sterility from changed

conditions of existence, which was so fully treated in the third chapter I should look at as only incidental; for this tendency to sterility could not have been acquired by natural selection; it may well be doubted whether it would have been created, as its only use would be to keep organic beings within certain limits & this would apparently be superfluous as climate & the struggle for life would be amply sufficient. <Nevertheless there are a few cases known of plants extending sometimes by the accidental dispersal of the seeds, into conditions where they are rendered sterile; but more cases could be given> Moreover several cases could be given of plants living in profusion, where they do not increase <propagate> by seminal reproduction.

Causes of the sterility which is incidental on Hybridism.—Very little light can be thrown on this subject. The following remarks will apply to animals as well as plants. There is clearly a fundamental difference between the sterility of Hybrids & that of the first cross between two pure species. In hybrids the sexual functions <elements> are deteriorated, as can be plainly seen at least in the pollen & spermatozoa. In first crosses between two pure species both sexual elements

are of course perfect; but either the pollen never reaches or does not penetrate the ovule, or reaching it does not cause an embryo to be developed; or an embryo is developed but perishes at an early age. When the pollen of a plant is placed on the stigma of a distantly allied genus, both the pollen & stigma are often, as before mentioned in some degree affected by their affinity, but the pollen-tubes do not properly penetrate the stigmatic tissues. Here we have the first of the three apparent causes of sterility. Thuret1 in attempting to cross distinct genera of Fucus, saw the antherozoids cling to the naked spores, but no germination ensued: here we have the second cause. But in some very rare cases Thuret observed the commencement of germination, the spore subsequently perishing;2 & this is our third cause.

The extreme difference of fertility sometimes observed, & in a lesser degree often observed, in reciprocal crosses amongst plants especially makes us feel how ignorant we are on the whole subject: are we to suppose that, though A can be readily fertilised by B, the pollen of A cannot reach the ovule of B; or reaching it that it does not cause an embryo to be developed or being developed that the embryo perishes at an early age? These points have

2 See Gaertner's account (Bastardzeugung s. 101) of his " Fructificatio subcompleta," in which only a small withered embryo is developed: when seeds of this character very rarely germinate, the seedlings are weak & soon perish.—

71v

That the early death of the embryo is in some cases is one very potent cause of the little fertility between two species when crossed, I cannot doubt from some facts communicated to me by Mr. Hewitt, who has had the largest experience during many years in making hybrid Gallinaceae. Mr. Hewitt has had in one year above 300 eggs from crosses between various pheasants & the common cock pheasant & Fowl; & he assures me that he has "opened hundreds of eggs, containing partially formed embryos of hybrid pheasant-fowls"; yet these two birds are so sterile together that only 3 or 4 per cent of the eggs produce chickens. Again out of 55 eggs from the hen Silver Pheasant, fertilised by the Gold Cock pheasant, he got only three hybrids, but on opening the bad eggs, he found "that many had germinated."—

71 insert

See Gaertner's account (Bastardzeugung s. 101) of his " Fructificatio subcompleta," in which only a small withered embryo is developed: when seeds of this character very rarely germinate, the seedlings are weak & soon perish.—

not been investigated as far as they might have been by microscopical dissection. Considering the double & unnatural condition of hybrid embryos, & that they have to be nourished within a seed or egg formed by one of the pure parents & therefore of a somewhat different nature to their own, & considering how easily the young both of plants & animals are affected by unfavourable conditions, & lastly considering that hybrid seedlings & young animals raised from between very distinct species are often tender & delicate, it does not seem improbable that in many cases the cause of infertility in first crosses lies in the early death of the embryo. But opposed to this notion, is the general health & vigour of hybrids when once produced. When two plants very remotely allied are crossed the probable cause of their infertility seems to be that the pollen does not reach the ovule; between plants more closely related, that the pollen does not cause an embryo to be developed, & perhaps in not a few cases that the embryo is developed & then perishes. But why the pollen-tube should not penetrate the stigma of a remotely allied plant, or why if penetrating it should not develop an embryo, is no more explicable, & apparently is no greater a difficulty, than why some trees can & some cannot be grafted on others.

72bis

With regard to the sterility of hybrids themselves which are imperfect in their sexual functions,—a fact which in itself has always seemed to me much more remarkable than the difficulty in effecting a first cross—I think we can perhaps see our way a little more clearly. In our third chapter, numerous facts were given showing that slight changes of condition & crosses between closely allied forms or varieties were good for all organic beings & their offspring; but that changes of conditions beyond certain limits affected the reproductive system in an especial & injurious

manner, independently of general health: therefore it completes the parallel that crosses beyond certain limits of affinity should injuriously affect the reproductive system. In this same chapter I showed in detail how similar in many respects the sterility superinduced by unnatural conditions was to that caused by hybridism; I will here recapitulate only the more important points. In both cases the sterility is often quite independent of general health; how healthy & how sterile is the common mule! In both cases the sterility occurs in various degrees: in both the male element is the most liable to be affected, but sometimes the female more than the male.1 In both, the tendency goes to a certain extent with systematic affinity; for whole groups of animals & plants are either greatly or very little affected by unnatural conditions of the same kind, without our being able to assign any adequate reason; & whole groups of species tend to produce sterile hybrids; but there are often marked exception in both cases in the same groups. No one can tell till he tries, whether any particular animal will breed under confinement, or any plant seed freely under cultivation; nor can be tell till he tries whether any two species of the same genus will produce more or less sterile hybrids. Lastly when organisms are placed during several generations under conditions not natural to them, they are extremely liable to vary, which

1 Gaertner Bastardz. s. 357. 360.

is due, as I believe, to their reproductive system having been thus specially affected though in, a lesser degree than when entire sterility is caused: so with hybrids, when they can breed, their successive generations are eminently liable to vary, as every experimentiser has observed.

Seeing how similar the results are in these two apparently very different cases, let us compare the secondary causes: in the one case the structure & constitution of the organism remain the same, but the conditions of life to which it is exposed have been changed, & hence results sterility: in the case of a hybrid, the conditions of its existence may remain unchanged, but its constitution & all the laws of its growth from its earliest days, from being compounded of two distinct forms, can hardly fail to have suffered disturbance, whatever may be the conditions of life to which it may be exposed; & hence sterility is the result. There can be but very few species in nature with their whole constitution & laws of growth so similar that the blending of the two would not cause a disturbance, different from, but we may suppose as great, for instance, as the giving a plant a little too much water during one season of the year, which we know will in some cases not in the least effect its general

health or prevent its flowering, but will render its pollen utterly impotent.

As the double & compounded nature of a hybrid is inherited by its offspring, it is not surprising that the infertility (subject, however, like the whole rest of the organisation to variation) should be likewise inherited: the gradual increase of infertility which has not rarely been observed in the successive generations of hybrids, I am strongly tempted to explain in large part for reasons already assigned, to the evil of close interbreeding; & we know that their infertility is highly susceptible to unfavourable conditions.

During the gradual reduction of a hybrid by successive crossings with either pure parent, the stain of the mixed constitution <foreign blood> is gradually washed out, & fertility is acquired. But no light as far as I can see, can be thrown on the very singular fact that amongst hybrids, "exceptional types" (or those which suddenly & abnormally closely resemble either parent-form) & "decided types" (or those which normally closely resemble either parent-form) are almost always extremely sterile; without, indeed, we might suppose that whilst the large part of the organisation took closely after the other, then I think we could understand how there would be a greater disturbance in the machinery of life, than if every part of the machine had a more strictly intermediate structure.

It will have been seen that I would explain somewhat differently the two cases of the sterility of hybrids themselves & of the difficulty in effecting first crosses;—namely in the case of hybrids by their double & heterogeneous nature producing closely analogous results to what changed conditions do when acting on pure species; & in the case of first crosses, either by obstacles perhaps of various kinds to the act of fertilisation—(somewhat analogous to those in making grafts) or sometimes by the early death of the embryo. Although the cause of sterility in first crosses & in hybrids seem to be, almost necessarily, somewhat different, I do not think that it is surprising that there should be a considerable degree of parallelism in the results; for in both cases the sterility is related to the amount of difference between the parent-forms. Moreover we should bear in mind that even within the limits of the same genus, the parallelism is by no means always close between the number of hybrid offspring produced by a first cross, & the fertility of the hybrids when obtained. How many cases there are, as with the common mule, in which there is no great difficulty in producing the hybrid, the hybrid itself being excessively sterile; & in plants there are a good many cases, exactly the converse. Even between the very same two species, when crossed reciprocally, we have seen that there is sometimes the utmost difference in the number of offspring obtainable, according as one or other is used as father or mother; whereas in the hybrids themselves obtained from such reciprocal crosses there is only occasionally

a considerable <slight> difference in fertility. <There is another important point of difference between the hybrid offspring of first crosses & the offspring from a first cross (taking of course species which have not been affected by culture) are generally, as we shall presently see, very uniform in character, for their parents have not been exposed to conditions tending to make them sterile & variable, whereas the offspring from hybrids which are newly [?] domesticated [?] in their successive generations, are eminently variable, like the offspring of species which have suffered during successive generations from conditions not natural to them.>

Finally we have seen that sterility occasionally ensues when two species are grafted; in a lesser degree from the close interbreeding of individuals of the same species; in a marked manner from exposing organic beings to conditions of life different from those to which nature had adapted them; & lastly from crossing species, I believe, as soon as we can explain why unnatural conditions make a pure species sterile without necessarily affecting its health, then & not till then, we shall understand the sterility of hybrids;—but that something different (& different in different cases) is required to explain why few or no progeny is produced from a first cross between two pure species. When our physiological knowledge is so perfect that we can explain why trees of different orders can never be grafted on each other, we may perhaps hope to explain the more difficult problem why the pollen of a given species will not penetrate or penetrating will not fertilise the ovule of another species.

Ch. 9 Animals, sterility of hybrids &c

Animals

In regard to the sterility of animals when first crossed & of their hybrid offspring, I shall discuss only a few points, & chiefly in comparison with plants. Carefully conducted experiments have seldom been made; & we have not here excellent Treatises1 like those on hybrid plants.

With animals the will & instinct come into play preventing or checking first crosses; but their importance has, I think, often been greatly exaggerated. No doubt some cases, as in the experiments recorded by Buffon & Hunter1 two species have shown a strong aversion to cross; but then, quite independently of any sexual relations, distinct kinds of animals often dislike each other.

If we lay on one side some few tribes of animals, as the Ruminants, which breed very. readily under confinement, it is hardly an exaggeration to say that in menageries hybrids are produced almost as easily as pure-bred species: let anyone look over the Reports of the Zoological Soc. for a number of years, in which both pure breed & cross-bred births are recorded, & he will see that this is true. How rarely do the Fringillidae breed in confinement,—yet at least nine species, belonging to three distinct genera have produced hybrids with the Canary: no species has been crossed oftener with the canary than the

78V[1]

1 In the several compiled lists, which have been published, little care seems to have been taken in sifting evidence. Several cases are recorded, as due to hybridism, which I can hardly doubt have been simply monstrosities. The statement by Hellenius that he crossed a common ram & female deer (Cervus capreolus) quite recently has been admitted; not withstanding that he found the hybrids perfectly fertile inter se! But I believe there is no doubt that the mistake arose from the Mouflon the reputed aboriginal of the sheep) being called a Roe in Sardinia; see Observations on Zoology by C. L. Bonaparte. [Wagner] in Ray Soc. 1841-42 p. 85. Again M. E. de Selys-Longchamps says (Bull. Acad. Royale de Bruxelles Tom XII [see pp. 341-3]) that the Baron de La Fresnaye procured seven hybrids from Anser cygnoides & Anser Canadensis : one coupled with canadensis & two others with other species all were sterile. Is not this the foundation of M. Chevreul's statement (Annal des Sc. Nat. 3 series Bot. 1846. Tom 6. p. 188.) that the Baron produced hybrids from these two species & that "il est remarquable que leurs hybrides se soient reproduits deja jusqu' a sept fois."? Morton, changes the A. canadensis into A. cinereus & says that Chevreul himself has seen "the progeny extend through seven generations".

78V[2]v

Animal Oeconomy p. 310. The she-wolf had to be held. But I have known the same thing to be necessary in choice fancy spaniels, which had long been closely interbred; the female having thus lost sexual passion.

78V[3]v

In one of the cases of aversion given namely, wolf & dogs other experimentisers have not observed any aversion,—& it is perfectly well known that a tame dog from Parry's ship coupled with a wild she-wolf.2

78V[4]v

2 It is scarcely possible to read M. Mauduyt little Treatise Du Loup et de ses Races 1851 without believing that in the Pyrenees crosses between free wolves & Dogs not very rarely take place.—See Supplement to Parry's Voyage 1819-p. CLXXXV: for additional case see Franklins narrative vol V. p. 140. [? See Appendix 5, p. 664] See also Pallas. on this subject.

Siskin (Fringilla spinus); yet instances of siskins breeding in confinement are extremely rare.3 I could give many cases of birds kept with others of the same species almost in a state of nature,4 yet pairing with distinct species. Under strict confinement this is still commoner: thus a female Bonnet monkey in the Zoological Gardens, I was often assured by the Keepers preferred the male of any other species to her own, & she produced a hybrid with the Rhesius monkey.—

The well authenticated cases, also, of hybrids bred between species, both in a state of nature, show that there cannot have been any strong aversion between them; though I do not doubt that generally these crosses have been caused by an inequality in the sexes of the pure species. Thus five species of wild Grouse have produced hybrids together; about 18 cases in Great Britain alone are now on record of hybrids produced between the Black hen grouse (Tetrao tetrix) & the cock pheasant & reciprocally.1 Several cases are known amongst Ducks; & distinct species of insects have often been caught in union.2 With rare exceptions, all that can be said with truth is that the sexes of distinct species, when very remote do not in the least excite (though even here

3 Mr. Hewitt (Poultry Book by Tegetmeier. 1857 p. 123) says that after a domesticated Cock Pheasant, has become attached to a Hen of the common Fowl, the intro duction of a female pheasant "will estrange all feelings of affection, which had before-times been indulged."

1 [Spicer] Zoologist vol 11-12—1853-1854. p. 3946 [4294].

79V[1]v

3 Mr. Milne says he has never known or heard of more than one instance: Loudon's Mag. of Nat. Hist. vol. 3. p. 440. See my Chapter 3 on the difficulty in making the Fringillidae breed in confinement.—

79V[2]v

4 Waterton says. Essays on Nat. History 2 series p. 42, 117. that a Canada Goose, living with 23 other birds of the same species, paired with a solitary Bernacle Gander, though of so different a size, & produced young. A wigeon (Mareca penelope) associated with others of same species, paired with Pintail (Dafila acuta). Loudons Mag. of Nat. History. vol. 9. p. p 616.—I have heard of many similar cases. Indeed the Rev. E. S. Dixon (Ornamental & Domestic Poultry p. 137)

79V[3]

2 A Gold Pheasant turned out & free in the woods at Henley Park produced hybrids with the common Pheasant; [Lowcock,] Annals & Mag. of Nat. History vol 6. 1841. p 73.—Numerous cases are on record of the Carrion & Royston crow pairing, see ch. 4.—Several hybrid Ducks have been shot at different times. see [Fennell] Loudon. Mag. of Nat. History. vol. 9. p. 616. See Macgillivray British Birds vol 1. p 398 on the breeding of Goldfinch & Green linnet in nature.—I can see no reason to doubt that the Black bird & Thrush have produced hybrids together: see [Berry] Loudons Mag. of Nat. History. vol 7. p. 599: this fact is corroborated by another case in Macgillivrays British Birds vol 2. p. 92. It would appear from Bechstein (Naturgesch. Deutschlands B t s. 950) that hybrids have been naturally produced from the Black & Brown Rat. For insects see Bronn's Ges[ch]ichte der Natur B. 2. p. 164. [Shuckard,] Annals of Nat. Hist. vol 7. 1841. p 526.; Westwood Transact. Entomolog. Soc. vol. 3 p. 195.

some strange anomalies have been recorded) each other passions; & that when more closely related, they excite each other, but in a less degree than natural;3 but not apparently that they cause more natural aversion than do the same sexes of the two species when confined closely together.—

With respect to the degree of fertility of first crosses in comparison with the fertility of the hybrid offspring from such crosses, few facts seem accurately known; & indeed from several causes can be made out only with much difficulty, more especially with those animals which produce only one or two young at a birth. The best evidence known to me refers to the common mule; in rearing which it has been found that only [ ] conceptions follow from 100 unions; whereas with the mare & horse [ ] per cent of the union are fruitful: Azara, moreover, states that the mare ceases to produce at an earlier age to the male ass than to the horse1

80Vv

In crossing Gold or silver pheasants with the common Pheasant "most of the eggs prove barren", & the chickens are difficult to rear, & when reared are almost invariably sterile2

On the other hand, hybrids from the common Duck & the Musk-Duck. (Anas boschas and Cairina moschata) are utterly sterile & even without any passion; but yet can be raised with great facility, & are raised in large numbers in the U. States for the table, as I was informed by Dr. Bachman; so that, as I infer, most of the eggs must be fertile

[a: below inserts here] in the same laying, however, if a single egg proves to have been fertilised, several can generally be hatched.4

Mr. Brent, tells me that in his crosses between the canary-bird, gold-finch, Linnet & green-linnet (Loxia chloris) he has often had the full number of eggs & every egg fertile: the hybrids from these birds will often breed with either pure parent, & some of them very rarely inter se. Mr. Brent, also, in crossing the Stock & common Pigeons (C. oenas & livia) found both eggs fertile, but the young were very difficult to rear, & (with him were) sterile. A Pomeranian bitch with a dog-wolf produced ten puppies.5

2 Mr. Hewitt: Poultry Chronicle 1855. vol. 3 p. 15 & Temminck Hist. Nat. Gen. des Gallinacees vol. 2. p. 828.

5 Pennants Quadrupeds 3 Edit. vol. 1, 1793 p 238.

81v

(a) text

a: but even in this case Mr. Garnett of Clitheroe who has raised many of these hybrids & from reciprocal crosses, informs me that the proportion of good eggs is not so great as with the Common Duck.—

Mr. Hewitt informs me that out of above 800 eggs from these two birds, he raised not above a dozen hybrid chickens:

81Vv

The hybrids from the Cock Pheasant & Common Hen are I believe universally quite sterile & do not even show any sexual passion; & there is considerable difficulty in producing them, for Temminck3 asserts that out [of] 100 eggs, only two or three young can be raised: Mr. Hewitt informs me that out of above 800 eggs from these two birds, he raised not above a dozen hybrid chickens: in the same laying, however, if a single egg proves to have been fertilised, several can generally be hatched.4

3 Hist. Nat. Gen. des Gallinacées Tom 2. p. 314.

4 Mr. Hewitt in the Poultry Book p. 125.

From these few facts I presume that there is, as with plants, some pretty close relation between the facility of getting offspring from first crosses & the fertility of the hybrids when raised; but I much doubt, more especially from the case of common mule & musk duck, whether the relation is uniform. The fertility of the first cross, when utterly barren hybrids are produced, does not seem so much impaired as with plants. <The frequency [of] several cases of very young hybrids being difficult to rear & of the first-laid eggs being addled perhaps indicates that the fewness of the progeny is in <largegt; part due to the deaths of the embryos at an early age.—>

With animals, it is difficult to decide whether in first crosses as with plants there is much or any unequal reciprocity, for here instinct comes into play. The reason why male Fringillidae alone are generally paired with female Canary-birds, is that hens of wild species, if not taken quite young from the nest, will seldom receive a male of any kind; & they will not build a nest, or use one when made for them.1 The greater facility of getting mules than hinnies I have heard attributed to a difference in a sexual instinct between the male ass & horses; & this perhaps accounts for the much greater frequency of crosses between the domestic dog & she wolf, than reciprocally.

But Bechstein who was well aware of the foregoing causes of difficulty, states that the male

82Vv

It is said further that though the he-goat crosses readily with the sheep; but <yet> that the Ram will not produce with the she-goat.2 Mr. Fink has found the reverse to hold good; namely that a Ram crosses more readily with she-goat than reciprocally.3 Mr. Garnett of Clitheroe informs me that the common Drake will seldom have any intercourse with the Musk-Duck, whereas, as is well known, the Musk-drake takes with perfect readiness to the common Duck; but when they are paired, Mr. Garnett tells me that he has not observed any difference in the number of young produced.

House sparrow succeeds better with the female treesparrow (Fringilla domestica & montana) than reciprocally.4 I could add other facts pointing to the same direction, but many more facts are wanted to draw any definite conclusion on reciprocal crosses amongst animals. <Hybrids themselves are only so far imperfect, as far as the most careful examination shows, that the spermatozoa in Mammals & Birds are in the same state as in pure species in the intervals of rut.5 > Gartener6 advances some evidence, <but hardly sufficient in my opinion,> showing that in hybrid animals; as with hybrid plants, the male sex fails easier than the female: I shall presently give a very striking case of this in hybrid Yaks. I do not know whether it is anyways connected with this tendency that male hybrids, <as remarked by Buffon & others & I believe truly,> are oftener produced than females.1

The fertility of first crosses & of hybrids seem to be much affected by favourable conditions of all kinds; the common mule is said3 to conceive more frequently in hot countries4 : I strongly suspect that the hybrids from Anser cygnoides & cinereus, are more fertile, as we shall presently see, in India than in Europe; the only known case

4 Stubenvögel s. 210, 224.

5 Annales des Sci. Nat. [ ] Gaertner Bastardz s. 340, 382.)

6 Bastardz. s. 340, 382.

3 Gaertner Bastardz. s. 381 & Morton on Hybrid Animals in Edin. New Phil. Journal vol 43. p. 264.

83v

1 Flourens de la Longevite Humaine 1855. p. 156. [Darwin added in pencil (one wonders when): 'See my paper on Species of Primula for additional facts.' See Darwin, 'On the Character and Hybrid-like Nature of the Offspring from the Illegitimate Unions of Dimorphic and Trimorphic Plants.' Linn. Soc. London. J. Rot., 10 (1869), 433-4.]

4 But an apparently well authenticated case of the mule breeding in Scotland is given in Smellies Edit, of Buffon 3 Edit. vol. 8 p 18.—[in note by Wm. Smellie.]

83Vv

BB text

A perfect gradation in the degree of sterility of first crosses & of the hybrids themselves could be given; from unions, which as between the Guanaco & Goat2 never produce offspring, to cases where a single instance is known of offspring produced after repeated couplings as in case of Peacock & Guinea-fowl, to such cases as the common mule, where they are habitually produced; & so with hybrids themselves, as we shall see in the cases presently to be given in more detail.

of these hybrids breeding inter se in Europe was effected by Mr. Eyton5 taking a male & female hybrid from different hatches, & thus in a slight degree lessening the ill-effect of the closest interbreeding. Age, either very slightly too great or too little, interferes most seriously with the fruitfulness of first crosses: we have seen that according to Azara mares fail to produce to the male-ass, earlier than to the horses. Mr. Brent informs me that it is an axiom with Canary fanciers never to put a hen bird over four years old to a male Gold-finch, as there would be no hope of produce. On the other hand more hybrids can be raised from between pheasants & common Hens, in their second year than in their first:6

So again a Canada goose crossed by a Bernicle gander7 for the two first years laid barren eggs, but in the third two young were hatched out of seven eggs. In the case of the hybrids themselves from the Canary & other finches, in the few instances in which they have bred the eggs produced the first year have been observed to be sometimes either very small or the young birds to be very weakly; but in the following years stronger hybrids have been produced.1

As with plants, a hybrid can be reduced or absorbed by successive crossing to the form of either pure parent form, of which I shall give a good instance under Phasianus versicolor.2 I have seen a triple cross amongst quadrupeds, namely a hybrid from a mare by a hybrid ass-zebra. Mr. Gould tells me that Phasianus, versicolor, torquatus & colchinus have blended together in the woods of Norfolk.

5 Charlesworth's Mag. of Nat. Nistory 1840. vol 4. p. 90.

6 Bechstein Naturgesch Deutschlands B. 3. s. 434.

7 Waterton's Essays on Nat. History 2 series p. 42, 117.

1 Bechstein Stubenvogel 4 Edit. s. 248. Mr. Brent informs me that he has had these small eggs from a hybrid canary-goldfinch, fertilised by pure Goldfinch.

84V[1]v

2 The hybrid Ph. versicolor & colchicus was reduced to the pure form of P. versicolor by two successive crosses with P. versicolor: i e P. colchicus was absorbed in three generations. Mr. Flourens, shows (Longévité Humaine 1855. p. 145.) that in four generations the Jackall was reduced into the Dog & the hybrid Yak seems to be reduced in three generations. From these facts animals would appear to be reduced by crossing with each other at a rather quicker rate than plants.

84V[2]v

QQ text

Mr. Hewitt tells me that eggs from these two birds, laid later than April & May, invariably failed to produce chickens.

Systematic affinity, though limiting the possibility of hybrids being produced, certainly does not, any more than with plants, absolutely regulate this capacity. But it is very difficult to draw any just conclusion with animals on this head; for under confinement so many pure species either will not unite, or uniting are quite sterile. How, for instance, can we compare the capacity for hybridising between the several Families of Carnivora, whilst the plantigrades, though freely coupling, so rarely breed in confinement3 ; or how can we compare hawks & gallinaceous birds, whilst the former have never been known to breed when tamed. We are almost driven to look to animals in a state of nature in order to judge of their tendency or capacity to produce hybrids; but then in a state of nature we can seldom form any opinion, on the degree of sterility of their hybrid offspring. By the foregoing remark I do not wish to doubt the common opinion that the Gallinaceae are eminently capable of hybridisation;4 I believe this to be case; but how much of this capacity to attribute to several species having been domesticated, & to most of the species breeding readily under confinement, & to the social habits of many, I know not. I strongly suspect that the great Pigeon Family, though several have crossed, are

3 See chapter 3. p. 76 et seq.

85Vv

4 The Anatidae have almost or quite equal capacity; Swans, Geese, Ducks of various subgenera have crossed very freely: see Selys-Longchamps, in Acad. Roy. de Bruxelles. Bull. Tom. XII. no. 10.—Nineteen cross were enumerated by M. Bartlett before the Zoolog. Soc. 1847. April 13. Amongst the Fringillidae Bechstein (Naturgesch. Deutsch. B. 4, s 468) enumerates nine species belonging to genera Fringilla (with its sub-genera) Loxia & Emberiza which have yielded hybrids together. These several Families include, with the exception of Pigeons, nearly all the Birds which have been domesticated.—

much less capable of hybridisation than the Gallinaceae. In a state of nature the grouse-genus (Tetrao) seems particularly inclined to cross; more especially Black-Game (T. tetrix) which has crossed with the pheasant, Fowl Capercailzie, Red Grouse and willow grouse, as may be seen in the following table; but we have no reason to suppose that the hybrids are in any degree fertile.

The pintail Duck (Dafila acuta) seems to have strong tendency to cross with the Wigeon & the common Duck,1 though these three Ducks are placed by modern systematics in distinct genera; & the hybrids of the Pintail & common duck are fertile, in an unusual degree; the hybrids again from Anser cygnoides & cinereus are far more fertile than hybrids between other & apparently much more closely related geese. Many cases could be given of very sterile hybrids produced by the crossing of species apparently very closely related; as between the common & collared Turtledoves; between Gallus bankiva & Stanleyi; between the common & changeless <Polish> swans2

2 Mr. Yarrell informed [me] that a hybrid from these two closely related swans was quite sterile at Lord Derbys: I shall refer presently to the sterile hybrids from the two very closely related species of Gallus: for the evidence about the Turtle-doves, see Ch. 2. of this work.

86Vv

1 On the fertility of Hybrids of Pintail & Common Ducks see Proceedings. Zoolog Soc. 1831. p. 158: and [Fennell] Loudon's Mag. of Nat. History vol 9. p. 616. I have heard of other instances & statements that the hybrids have bred inter se. At the Zoological Gardens I saw hybrids, which I was informed by the Keeper were descended from half-bred pintail by a Duck & then by a Pintail; so that these hybrids had 10/16 of pintail in them & 6/16 of common Duck.

87a

The Families & genera are arranged in accordance with Mr. G. R. Gray's classification generally acknowledged as one of the best. The Brackets, imply that hybrid offspring has been produced by the two forms so connected. The degree of fertility of the hybrids, is given in the notes, where nothing is said, nothing is known & in this case generally, their sterility may be safely inferred.

As far as our present imperfect state of knowledge serves, I conclude with Gaertner that the laws regulating the fertility of first crosses & of hybrids amongst animals are nearly the same as with plants; but my impression is (notwithstanding the obstacle sometimes opposed by the will & instinct of animals) that first crosses are more easily effected between animals considerably remote in the scale of nature, than between equally remote plants: of course nothing can be more vague than the comparison of the differences in plants & animals; but if we assume (though it would be a very bold assumption) that the genera of birds, for instance amongst the Rasores, are of at all equal value with the genera in any order of plants, then I think my remark holds good. On the other hand, I believe that amongst animals, the hybrids themselves are generally more sterile than with plants.We shall see some proof of this conclusion by comparing in the following table the very wide limits within which amongst the Rasores, first crosses have been effected, with the very general sterility of the hybrids I have given the following Table of all the well authenticated crosses which I have heard of in one order of Birds, the Rasores; in order that those who have not attended to the subject, may see how numerous the crosses have been, & between what different forms.

87Vv

We shall see some proof of this conclusion by comparing in the following table the very wide limits within which amongst the Rasores, first crosses have been effected, with the very general sterility of the hybrids produced. This latter fact is, also, strongly brought out in the few cases, immediately to be given of the highest degree of fertility observed in any hybrids. This conclusion, moreover, seems quite conformable with the analogy, which I have drawn between the sterility of hybrids themselves (not of first crosses) & that in pure species from changed conditions of life: for the complex organization of an animal might well be more disturbed by being blended with that of another species, than in the case of a plant; & we have seen that very many animals if confined in their own country are rendered sterile, where with plants this is rare without they suffer from some very marked change of conditions.

Order Rasores, Fam. Cracidae Sub. Fam. Craeinae Crax rubra1 —globicera—alector

Fam. Phasianideae Sub. Fam. Pavoninac

Pavo muticus2 Japan Peacock—cristatus3 Peacock

Sub. Fam. Phasianidae

[remainder not transcribed]

88a

7 Hybrids from the male of this Bird & the common pheasant were raised during two seasons in Zoological Garden. A male hybrid was paired with hen common & hen Rees pheasant < (Penny Encyclop. vol. 18, p 61)> but was quite sterile. M. S. Report sent me from Zoological Society.

8 For crosses between Gold & Silver, each with the common see Proc. Zoolog. Soc. 1836, p. 84—& See also Is. Geoffroy Saint Hilaire Essais de Zoolog. Generale 1841. p. 493. The hybrids are said to be always sterile, but Temminck (Hist. Gen. des Gallinacées vol 2. p. 323) says that one hybrid from Gold & common Pheasants produced with pure P. colchicus a sterile hybrid grandchild of the two pure species. I give the cross between the Gold Pheasant & common Fowl on the authority of the M. S. return from Zoological Society: this hybrid was quite sterile. Mr. Hewitt has repeatedly tried to make this cross, & likewise between the Silver Pheasant & common Hen without success, though the birds coupled freely.—

5 The account of the hybrid Pheasant-Fowl, being fertile with pure Pheasant in Proc. Zoolog. Soc. 1836. p. 84—does not at all satisfy me for I saw the Birds But Dr. P. Lucas quotes Bomare (L'Heredité Naturelle Tom. 2. p. 307) for the same fact of the occasional fertility of these Hybrids with the pure pheasant; butI doubt the account, for Mr. Hewitt has had much extra-ordinary experience in making these hybrids; never once saw them show any sexual feeling.—

2 On the authority of Mr. Blyth in Rev. E. S. Dixon's Dovecot & Aviary 1853 [sic] p. 88. They perished when a few days old. See [Blyth] Indian Sporting Review New Series vol. 2. p 253.

14 Morton describes their hybrid in Proc. Philadelphia Acad. of Nat Sciences. Quoted in Annals of Nat. Hist. vol 10 [19], 1847. p. 210 The Zoological Society once possesed specimen of this hybrid, which was quite sterile. The cross from between Guinea-Fowl & common Pheasant, I give on authority of M. S. return from the Zoological Society: this hybrid was quite sterile.

11 Layard in Annal of Nat. Hist vol XIV. 2 series 1854. p. 63. The one hybrid raised was quite sterile. This is a very remarkable fact seeing how close this species is to G. Bankiva.

16 The reappearance of Tetrao medius of some authors in Scotland after the reintroduction of the Capercailzie is the best & most curious proof of its hybrid origin from this Bird & the Black cock. See [J. Wilson] Proceedings Royal Soc. of Edinburgh December 19. 1842 [vol. 1, p. 395.]

13 Wagner's Report on Zoology for 1843-44 in Ray Soc. for 1847. The hybrid from this species with common Fowl has been called G. aeneus. They are commonly raised in Java as I am informed by Mr. Crawfurd, but are believed to be always sterile.

1 Temminck Hist Nat. Gen. des Gallinacees vol 3. p. 13-21. The hybrids raised from these species are sometimes sterile & sometimes fertile with the pure parent species.

88a

I give the cross between the Gold Pheasant & common Fowl on the authority of the M. S. return from Zoological Society: this hybrid was quite sterile. Mr. Hewitt has repeatedly tried to make this cross, & likewise between the Silver Pheasant & common Hen without success, though the birds coupled freely.—

The cross from between Guinea-Fowl & common Pheasant, I give on authority of M. S. return from the Zoological Society: this hybrid was quite sterile.

88b

4 The hybrids from P. torquatus & colchicus there is good reason for believing are quite fertile see text infra.

6 With respect to the high fertility of the hybrids between P. versicolor & Colchicus, see text infra.

17 Lloyd states (Field Sports of N. of Europe vol I. p. 314) on the authority of Nilsson that the Black-cock has crossed with the Fowl, but the chicks survived only a few days; but Nilsson is so excellent an authority that his statement, may, I should think, be trusted.

9 Mr. Hewitt informs me that he raised three hen hybrids from the cock Gold Pheasant & Hen Silver Pheasant.

88c

18 Bronn Ges[ch]ichte der Natur B. 2. s. 166 and Wagner Report on Zoology for 1844 in Ray. Soc. for 1847 [p. 293.] <I presume that the translation is correct & that the term Ptarmigan is meant.> The Hybrid is the Tetrao hybridus lagopides of Nilsson.

12 Mr. Blyth raised many hybrids from this bird & a domestic Hen from Aracan, but these were quite sterile inter se & with the domestic cock or Hen. But several years ago I myself saw at the Zoological Gardens young birds, which were the offspring of hybrids inter se from a Sonnerat cock & Bantam Hen; some of these had returned most closely to the pure Sonnerat & others, as I was told, to the pure Bantam: there could hardly be any mistake here; for they had then only one Sonnerat cock; & the fact of the young (or grand-children) returning towards both pure species strongly corroborates their asserted parentage.—I was told by the curator Mr. Miller at the Zoological Gardens that hybrids have been raised between this Fowl & the Pheasant.

15 Temminck Hist Nat. Gen. de Gallinacees vol 3 p. 301. The hybrids are said to be always sterile. This remarkable cross is confirmed by Mr. Blyth, on the authority of Major W. Sherwill in the Indian Sporting Review new S. vol 2. p 241.—

19 There seems no reason to doubt from Mr. Macgillivrays account in his British Birds vol I. p 162. that hybrids from the Black & Red Grouse of Scotland have been produced in a state of nature.

3 Mr. Mitchell. Sec. Zoological Soc. saw at Amsterdam a hybrid between the Peacock & Guinea Fowl & brought home a drawing of it. A clergyman in England informs me that he has seen his Peacock though having a Pea-Hen frequently unite with a Guinea-hen; but no offspring was produced—

88cv

15 Temminck Hist Nat. Gen. de Gallinacees vol 3 p. 301. The hybrids are said to be always sterile. This remarkable cross is confirmed by

I will now discuss in some detail the degree of fertility of certain hybrids in successive generations.

(Small type)

Dogs—Wolves Jackalls Foxes: that hybrids between dogs & wolves & Jackalls possess a certain degree of fertility is notorious <I think evident from the frequent practice amongst savages, as has been alluded to in a previous chapter of universally crossing their dogs with wolves): M. Flourens (Longévité Humaine 1855. p. 143, 156) has made laborious experiments on a large scale with these hybrids having raised no less than 294 individuals. He finds that when bred inter se the dog-wolves invariably became sterile at the third generation, & the hybrids of the Dog & Jackall at the fourth generation.

But nothing is said to show that care was taken to prevent the ill effects of close interbreeding: if the hybrids were all raised from one dog & wolf & one dog & Jackall, & all were placed under the same conditions of life, then I think, starting with some degree of sterility, it is not at all surprising that their sterility should have been increased by the interbreeding, to such a degree that Mr. Flourens could not rear any offspring beyond the fourth generation. I saw a female hybrid from a Dog & Jackall in the Zoological gardens which even in this its first generation was so sterile that it only comes periodically into an

89Vv

<He raised no less than 294 hybrids between these three species.) This increasing sterility in the successive generations is curiously parallel to the same fact, as observed by Gaertner, in plants.

imperfect state of heat. Hybrids from the Fox & Dog will breed with the Dog, as shown in Pennant's Quadrupeds 3 Edit. vol i. p 239. & in Herbert's Amaryllidaceae. p. 338.

Goat & sheep. The fertility of the hybrids from these two animals, classed in distinct genera, has been exaggerated by some authors, see for instance < (Nott & Gliddon, Types of Mankind 1854. p 379)>

M. Chevreul (Annal des Sci. Nat. 3 ser. Bot. Tom. VI. p. 188) affirms on the authority of Mr. Gay who must have had excellent opportunities of ascertaining [?] the truth that the male hybrid from the male Goat & female sheep are paired with female sheep, & that the offspring, with 3/4 of the blood of the sheep, are propagated inter se, but that the character of the fleece deteriorates, so that after some generations they have to take a fresh cross with the half-blood or first hybrid. This seems a very remarkable degree of fertility, for it implies that after several generations the hybrids with one-fourth of the goats blood in them are fertile with the hybrid having half goats blood.

Bactrian & common Camel. There seems no doubt that hybrids are commonly raised (Burnes Travels in Bokhara vol. 3. p. 154) between these species though so remarkably distinct in structure & constitution. Eversmann enters into details & clearly distinguishes the Bactrian from common camel, &

90v

It is stated that Goats & sheep are habitually crossed in Chile (Molina Historia Geograph. del Reyno de Chile. 1788. P[art.]. I. p 376); & similar crosses have been effected in Europe.—

90bis

states that the hybrids have either one or two humps, & he expressly adds that these hybrids are fertile (Quoted in C. Ritter's Erdkunde. B. 8. Th. I. s 655, 659: I was guided to this reference by Nott & Gliddon Types of Mankind [p. 380]); but it is not expressly stated whether these hybrids are fertile inter se or only with either pure parent-species. The old traveller Olearius, as quoted by Desmarest in Encyclop. Method, says these hybrids are sterile, like the common mule. The great facility with which crosses can be made between the several forms of Llamas (Auchenia) & the reputed fertility of the hybrids, renders Eversmann's statements somewhat more credible.

Muntjac Deer. At Lord Derbys the Cervulus vaginalis from the Malayan archipelago & C. Reevesii from China bred together (Gleaning from the Menagerie Dr. J. E. Gray 1850. p. 65); & Mr. Thompson the intelligent Curator (& now curator of the Zoological Gardens) assures me that he is certain that the hybrids were perfectly fertile inter se, & that a herd was thus reared. No doubt these are close species; & Kölreuter in such cases would have at once called them varieties.

Bos grunniens (Yak) & taurus var. Indicus. Various hybrids have been produced between several species of oxen & Buffaloes, but all, as far as I can find out are quite sterile, with the exception of those between the two above-named apparently very distinct species, which indeed by many naturalists are ranked in distinct genera. These hybrids are raised in large numbers in the Himalaya. The case is interesting, in as much as the two parent species have remarkably different constitutions, the yak enduring extreme cold & the Indian cattle extreme heat: the yak, moreover, browses on very different plants.—Mr. Schlagintweit has lately stated (British Association: Dublin 1857)

that these hybrids are perfectly fertile together & that he examined some bred inter se to the seventh generation! But I must think that this is incorrect; for Dr. Falconer has given me the following precise details obtained from the best possible authorities in Little Thibet: the male Yak is habitually crossed with the female cow, but the reciprocal cross can also be made: the female hybrid (called Bsohn) can be fertilised by the male yak; & the female offspring (with 3/4 of Yak blood, called Gurmoh) can be crossed again with the male Yak; & the progeny (namely great-grand-children of the two pure species first crossed) are then quite fertile; both on the male & female side, & can be crossed inter se or with either pure parent. But in the earlier stages of reduction, though the female hybrids are fertile with the pure Yak, the males are impotent.

Fringillidae.: Bechstein states (Stüenvogel 1840. p. 248) that the hybrids from the canary birds with the Goldfinch, & with the siskin (F. carduelis & spinus) have certainly interbred inter se. I have heard of another case. Capt. Hutton, a good observer says (Calcutta Journ. of Nat. Hist. vol 2. 1843. p. 530) he has himself reared the young from a pair inter se. of mule birds from the canary & linnet. (F. canaria & linota), which young (grandchildren of the pure species) "again in turn produced & reared a brood". The cases, however, of hybrids from the Canary & other Finches breeding inter se are undoubtedly rare.. But the
experiment, cannot be considered as fairly tried until, some of the species with which the canary-bird is crossed breed quitereadily in confinement which occurs with none of them I have seen it stated but have failed to find the original account that Vielliot found the hybrids from the F. citrinella & canaria, which are very closely allied species, perfectly fertile.

Hybrid Pheasants: I am assured by Mr. Thompson (see also the Sale Catalogue [by E. S. Stanley]) formerly curator at Lord Derbys that he is certain that the hybrids from P. versicolor & Colchicus (manifestly distinct species) bred quite freely inter se, & that their progeny (grandchildren of the two pure species) again had offspring. The hybrids were also reduced by crosses with pure P. versicolor, & the great grand-children of the two pure species are undistingushable from P. versicolor (Report Zoolog. Soc. April 1855. p. 17): In the summer of 1856 I saw the latter birds & two of the hens had laid together 24 eggs, & 15 young had been reared from them.

Phasianus torquatus has of late years been turned out in several places into our woods, & has certainly crossed with the P. Colchicus. These species are closely allied, but I believe are universally admitted to be distinct (a)kind of pheasant, called the ring-necked is not uncommon in some parts of England, & is stated by Selby to have spread within a comparatively short time over nearly the whole of Northumberland. (Montagu's

94 insert

Temminck (Hist. Nat. Generate des Gallinaceas vol. 2. p. 326) distinctly states that the hybrids from them are fertile inter se. We have, also, indirect evidence on this head: the

Ornitholog. Dict. Rennies Edit. p 370). From different statements which I had read, I had always considered the ring-necked pheasant as a simple variety not due to crossing until Mr. Gould was so good as to show me the several points of resemblance between the common ring-necked pheasant & an undoubted hybrid from P. colchicus & torquatus. Besides the white collar, a trace of a white line over the eyes, a slightly more fulvous tinge on the flanks, & sometimes a trace of green on rump, & the bars on the tail all seem to indicate a cross from P. torquatus, as is likewise the conclusion arrived at Mr. Blyth after careful examination of the subject; it seems highly improbable that these several characters should have all concurred from simple variation. If this be so, the case is interesting, for as we have seen that the ring-necked pheasant has rapidly increased in some districts, its fertility must be very great; & although it is not at all known what is the exact proportion in which the two species are blended in the English ring-necked pheasant; yet there can hardly have been repeated crosses for many generations with the pure P. colchicus, otherwise the characters of P. torquatus would have been much more completely obliterated.

Chinese & common Goose (Anser Cygnoides & Cinereus domesticus). These birds are so distinct that most ornithologists place them into distinct genera; many cases are on record of hybrids from these birds breeding readily with the pure parent species; Mr. Eyton once succeeded by pairing birds reared from different hatches, in rearing no less than eight young birds from half-bred hybrids inter se, though both the parents were young birds, which would naturally have lessened their prolificness. (Charlesworth Mag. of Nat. Hist. vol I. 1837. p. 358. & vol 4. 1840. p. 90)

96 insert

(a text)

but they are generally quite sterile inter se; & one case is on record (["Zenas,"] Poultry Chronicle vol. 3. 1855 p. 487) of a hybrid with 3/4 of pure China-goose blood in them, being sterile inter se. Nevertheless

In India over large tracts of country, in the N. W. Provinces, in Assam, & near Calcutta, the geese as I am informed in letters from Mr. Blyth1 & Capt. Hutton2 are of a mixed breed, clearly intermediate in all their characters, even in their voice, between A. cygnoides & cinereus: Such acute & experienced ornithologists as these two observers could not possibly have been mistaken on this head. In many of these districts neither pure species is kept. Therefore there can be no doubt that, as I am assured by these gentlemen that these geese breed inter se; & Mr. Blyth says he believes they are fully as prolific as the common English Goose. It is indeed I think, obvious that their prolifickness, must be great; otherwise the breed would not be commonly kept for profit. We may perhaps attribute the much greater fecundity of their hybrid when bred in India than in Europe, to the difference of climate, & probably in large part of the numbers raised,—the ill-effects of close inter-breeding being thus wholly removed.

97 insert

2 Mr. Sundevall (Annals of Nat. Hist. vol. 19. 1847. p 171) noticed these crossed geese near Calcutta. Pallas in the Act. Acad. St. Petersburgh. 1780 p. 83. speaks of the hybrids from these two genera as being very prolific.

I have now given in some detail all the cases which I have collected on the degree of fertility of the most fertile animal hybrids. But I think that hardly one of these cases has been sufficiently investigated. The scantiness of the facts plainly shows how rarely there is with hybrid animals any approach to perfect fertility. In considering, however, the subject we should always bear in mind that the experiment is never fairly tried, without both parent-species breed perfectly well under domestication or confinement, & without both are placed under favourable conditions & without several hybrids, not related to each other, are raised at the same time, so that the ill effects of close interbreeding in the successive generations may be excluded. Very few cases are on record in which all these conditions have been fulfilled. Nevertheless, the hybrids from Phasianus Colchinus with P. torquatus & with P. versicolor, & in India from Anser cygnoides & cinereus probably make a very close approach to perfect fertility, or perhaps are perfectly fertile together.

Most naturalists now believe that many of our domestic

animals, as dogs, cattle, sheep &c, have descended, each from several aboriginally distinct species. In some cases, this seems to me the most probable view. Those who admit this view, must suppose either that there once existed several distinct species, which were capable of uniting & of producing perfectly fertile hybrids, which we have now domesticated around us; or they must suppose, in accordance with the view first broached by Pallas,1 that species originally infertile together, become quite fertile through a long course of descent under domestication. That the making of the first cross should be facilitated by both species having been thoroughly domesticated seems extremely probable; but I know of no actual facts to support this view, except the statement by Dureau de la Malle who has so closely studied classical literature, that the common Mule was produced with more difficulty in the time of the Romans than at present:2 on the other hand, Gaertner3

1 Act. Acad. St. Petersburgh. 1780 Part H. p. 84, p. 100.

2 Annales des Sc. Nat. Tom 21. (1 series) p. 61.

3 Bastard[er]zeugung s. 11, 12. [At top of next folio, Darwin added in pencil: 'Aegilops being more fertile N. Hist. Review.']

100

could perceive no difference in the facility of hybridising cultivated & wild plants & on asking W. Herbert he expressed to me the same opinion: but neither of these botanists have experimentised on the very same species in its wild & cultivated state. Even if the first cross between two animals could be effected more easily, when both were domesticcated, I know of no fact whatever countenancing the view that the fertility of the hybrids thus to be produced, would be greater, after the parent species had long been domesticated, than at first. Nevertheless, I must confess that there seems to me much probability in this hypothesis. Believing as I do, that our dogs, for instance, have descended from several distinct wild stocks: analogy prevents my believing that if their wild stocks had been caught & paired that their offspring would have been as fertile as are our mongrel dogs:1 but how much of the infertility in this hypothetical case would have to be put down to the wild parent stocks not breeding readily under confinement, & the many hybrids

100 insert

Mr. Blyth ['Zoophilus'] has remarked (Calcutta [actually India] Sporting Review vol 2. new series p. 133) that the N. American wolf & the Canis latrans keep distinct in their native state, & yet it is believed that the Indian dogs, which I think there can be hardly a doubt have descended from these two wild species, mix readily together; & Richardson has described (Fauna Boreali-americana [I] p. 80) an intermediate race.—For an interesting discussion on this subject, see Nott & Gliddon, Types of Mankind p. 384.—

101

not having been raised so as to prevent the ill-effects of continued interbreeding, it would be hard to conjecture. If this hypothesis could be proved true, it would throw considerable light on the history of our domesticated animals; & would be interesting for our theory as it would show that the sterility of hybrids was a varying quality, which would in some degree lessen its importance as a diagnostic character between Species & Varieties.

Fertility of crossed Races <Varieties>, in Animals, & their Mongrel Offspring.—As I have already more than once remarked, after seeing the almost universal lessened fertility of even very closely related species when crossed, the perfect, nay very generally increased, fertility both of the first cross between the most widely different varieties & of their mongrel offspring is a highly remarkable fact. In plants I have been able to give a few cases pretty well authenticated of some degree [of] sterility in crossed varieties: with animals I can advance no satisfactory evidence on this head. From the facts

102

given in our sixth (?) chapter, I do not doubt that animals of the same variety sometimes prefer pairing with each other; but this, though an interesting fact, & bearing on the reputed repugnance of distinct species to unite, does not concern their fertility when crossed.

When we hear that certain domestic breeds of native American dogs3 do not pair or even readily associate with other breeds; when we hear it said, but on what grounds I know not,4 that certain breeds of dogs, are more fertile when crossed together than other breeds, the explanation probably is in the case of the American dogs, & perhaps in the latter case, that these dogs have descended from primordially distinct species, & not that any degree of relative sterility has been acquired during domestication. The same explanation probably applies to Bechstein's statement5 (if to be trusted) that dogs of the Spitz breed can be easiest crossed with Foxes: in plants, however, I may remind the reader that we apparently have good cases of varieties of the same species uniting with different degrees of

102 insert

Nor can physical obstacles from great difference in size between two varieties be strictly considered as causing lessened fertility. Thus it is well known that bitches paired with dogs of large size, often die during parturition. I presume very unequal size would sometimes interfere with the union of varieties; though A. Knight got offspring from a Dray stallion & Norwegian Pony1 & chickens from a Cochin cock & Sebright Bantam Hen were exhibited at Manchester2

2 Poultry Chronicle vol. 2. 1854. p 446.

102 insert 2

3 Rengger Saiigethiere von Paraguay 1830 s. 153, on the Hairless dog.—Gosse's Sojourn in Jamaica p. 339. on the Alco or Mopsy Dog.

4 Gaertner, Bastardzeugung s. 577.

5 Naturgeschichte Deutschlands. 1801. B. l. s. 638.

103

facility with distinct species. Believing as I do that some of our dogs are descended from the European wolf, & seeing that the hybrid from the wolf & dog shows some sterility, I should have been tempted to surmise had the experiment been made with a breed like the Hungarian sheep-dog, which is extremely like the wolf, that the dog had become so much modified by domestication that the fertility of its offspring when crossed with the wolf had become impaired.

Many naturalists believe that the Llama & Alpaca are only varieties of the wild Guanaco: Mr. W. Walton, who has particularly attended to these animals6

6 The Alpaca. 1844. p. 29.

104

says the two first named often breed together, but that their offspring are quite sterile & are hence called "Machorras". From other statements which I have heard, I doubt the fact, but supposing it to be true, the inference would be that those naturalists are right who view the Llama & Alpaca as distinct species, now utterly exterminated in their wild state. A good authority says that the first cross from the Long-& Short-horned cattle is excellent, but that in the third or fourth generation "there is much uncertainty whether the cows will hold to the bull; & full one-third of the cows among some of these half-breds fail of being [in] calf."1

Supposing that this remarkable statement could be trusted, I do not doubt that some naturalists would immediately argue that our Long-& Short-Horns have descended from two distinct species.—

I have given the foregoing details to show how much inherent & almost insuperable difficulty there must always be, from our ignorance of the

104 insert

In the case of all our domestic animals, though crosses are very frequently made, yet from their extreme uncertainty, they are very rarely propagated for many generations, so that it is quite possible that a slight degree of infertility might long remain undiscovered. But in the case here given it so happens that the cross has been systematically made, & according to Mr. J. Wilkinson2 a half-breed has been fully established; & this makes me doubt Youatt's statement; for any marked degree of infertility would surely have prevented any one establishing the half-breed.

105

history of our domestic breeds, in this subject.—If any two breeds had ever become so different as to be in the slightest degree sterile together; scarcely any amount of evidence would convince naturalists that both had descended from the same parent stock.

The case of perfect fertility between varieties, which has struck me most, is that of Pigeons: I have myself largely experimentised on the fertility both of simple & the most complicated crosses between the most distinct breeds; & I have given my reasons for fully believing that all are descendents of one species. Compare a Pouter, Tumbler, Carrier, Fantail & Barb, which produce together quite fertile mongrels, & see in how marked a manner they differ from each other, in comparison with Gold, Silver & common Pheasant, or with the Java & common Fowl. (Gallus domest. & varius), from which it is often difficult to rear any young

106

& these when reared are utterly sterile! Such cases as this of the Pigeons are very surprising & seem to stand in direct opposition to our view that species do not essentially differ from varieties. But there are some considerations which make the case not quite so contradictory, as it at first appears. In the first place, it has, I think been clearly shown, in accordance with Gaertner's conclusion, that the power in any two species of easily producing hybrids & these more or less fertile does not strictly run (as well seen in the different results from reciprocal crosses) with their systematic affinity, that is with the amount of resemblance which can anyhow be detected by the eye: hence we should err if we were, under the view of varieties not differing essentially from species, to infer that because a Powter-pigeon apparently differs more from a Tumbler than the common pheasant does from the Silver, that there ought necessarily to be fully as great difficulty in getting fertile hybrids from the two pigeons, as from the two pheasants.

107

Secondly, those many naturalists who believe that most of our domestic animals are descended each from several aboriginal species, & who, therefore, must believe either that perfect fertility between species when crossed is very far from uncommon in nature, or is a quality readily acquired under domestication, will feel little surprise at the fertility of varieties when crossed. On the view, indeed, of sterility being commonly lost between species when under domestication, it would be most strange if sterility were, also, to supervene between varieties under raised domesticcation.

Lastly: sterility from hybridisation, like that from other & quite distinct causes, must be looked at, not as a specially endowed quality, but as incidental on other & unknown differences in the sexual organisation of the species which are crossed,—as we see very plainly in the frequent & great differences in the results of reciprocal crosses. Now man both in his unconscious & in his methodical selection of varieties useful or pleasing to him, selects almost exclusively by the eye; he has neither the power or the wish to affect by continued selection those

108

obscure & inner constitutional differences, on which the sexual affinity of distinct species seems to depend. Moreover man does not select each variety in exact relation to the conditions to which he exposes it; nor does he keep the conditions as constant as possible; nor is his selection uniform in direction & extremely slow. How differently nature acts! She keeps her conditions uniform or nearly uniform for thousands & tens of thousands of generations: if she modifies her productions she modifies them most slowly & uniformly only for the good of the selected variety. And who can say what may be the difference in the results of these two kinds of selection? In a somewhat analogous manner, as species which are so generally adapted by nature to a certain limited climate, would appear, as before shown, when domesticated by man to lose to a large extent this close adaptation, & as the varieties raised by him acquire only in a very limited degree such kind of adaptation; so it seems not very improbable that species domesticated by man may lose, as some have thought, this tendency to sterility when crossed, & the varieties

109

raised by him not acquire it, or only, as with certain plants, in a very slight degree. Hence I conclude that it is not so surprising, as on our view of the nature species it must at first appear, that varieties formed under the care of man, should not have become modified in their sexual organisation, in that mysterious manner on which their greater or lesser power of crossing with other forms & of producing more or less fertile offspring depends, in the same remarkable manner as is so generally & eminently characteristic of Species.

110

Comparison of Hybrids & Mongrels, independently of their fertility: Plants We must now compare the hybrid offspring of species & the mongrel offspring of varieties, & see how far they resemble or differ from each other, independently of their fertility. This subject is of some importance for us; for it would be strange if the union,—that most mysterious problem in physiology—of two varieties, produced by natural causes & the offspring of two species, supposed to have been separately created, yielded offspring, which followed the same laws in their likeness to their parents, & in other respects; & if this can be shown to hold good, in all essential particulars, the fact may be used as an argument that the origin of species & varieties is essentially alike. For several reasons I will discuss this subject separately & first for Plants & then for Animals.

As some authors have considered hybrids as monsters, & beyond the pale of law, it will be well to premise that Gaertner & Kölreuter have clearly shown that this is quite a false view. <Undoubtedly the hybrid progeny of two

111

species follows, at least approximately, the same laws of resemblance as do the individual offspring of the two sexes of the same species.> When two not-cultivated species have been repeatedly crossed, their hybrid offspring of the first generation are as a general rule found closely to resemble each other;1 & when rarely "exceptional types" are produced, which abnormally closely resemble either parent-species, such types when reproduced at long intervals from the same two species, are alike. Moreover, when these exceptional types are sufficiently fertile to propagate themselves, their offspring generally revert to the normal hybrid type of the two species in question.2 Nor, I may add, do malconformations3 occur oftener in hybrid plants than with pure species: even their reproductive organs, as we have seen, are only functionally imperfect.

1 Gaertner Bastardzeugung s. 284.

2 Bastardzeugung s. 238, 424.

3 Bastardzeugung s. 518 557.

112

Gaertner with his immense experience has compared in detail1 hybrids & mongrels: it is evident that he would gladly have seized on any difference in the progeny of species & varieties, for such is the clear tenour of his whole admirable work. He reduces, here leaving out the question of fertility, the main differences to two,—namely that mongrels, especially in the first generation, but likewise in the succeeding generations are much more variable than hybrids in the first & even in the succeeding generations: secondly, that mongrels evince a much stronger tendency to revert to either pure parent-form than do hybrids; but this latter difference in fact is only a part of that first specified of greater variability & less antiquity. As a general rule, I think there can be no doubt that mongrels of the first generation are much more variable than hybrids of the first generation. But "exceptional types"2 , which are nothing but strongly marked variations of a definite kind, do occasionally occur amongst hybrids of the first generation3 : and Gaertner further admits that lesser variations do likewise sometimes occur, but he adds, & the observation is an

1 Bastardzeugung s. 582.

2 [Darwin later pencilled: 'Are not these only reversions?']

3 Bastardzeugung s. 238.

113

important one, that he has noticed these lesser variations only in hybrids from species which have long been cultivated in gardens.

It is, also, an important observation that these lesser variations in the first generation have been principally observed by Gaertner6 in hybrids between species so closely related that they have been thought by some authors to be varieties, though ranked by Gaertner by the test of fertility, as true species. In this greater variability in hybrids from between very closely related species we plainly see a gradation towards the strongly marked variability of mongrels in their first generation. Moreover, it is well to observe that there are causes [?] of variability in hybrids, which seem to escape all law; thus in the following combinations, Dianthus barbatus fertilised by the hybrid D. barbato-carthusianorum gives many more varieties than the reciprocal cross of D. barbato-carthusianorum by D. barbatus; so again with these distinct species, Lobelia fulgens fertilised by L. cardinali-syphilitica yields a more variable hybrid progeny, than L. cardinali-syphilitica fertilised by L. syphilitica.1

6 Bastardzeugung. s. 247, 249, 577.

1 Bastardzeugung s. 445, 507, 513.

113v

I saw at Spofforth4 two hybrids between Rhododendron & Azalea raised by Herbert from the same seed-capsule, & they differed greatly in appearance

4 [At Spofforth, Dean Herbert gardened and experimented with hybrids; see his Amaryllidaceae, p. 359.]

113 insert

In fact there cannot be a doubt that hybrids of the first generation from between two species both long cultivated often differ greatly from each other as in hybrids from distinct species of Rhododendron, Passiflora, Fuchsia &c.5

5 Herbert Amaryllidaceae. p. 359: Mr. Sabines account of hybrid Passiflora in Hort. Transact. vol Iv. p. 261, 266. In Hybrids from Fuchsia coccinea fertilised by fulgens, Mr. Thwaites says "scarcely could any two be found so much alike as to be undistinguishable": he adds a truly remarkable case of one single seedling, half of which more resembled the one species, & the other half the other species.

114

Now considering these facts, can it be considered as a surprising or important difference that mongrels in the first generation should be more variable than hybrids? In the first place the far greater number of varieties have been produced by longcontinued cultivation, & this we have seen makes true hybrids variable in their first generation. Secondly, according to Gaertner, hybrids are more variable from between closely related species than from between those which are very distinct; & of course varieties are closely related to each other.

But we do not know that mongrels, especially from varieties in a state of nature are universally highly variable in the first generation: we must remember that extremely few experiments have been systematically made & recorded on varieties: I find in Kölreuter's works2 a few cases of crosses some made reciprocally, between several varieties of Mirabilis, Matthiola & Nicotiana; & no mention is made of any extreme degree of variability, as from the tenour of his works

2 Zweite Fortsetzung s. 56, 126, 128: Journal de Physique 1782. p. 285: Nova Acta Petrop. 1795, p. 333 & 1797. p. 393.

114

Thirdly, variability in itself is certainly inherited, & as varieties are in many cases only recognised as varieties from this very quality, it would be strange if their mongrel offspring were not commonly thus characterised.

115

might most safely have been expected, had such variability occurred.

Turning now to the generations succeeding the first in hybrids, every observer has been struck with their extraordinary variability, & some authors have even thought this a more important characteristic of hybrids than their lessened fertility Gaertner freely admits this variability in the successive generations of hybrids whether fertilised by their own pollen, or by that of either pure parent-species: Kölreuter used the strongest expressions on this head, as does W. Herbert.3

If mongrels in their successive generations are more variable than hybrids in the corresponding generation, which I think probably is the case, at least with cultivated varieties, the difference is only one of degree; the comparison moreover can only have been vaguely made for I know of only one case on record, in which the offspring of two varieties have been carefully observed

3 Gaertner's Bastardzeugung s. 518: Kölreuter in Nova Acta Petrop: 1794. p. 391: Herbert Amaryllidaceae. p. 348.

115V

B text)

According to the view, which seems to me most probable, on the cause of variability, namely the influence of the condition of life which either one or both parents are exposed, on their reproductive system, & consequently on the ovules & pollen

[Not transcribed by Stauffer. JvW]

115Vv

B text)

This highly remarkable & unexplained difference in the degree of variability of hybrids in the first generation, compared with that in the successive generations is quite conformable with the view of the cause of ordinary variability (independently of crossing), which as stated in the first chapter seems to me by far the most probable; namely that the reproductive system is affected by the conditions of life to which either one or both parents have been exposed, in the same manner but in a lesser degree, as in those many cases in which actual sterility supervenes from cultivation or confinement.—For in hybrids of the first generation from between two species which have not been modified by cultivation, the reproductive system of neither parent has been in any way affected; whereas in the successive generations from hybrids, we well know from the lessened fertility of the latter how seriously their reproductive system has been generally affected.

116

for several, in this one case for four generations1 . Nor is <the> rule universal that the successive generations of hybrids are highly variable; for Gaertner has given2 five cases in which the progeny kept constant, & one, namely the offspring of the hybrid Dianthus armeria-deltoides was observed even to the tenth generation. In the same manner it would appear that occasionally, though very rarely, the mongrels keep true; I am assured by an intelligent nurseryman that "Dale's hybrid turnip" has every appearance of being a hybrid, & that it does not vary; & Mr. Beaton,3 has remarked that "Melvilles most extraordinary cross between the Scotch Kale & an early cabbage is as true & genuine as any on record". In these cases there may probably have been some selection in the early generations; but had the variability been as extreme, as it generally is, no one would

1 Kölreuter on varieties of Mirabilis in Nova Acta Petrop. 1797. p 393.

2 Bastardzeugung s. 553. Dean Herbert showed me some hybrids from two species of Loasa, which had kept constant for several generations.

3 Cottage Gardener. 1856 p. 110; Wiegmann in his Bastardzeugung s. 33 says that seedlings from his mongrel cabbages, as a general rule, retained their blended nature.

117

have had patience to have raised a true mongrel race.

With respect to Gaertner's statement that mongrels show a greater tendency to revert to either pure parent state, nearly all the foregoing remarks are applicable, as indeed this reversion is only a form of variability. Even if proved strictly true & I must repeat my remark on how few mongrels have been carefully observed during several generations; it would be only a difference in degree, for Gaertner gives many cases of true hybrids reverting to ancestral forms an anomalous circumstance that according to Gaertner the most fertile hybrids which according to our view, differ least from varieties do not show this tendency to reversion.1 I may give one remarkable instance from Kölreuter2 in the offspring of the hybrid Mirabilis Jalapa-longiflora, fertilised by several varieties of M. Jalappa, in which some of them, though having in their composition only one quarter of M. longiflora, yet more closely resembled that species, than their hybrid mother.

1 Bastardzeugung s. 236, and 420-446, 474.

2 Dritte Fortsetzung s. 53, 59.

117 insert

; & it deserves notice that this tendency to reversion was observed much oftener with cultivated than with wild species. Nevertheless it is

118

With the exception of the new specified differences, though these seem to me to be of extremely little importance, hybrids & mongrels in nearly every other respect apparently have the closest resemblance,—the degree of fertility, as before, not being here considered. Both are remarkably luxuriant, hardy & precocious. Both generally come alike from reciprocal crosses.3 Both follow nearly the same rules in their resemblances to their two parents: Gaertner4 has taken much pains in classifying the resemblance of hybrids to their parent species, both in the first & successive generations; he makes three classes but which he fully admits, blend together in an inextricable manner; (1st) hybrids very nearly intermediate in their whole structure; (2nd) hybrids, (& these are extremely common) resembling, but not identical with, one parent in one part, & the other parent in another part; (3rd) hybrids decidedly resembling one of their parents. In the first generation any particular hybrid may generally be classed under one of these heads, but in the succeeding generations the same hybrids often break into all three classes: so it is with mongrels. Gaertner asserts that hybrids following the first or intermediate class of resemblance are generally

4 Bastardz. s. 277-94, 580.

118 insert

3 The only cases, carefully recorded of reciprocal crosses amongst varieties, which I know of, are given by Kölreuter in Mirabilis, Matthiola & Digitalis thapsi & purpurea. (Dritte Fortsetzung s. 126, 128: Nova Acta Petrop. 1795 p. 333. 1797. p. 393: Journal de Physique 1782 p. 285). I should not expect the rule of the similarity of reciprocal crosses to hold strictly with long cultivated plants (see Wiegmann, Bastardz. s. 10, 11 on mongrel cabbages) for Gaertner. (Bastardz. s. 223) says it is more particularly true in regard to hybrids from species growing in their wild & natural state.

119

raised from between very closely allied species; & he further remarks that mongrels commonly belong to this same type.1 But I must think, from what I have myself seen in mongrel cabbages & raddishes, & from some of Mr. Knights descriptions of mongrel apples & grapes2 that with long cultivated plants the variation of the mongrels is so excessive that seedlings even from the same pod might be generally ranked in all three classes.

Gaertner has clearly shown that certain species possess a pre-potent power over other species with which they are crossed (distinct from their fertilising power) of impressing their likeness on their hybrid offspring.3

This difference of prepotency has not been proved to exist in varieties, owing as I believe, to the fewness of the experiments tried; but it holds good, as we shall see, with the varieties & apparently even with the indi[vi]duals of animals, & I cannot doubt that it does also hold with plants; for this prepotency is closely connected with the power of one species reducing another

1 Bastardzeugung s. 282, 578. This was, also, Mr. Knights opinion p. 39 Treatise on the Culture of the Apple & Pear.

2 Philosophical Transactions 1799 p. 201.

3 Bastardz. s. 290 256.

119v

(B note continued)

This was, also, Mr. Knights opinion p. 39 Treatise on the Culture of the Apple & Pear.

119 insert

To give one single instance from Gaertner of Nicotiana

120A

Some of the special cases of resemblance of hybrids to their parents are curious. If we put aside the species having a prepotent power of transmitting their likeness, then in complex crosses of two species, the appearance of the hybrid depends on the proportions in which the parent-species have been blended together. Thus in Dianthus barbato-barbatosuperbus, in D. barbato-superbobarbatus, & in D. barbatosuperbo-barbatus we have two species differently mixed, but in the same proportion, namely with three-fourth of D. barbatus in each & they closely resemble each other. (Gaertner Bastardz. s. 504) But the result is wholly different when a hybrid is fertilised by the pollen of a third pure species (for instance Lobelia fulgenti-cardinalis fertilised by L. syphilitica), for in this case the triple hybrid always closely resembles (though having only half blood) its pure father, so closely that it might often pass as a mere variety of it: Kölreuter says he was almost as much astonished at one such case in Nicotiana, as if he had seen a cat born with the form of a lion. (I Fortsetz. s. 42. Gaertner Bastardz. s. 511. gives several cases. M. Regel quoted in Journal de la Soc. Imp. d'Horticulture vol 1855. p. 251. makes a similar remark on some hybrid Achimenes.) On the other hand, in the exactly reversed case, namely of a pure species fertilised by the pollen of a hybrid from two other species (for instance, Lobelia fulgens fertilised by L. cardinali-syphilitica), the triple hybrid does not take after its pure mother or after its hybrid father (Gaertner Bastardz s. 507).

120

by successive crossings; & the power of one species reducing another in fewer or more generations depends not only on its specific difference, but on that of the variety or even on the indi[vi]duals used, & likewise on whether the species has been long cultivated.5

Gaertner1 adduces, the fact that one species can. be made by repeated crossings to reduce or absorb another, as an "unequivocal proof' that species have fixed limits. This seems to me singular reasoning, for supposing that the case had been exactly the reverse, namely that it had been found impossible to reduce by crossing one species into another, might not this with much greater force have been advanced as a proof of the aboriginal & immutable difference of the two species? This argument was indeed used to me

At one time I had thought it probable that if a variety produced by culture were crossed with an unaltered & distinct species, that the artificial variety would have less power

5 Gaertner Bastardzeug. s 458, 461, 465.

1. Bastardz. s. 475.

120v

Gaertner assuredly would not have disputed that one variety might be reduced by another; and

[illeg] Pigon Book

120 insert

by an acute observer, on my telling him of a case, where the effects of a single cross from the Malaya breed of Fowls was occasionally perceptible in a stock of poultry after an interval of 40 (thirty) years: he argued from the stain of the Malay blood being so permanent that, it could be only due to its being an aboriginally distinct species. On the other hand M. M. Boitard & Corbié2 have argued that because in crossing certain breeds of Pigeons, their characteristic features are lost even in the first generation & cannot be recovered without extreme difficulty, by crossing the mongrels repeatedly with pure birds, that such breeds are true species!)

So we see how this argument may be turned round & round to do any duty.

2 Les Pigeons. 1824 p 198.

121

than the unaltered species of impressing its likeness on their mutual hybrid offspring. But for some few crosses made by Kölreuter3 between the varieties of one tobacco with another species, & between several long-cultivated forms of Dianthus with wild species, this does not seem to be the case; & a hybrid from between a variety of one species & a second unaltered species seems as often to come intermediate as from between two unaltered species. Gaertner lays great stress1 on the fact that when two distinct, but closely allied species are crossed with a third species, the two sets of hybrids are very distinct from each other, even more distinct than the two closely allied pure species are from each other. On the other hand if two varieties of one species are crossed with a distinct species, he asserts that the two sets of hybrids differ very little from each other. In regard to this latter statement, no other facts are given, but two sets of crosses from between several varieties of two species of tobacco with

3 Fortsetzung s. 29. Dritte Fort. 72. 79, 83, 87, 108: in the Zweite Fortset. s. 116, there is, however, one somewhat opposed case.

1 Bastardz. s. 273, 581.

122

two other distinct species. Now Kolreuter2 also crossed several varieties of tobacco with a distinct species & he expressly states that the hybrids differed as much from each other as might have been expected from the difference of the varieties, & more than this could hardly have been expected. This same result seems to have followed crosses of differently coloured varieties of Verbascum with distinct species & of two varieties of Digitalis purpurea with D. lutea3

2 Fortsetzung s. 81. Zweite Forts. s. 56.

3 Journal de Physique 1782. p. 291. Dritte Fortsetzung s 6, 35. Acta Acad. Petrop. 1781. p. 249, 257 [Darwin cancelled the next sentence, before completing it: '<Even Gaertner would not have disputed that a common or wild cabbage, if crossed for instance with a Kohl-Rabi & Brussel Sprouts (as I have seen) would yield remarkably different>' He later pencilled: 'Probably here Laburnum case.']

123

Comparison of hybrid & Mongrel Animals, independently of their fertility.—In comparing hybrids & mongrels together, & both with their two parents, we meet with great difficulties, besides those necessarily inherent in all such comparisons. In animals we have quite commonly secondary sexual differences, so that a hybrid or mongrel has to be compared more or less with both sexes of both parents;

In the next place, differently from in plants, the progeny from reciprocal crosses between two species, or two races, is generally unlike, & this greatly complicates the case. As with plants, one species or one race is prepotent over another in transmitting its likeness to its crossed offspring; but with animals the prepotency seems often to run in one sex, which probably accounts for

123 insert

& in hybrids, owing I presume to their sterility, the secondary male characters are developed late in life & apparently not fully at any period; for instance Mr. Hewitt informs me that he has never seen even in old hybrid Pheasants & fowls, full-sized spurs.

124

the very frequent dissimilarity of the offspring from reciprocal crosses. But besides this general prepotency certain parts or characters, in certain species or even genera, appear to be more readily transmitted by one than the other sex, perhaps most frequently by the male sex; & they may be transmitted to both sexes, or only to one sex; & that sex may either be the same or different from the one which transmits the character.1 Moreover some authors believe that the age & vigour of the male influence the character of the crossed offspring, but I have not met with any satisfactory evidence on this head. Altogether it is not easy to exaggerate the complexity of the subject, & I will in consequence make only a few remarks on some of the points of comparison between hybrids & mongrels.

Isidore Geoffroy Saint-Hilaire has stated2 that hybrids from between two species generally present fixed & constant characters

2 Dict. Class. d'Hist. Nat. Tom x p 121.—1826; & subsequently in other publications, as Essais de Zoologie Generate 1841 p. 516.

124v

1 Numerous facts confirming these propositions may be found in Dr. Prosper Lucas' work on L'Heredite Naturelle. I shall have to give illustrations on several of them.

125

partly those of the father & partly those of the mother: on the contrary that mongrels are either intermediate like hybrids, or resemble entirely one of their parents. Two somewhat different considerations are here excluded, namely the resemblance of crossed offspring to their parents, & their homogeniety with respect to resemblance. I think there can be no doubt that hybrid animals, exactly as with plants, are either intermediate in structure, or take more after one parent in one part & the other parent in another part, or are altogether more like one parent than the other. But hybrid animals, in the first generation perhaps hardly ever so closely resemble either parent as do mongrels: Bechstein, how-ever,3 says that hybrids from the Canary & Fringilla spinus, always have both the colour & form of the latter. With respect to homo-geneity, hybrid animals from between the same species as with plants not long cultivated, seem generally to be alike; but there are marked exceptions as in the offspring from a Dog & Wolf for instance those described by Wiegmann4 two of which resembled the ordinary wolf hybrid, but a third took closely after the pointer: in a flock of hybrids from the common & Chinese goose, I saw some with

3 Stubenvögel 1840. s. 239.

4 Bastardzeugung s. 21:

126 insert

a text)

Hybrids, however, from the Carrion & Hooded crow resemble in their colour either parent or an intermediate.2 Hybrids from the Capercailzie & Black-Game differ in size & colour from reciprocal crosses, but this is not sufficient to account for the amount of difference; for

(B text)

Of two hybrids from the Guinea & Common Fowl, examined by Morton, "one looked more like the fowl, the other had much stronger resemblance to the Guinea Fowl."1

126

black & some with yellow beaks like one or the other parent; & the Rev W. D. Fox informs me that in some other hybrids which he had seen there was considerable diversity in the degree of resemblance to either parent goose. Hybrids from between the Canary & linnet are said to differ; & from between the common Pheasant & P. torquatus Temminck says some are like one parent & some like the other, & some intermediate.

But in all these cases one or both of the parents have been more or less domesticcated.
out of twenty of these hybrids, and all of the male sex, not two are said to be quite alike.3

1 For Guinea Fowl, see [Morton] American Journal of Science 2 Ser. vol 3. p. 204. For the Pheasants see [Temminck] Hist. Nat. Generale des Pigeons et des Gallinacees Tom 2. p 330. I give the linnet case on the authority of Dr P. Lucas Tom. i p. 211; but this diversity in hybrid canaries I do not believe from enquiries which I have made to be common.

3 L. Lloyd. Field Sports of the North of Europe vol. I. p. 285; on the authority of Mr. Falk: Latham, also, has in his Synopsis (Supplement) noticed the great variability of these hybrids.

127

So extremely few cases are on record of hybrid animals breeding inter se, that it is not known what rules the successive generations of hybrids would follow in their resemblance to their parents; but there is a high degree of probability that they would vary like the successive generations of hybrid plants, & take after either one or other parent-species. In the offspring from the hybrids from a hen Bantam & Gallus Sonneratii, which I saw, some resembled in an extraordinary degree the pure grandfather G. Sonneratii; & I was told that others had taken closely after their grandmother, the Bantam Hen.

Now let us turn to mongrels: the general rule seems to be that they are in some degree intermediate, between their two parents, & homogenious [sic]4 It is notorious that breeders, who have once crossed two breeds of cattle or sheep can gradually foretell what the character of the

4 See Colins Traite de Phys. Comp. des Animaux Domestique. Tom 2. p. 356, who has well treated this subject.

128

offspring will be in the first generation, which shows that there is no such great variability. I have crossed in Pigeons Barbs & Fan-tails, Pouters of two sub-breeds & Fantails, & in this latter case reciprocally, & I have been surprised at the similarity of the mongrels, even in colour; as I have found likewise to be the case from several other crosses. It is chiefly in dogs, pigs, Fowls & other animals producing many at a birth, that such great diversity has been observed, even in the young produced at the same time; but I can hardly persuade myself that there is not some exaggeration, when the young have been said to perfectly resemble one of the parents. Mongrels, bred inter se, after the first generation, no doubt present the most extraordinary diversity & reversions to their pure grand-parents, as I have myself seen from the very uniform pigeon-mongrels when bred inter se. Occasionally, however, characters immediately become fixed in a mongrel breed;
The Editor of the Poultry Chronicle3 bred some blueish fowls

3 Vol. i. 1854 p. 101.

128v

(Used)

128 insert

Boitard & Corbie1 who have had immense experience in crossing Pigeons, assert that from a Pouter & Runt

129

from a Black Spanish & Malay hen, & these remained true to colour "generation after generation"4 By the aid of some selection several intermediate mongrel breeds of sheep, as the Oxford & Shropshire Downs have been firmly established, & amongst cattle a breed, before mentioned from Wilkinson between Long & Short-Horns.—

With respect to the rules of resemblance of hybrids & mongrels to their parents, it deserves notice that very many attempts have been made to give laws such as that the Father gives external characters, & the mother internal or vital organs &c &c None of these rules, if widely extended to all animals seem to hold good, as has been ably shown by Dr. P. Lucas & Gaertner5 by merely contrasting the diversity of the rules given by different authors, who have had ample means to form an opinion. Similar rules have been enounced for plants, & have I think been con-clusively shown by Gaertner not to be true. Prepotency of one species or race over another, has been generally confounded with the influence of sex. If we confine our view to the races of one species, or perhaps even to the species of the same group, some such rules may

Bastardzeugung s. 264-266. L'Heredite Nat. Tom. 2. B[ook] 2. Ch. I. I could add other rules to those given by these authors.

(a) I crossed Penguin [ducks] & Black B[uenos] Ayres & the offspring kept not perfectly but very nearly true of a brown colour, a few darker again to blue, with center white mark on breast & even the bill.

(note continued)

In not a few cases examples have been given of crosses, without the manifest necessity of a reciprocal cross having been made.

130

hold good; for instance it seems in crossing reciprocally different breeds of poultry, that the cock very generally gives colour;6 & in sheep that the Ram gives the character of the fleece & horns, & the Bull its horns or absence of horns.
But what concerns us, is that I have never observed that a different rule has been given for hybrids & mongrels

One case, however, seems to occur frequently with mongrels, almost in accordance with Is. Geoffroy's remark, and which as far as I am aware has not been noticed in hybrid animals from between species in a state of nature; or only in a very slight degree as in case of carrion & hooded crow; namely either the perfect transmission or entire absence of some marked character of one of the parents in the mongrel

' Cottage Gardener 1856 p. 101, 137 for Poultry.

130v

*For other cases see P. Lucas L' Heredite Naturelle
Tom I p. 212

[not transcribed by Stauffer. JvW]

130 insert

And I think we may safely follow Lucas1 that the same wide & diverse rules of resemblance are common to the crossed offspring between species, varieties & individuals of the same race.

1 L'Hérédité Naturelle Tom. 2. p 179-184.

131

but intermediate states also appear: so it is sometimes with the condition of the hair. The dwarfed & turnspit like structure of the Ancon sheep when crossed with others seems to have either not at all or almost perfectly transmitted their characters. Piebald animals, & such cases as the mongrel offspring from the Dorking & other fowls, having five toes on one foot & four on the other—the cross from the <solid> whole-hoofed & common pig, which with Sir R. Heron had two feet whole & two normally divided—are probably due to this same difficulty of fusion in certain characters.2

I strongly suspect that characters which refuse to blend have first appeared suddenly & perfectly developed: I do not believe that any structure slowly acquired through selection, whether artificial or natural, can be transmitted in this entire or quite negative manner

131 insert

Black, white & other coloured varieties of several kinds of animals have been observed in a state of nature, far oftener than piebald individuals, which shows the same tendency for certain colours to appear, independently of crossing, fully developed or not at all.

132

Prepotency. As with plants, one species of animal seems to be prepotent over an other in impressing its likeness on its hybrid offspring. This according to Flourens3 is the case with the Jackall over the Dog, & seemed to me to be so with one of these hybrids which I examined. I cannot doubt that this is strongly the case with the ass over the horse; the prepotency here running more strongly through the male ass. The Pheasant preponderates over the fowl, in those hybrids which I have seen.1 <but it is most difficult to form any accurate judgment on this head.> In races of our domestic animals, numerous instances can be given of prepotency of one over the other. Godine2 has given a very curious case of a strongly characterised goat-like race of Cape Sheep; the Ram of which was crossed with 12 other breeds, & the offspring were so like the ram, that they could hardly be distinguished from it; the lambs, however of two ewes, of this breed, when crossed for six successive years with a Merino-ram, resembled most closely merinos.3 Sturm states positively from repeated experiments that a Holland Bull //4 [was much less prepotent than a Holland (i.e. Friesian) cow in crosses with Swiss cattle.]

134

Prepotency seems, also, to be characteristic of individuals, of either sex, of the same race; for we can understand in no other way the manner in which marked features are trans mitted in certain families, after marriages with different females. So amongst our domestic animals, certain individuals have been notorious for transmitting their characteristic qualities.—

134 Bis

I have met with several observers who have expressed a strong opinion that when an ancient or naturally formed breed or species is crossed with a modern domesticated one, that the characters of the former preponderate in a cross. That a character which has long been inherited should continue truer than one recently acquired is almost self evident; but when we cross one breed with another, the separate question of prepotency (as we have just seen in the case of Trumpeters) comes into play. Ancient breeds may be prepotent over modern breeds when crossed with them; & I have met with some facts which countenance this view; <but many others seem strongly opposed to it.> the almost monstrous Nata cattle, of S. America before alluded to, have arisen within the last three centuries, but yet are prepotent over other cattle: no breed is more modem or artificial than the Improved Short-horn, yet I observe in all accounts from the continent, that no breed is more potent in impressing its character on other

On the other hand, I have met with several facts which seem strongly opposed to the foregoing supposed rule. Thus,

134 Bis v

This opinion may have partly arisen from a difficulty or impossibility of improving old established breeds in wild mountainous countries by crosses with improved artificial breeds; but the unfavourable conditions for tender animals here come into play. *

135

native breeds, & hence partly its great value for exportation. Drooping ears are no doubt due to domestication, yet in a hybrid from a Jackall & Terrier, which I saw in the Zoological Gardens, though the Jackall preponderated, yet the ears drooped; & this hybrid wagged its tail.

The Penguin Duck is an almost monstrous race, but in some mongrels from this bird & the common Duck, & in some remarkable hybrids with the Aegyptian goose in the Zoological Gardens (Tadorna Aegyptiaca), the upright & singular gait of this breed seemed to me equally to prevail.

It is notorious that both a species & race can be reduced or absorbed by repeated crossings with a distinct species or race. The number of generations required probably differs in different species, as we have seen to be the case with plants; & this probably accounts for the great diversity of opinion of breeders on this head; some saying that 12 or 13 or even 20 generations are required, others more commonly saying (as with plants) that 8 or 9 amply suffice. It is certain that the rate of reduction differs according as the male or female of the reducing race

135 insert

Mr. Hewitt1 describes hybrids from the Cock Pheasant & five differently coloured breeds of the common hen; & these hybrids differed greatly from each other, in colour, showing that the Cock pheasant had no marked degree of prepotency at least in colour over these several domesticated varieties.

1. Poultry Book by Mr. Tegetmeier 1856 p. 124.

136

be used;2 & this naturally follows from prepotency running in animals more in one sex than the other. I know of no facts showing that one strongly marked race can reduce another either more quickly or more slowly than one species can another.. The great grandchildren of the common pheasant were reduced to the perfect appearance of P. versicolor by three crosses; & so it has been with a mongrel Fantail-pigeon, reduced by a pure Fan-tail, but as Boitard & Corbie have remarked three or four more generations would be required to make sure of the purity of the offspring.

Since Lord Morton's famous case of the Quagga & Arabian mare, it has been universally admitted that the subsequent offspring of a female mammal is affected in an incomprehensible manner by a first cross from a distinct species. And there is copious evidence that this is likewise the case between different races of animals & even different individuals of the same race!

Finally it seems that the same rules hold generally good for crossed plants & animals whether distinct species, or varieties or merely individuals of the same race are crossed in regard to their resemblance to their parents, their variability

136v

2 Note used in Domestic Animals Chapter 15. Crossing

137

prepotency & reduction. <The same rules seem to hold good for hybrids & mongrels & for the offspring of individuals of the same race when they present recognizable differences.> The chief difference between plants & animals seems to be that prepotency, or an extraordinary power of impressing resemblance & consequently of reduction, commonly differs in animals to a large extent in the two sexes of the same race or species.

The chief difference between mongrels & hybrids, whether in plants or animals, seems to be that in the first generations hybrids are generally uniform in character, but this is not universally the case; nor are all mongrels very variable.

In the succeeding generations, mongrels probably are more variable than hybrids in the corresponding generation but this does not seem to me, considering their origin, to be at all surprising; occasionally, though very rarely, both hybrids & mongrels keep true in their successive generations. <Lastly when races are crossed, those characters, which, as I suspect have originally appeared suddenly, are much oftener transmitted either perfectly or not at all to their crossed offspring than in the case of crossed species, of which the characters have not been formed in this sudden and monstrous manner.>

137 insert

Domestic races of animals often have characters which have originally appeared in a sudden & monstrous manner, & these I suspect are frequently transmitted either perfectly or not at all to their mongrel offspring; & this seems rarely to be the case with hybrids from between two species, neither of which have been modified by domestication.

138

Summary on Chapter.—Weighing all the evidence given in this chapter, I think we must conclude that the first cross between the forms, called by naturalists species, & their hybrid offspring are with rare exceptions sterile in some degree. But when closely related forms are tried, the sterility so graduates away, that the two best observers, who ever lived differ diametrically whether or not they are perfectly fertile together. The attempt to measure fertility by so nice a process as counting the number of seeds is seriously interfered with the ill-effects of manipulation & the seclusion of the specimen & culture. We probably see the importance of the latter, in the difference of the results obtained by Gaertner & Herbert when experimentising on the same two species. Gaertner's failure to obtain full fertility between many forms, ranked by all the best botanists as varieties must shake our confidence in his conclusion that species are universally in some degree sterile together. The increasing sterility of hybrids when naturally self fertilised for successive generations may I think be safely attributed in large part to the ill-effects of close interbreeding; for it seems otherwise impossible to understand, how artificial fertilisation, in itself injurious, should aid, as Gaertner asserts it does, their fertility.

139

But it is, I think, impossible to admit that species when first crossed & their offspring are invariably sterile together even in the slightest degree, after Herberts repeated observations on, for instance Crinum, in which he found that every crossed ovule produced a seed, which never happened with natural fertilisation. Nor should we pass over the apparently perfect fertility of several florists flowers, as in the genus Rhododendron, which have of late years been crossed in so complicated a manner: in these cases alone have the experiment been quite fairly tried, for here there has been excellent culture, no manipulation, & natural intercrossing allowed in whole beds of the same hybrid kind.

With animals, though first crosses can in some cases be so easily effected, yet it cannot be said that the perfect fertility of the successive generations of any one hybrid has been fully established; though we have no reason to doubt it between certain pheasants. But how few experiments have been tried between closely allied species, both of which will breed perfectly under confinement. Nor must the ill effects of close interbreeding be overlooked in those few cases in which hybrids all descended from the same two parents have been bred for some generations inter se.

140

When we consider that the Fertility both of first crosses & of hybrids graduates in different cases from zero to the normal degree of perfection; that it is in all such cases eminently susceptible of favourable & unfavourable conditions; that it fails easier on the male than female side; that the degree, in the same hybrid in the first & successive generations, is innately highly variable; that the degree does not run closely parallel with the amount of systematic difference between the two parent-species, even within the limits of the same genus; that it often differs widely in reciprocal crosses from between the same two species; that there is no absolute relation between the facility of getting a first cross & the degree of fertility of the hybrid offspring; that there is no close relation between the likeness, whether abnormal & occasional, or normal & regular, of the hybrid to one of its parent-species; when we consider all these & other such singular facts, I cannot believe, that the lessened fertility of first crosses & of hybrids has been a specially endowed quality to prevent those forms which coexist in the same country from becoming blended together. The complexity

140v

[deleted start to 140]

141

& singularity of the rules seem to me to indicate that they are incidental on differences in the organs & functions of Reproduction in different species, in some degree analogous to the differences in the organs & functions of vegetation, on which the capacity for grafting depends, & which, I presume, no one would suppose were specially endowed to facilitate or prevent one tree being grafted on another. No doubt, differences in different species in the organs & functions of Reproduction & of Vegetation will follow pretty closely as a general rule, systematic affinity, which means the sum of all resemblances of all kinds, & not of any one particular organ or function.

In first crosses, the sterility must depend on different causes: in plants widely different the pollen-tube does not penetrate the stigmatic tissue; in more closely related plants, though reaching the ovule it does not cause an embryo to germinate: in other cases, a large part of the sterility depends on the early death of the fertilised embryo: in these cases we can no more offer any explanation than we can why some trees belonging to the same genus cannot be grafted on each other.

142

In hybrids, the cause of sterility is widely different from that of the first cross between two pure species; for in hybrids the male & female sexual products are manifestly deteriorated, whereas in the parent species they were of course perfect. When we bear in mind the numerous facts given in our third Chapter, showing how eminently susceptible the sexual functions are to any change in the conditions of life, I think we need feel little surprised that hybrids with their double constitution & laws of growth confounded together should have their fertility affected in a somewhat analogous manner, whatever may be the conditions of life to which they may be exposed. Nor is it, I think, any more strange, that the sterility of hybrids & of first crosses though depending on very different causes, should run in some degree parallel, than that both should run in some, but far less close, degree parallel with the capacity of grafting; for all these depend though in different ways, on the amount of

143

resemblance & dissemblance in the species experimentised on.

The most surprising circumstance in our whole present subject is the almost universal fertility of the most distinct varieties when crossed, as in the case of the several breeds of fowls & pigeons. But in plants we have as good evidence, as we can ever get on the slighter degrees of sterility between closely related species; that varieties are in some few cases slightly infertile together. In the case of animals, it may be as Pallas hypothetically concluded, that domestication eliminates the tendency to sterility in crosses; if this be so, we could not expect that sterility should appear between the most distinct varieties if produced under domestication. We have seen that slight changes in the conditions of life are favourable to fertility, though greater changes or. changes of another kind affect in so decided a manner the reproductive functions; hence it is not

144

surprising that crosses within certain limits should be favourable to fertility, though beyond such limits they should cause sterility. But it is surprising & could never have been a priori anticipated, that crosses between, for instance, such extremely distinct breeds, as those of the pigeons, should not have been in the least degree unfavourable to their fertility: yet we should bear in mind, that man by his selection,—the great agency in the production of domestic breeds—has no power or wish to modify either directly, or indirectly by selecting consti-tutional differences in the reproductive system; & it is on differences in the reproductive system, on which the sterility of species when crossed, seems incidentally to depend.

Finally in all other respects, besides fertility, the offspring of species & varieties seem to follow, often absolutely & always very closely, the same laws namely in their resemblance to their parents, their variability, equal or unequal reciprocity prepotency, reduction &e:—but it is needless to sum up the conclusions, just arrived at, on these heads. I will conclude by remarking, that

145

if the difference in fertility between species & varieties when crossed, has been rightly urged as so very important a distinction; then on the other hand, so close a resemblance in their progeny in all other respects, ought to weigh with us as an argument of not a little weight that Species & Varieties have not had an essentially different origin.

Citation: John van Wyhe, ed. 2002-. The Complete Work of Charles Darwin Online. (http://darwin-online.org.uk/)