The storage of a supply of nectar in a protected place is manifestly connected with the visits of insects. So is the position which the stamens and pistils occupy, either permanently or at the proper period through their own movements; for when mature they invariably stand in the pathway leading to the nectary. The shape of the nectary and of the adjoining parts are likewise related to the particular kinds of insects which habitually visit the flowers; this has been well shown by Hermann Muller by his comparison of lowland species which are chiefly visited by bees, with alpine species belonging to the same genera which are visited by butterflies. (10/5. 'Nature' 1874 page 110, 1875 page 190, 1876 pages 210, 289.) Flowers may also be adapted to certain kinds of insects, by secreting nectar particularly attractive to them, and unattractive to other kinds; of which fact Epipactis latifolia offers the most striking instance known to me, as it is visited exclusively by wasps. Structures also exist, such as the hairs within the corolla of the fox glove (Digitalis), which apparently serve to exclude insects that are not well fitted to bring pollen from one flower to another. (10/6. Belt 'The Naturalist in Nicaragua' 1874 page 132.) I need say nothing here of the endless contrivances, such as the viscid glands attached to the pollen-masses of the Orchideae and Asclepiadae, or the viscid or roughened state of the pollen-grains of many plants, or the irritability of their stamens which move when touched by insects etc.—as all these contrivances evidently favour or ensure cross-fertilisation.

All ordinary flowers are so far open that insects can force an entrance into them, notwithstanding that some, like the Snapdragon (Antirrhinum), various Papilionaceous and Fumariaceous flowers, are in appearance closed. It cannot be maintained that their openness is necessary for fertility, as cleistogene flowers which are permanently closed yield a full complement of seeds. Pollen contains much nitrogen and phosphorus—the two most precious of all the elements for the growth of plants—but in the case of most open flowers, a large quantity of pollen is consumed by pollen-devouring insects, and a large quantity is destroyed during long-continued rain. With many plants this latter evil is guarded against, as far as is possible, by the anthers opening only during dry weather (10/7. Mr. Blackley observed that the ripe anthers of rye did not dehisce whilst kept under a bell-glass in a damp atmosphere, whilst other anthers exposed to the same temperature in the open air dehisced freely. He also found much more pollen adhering to the sticky slides, which were attached to kites and sent high up in the atmosphere, during the first fine and dry days after wet weather, than at other times: 'Experimental Researches on Hay Fever' 1873 page 127.)—by the position and form of some or all of the petals,—by the presence of hairs, etc., and as Kerner has shown in his interesting essay, by the movements of the petals or of the whole flower during cold and wet weather. (10/8. 'Die Schutzmittel des Pollens' 1873.) In order to compensate the loss of pollen in so many ways, the anthers produce a far larger amount than is necessary for the fertilisation of the same flower. I know this from my own experiments on Ipomoea, given in the Introduction; and it is still more plainly shown by the astonishingly small quantity produced by cleistogene flowers, which lose none of their pollen, in comparison with that produced by the open flowers borne by the same plants; and yet this small quantity suffices for the fertilisation of all their numerous seeds. Mr. Hassall took pains in estimating the number of pollen-grains produced by a flower of the Dandelion (Leontodon), and found the number to be 243,600, and in a Paeony 3,654,000 grains. (10/9. 'Annals and Magazine of Natural History' volume 8 1842 page 108.) The editor of the 'Botanical Register' counted the ovules in the flowers of Wistaria sinensis, and carefully estimated the number of pollen-grains, and he found that for each ovule there were 7000 grains. (10/10. Quoted in 'Gardeners' Chronicle' 1846 page 771.) With Mirabilis, three or four of the very large pollen-grains are sufficient to fertilise an ovule; but I do not know how many grains a flower produces. With Hibiscus, Kolreuter found that sixty grains were necessary to fertilise all the ovules of a flower, and he calculated that 4863 grains were produced by a single flower, or eighty-one times too many. With Geum urbanum, however, according to Gartner, the pollen is only ten times too much. (10/11. Kolreuter 'Vorlaufige Nachricht' 1761 page 9. Gartner 'Beitrage zur Kenntniss' etc. page 346.) As we thus see that the open state of all ordinary flowers, and the consequent loss of much pollen, necessitate the development of so prodigious an excess of this precious substance, why, it may be asked, are flowers always left open? As many plants exist throughout the vegetable kingdom which bear cleistogene flowers, there can hardly be a doubt that all open flowers might easily have been converted into closed ones. The graduated steps by which this process could have been effected may be seen at the present time in Lathyrus nissolia, Biophytum sensitivum, and several other plants. The answer to the above question obviously is, that with permanently closed flowers there could be no cross-fertilisation.

The frequency, almost regularity, with which pollen is transported by insects from flower to flower, often from a considerable distance, well deserves attention. (10/12. An experiment made by Kolreuter 'Forsetsung' etc. 1763 page 69, affords good evidence on this head. Hibiscus vesicarius is strongly dichogamous, its pollen being shed before the stigmas are mature. Kolreuter marked 310 flowers, and put pollen from other flowers on their stigmas every day, so that they were thoroughly fertilised; and he left the same number of other flowers to the agency of insects. Afterwards he counted the seeds of both lots: the flowers which he had fertilised with such astonishing care produced 11,237 seeds, whilst those left to the insects produced 10,886; that is, a less number by only 351; and this small inferiority is fully accounted for by the insects not having worked during some days, when the weather was cold with continued rain.) This is best shown by the impossibility in many cases of raising two varieties of the same species pure, if they grow at all near together; but to this subject I shall presently return; also by the many cases of hybrids which have appeared spontaneously both in gardens and a state of nature. With respect to the distance from which pollen is often brought, no one who has had any experience would expect to obtain pure cabbage-seed, for instance, if a plant of another variety grew within two or three hundred yards. An accurate observer, the late Mr. Masters of Canterbury, assured me that he once had his whole stock of seeds "seriously affected with purple bastards," by some plants of purple kale which flowered in a cottager's garden at the distance of half a mile; no other plant of this variety growing any nearer. (10/13. Mr. W.C. Marshall caught no less than seven specimens of a moth (Cucullia umbratica) with the pollinia of the butterfly-orchis (Habenaria chlorantha) sticking to their eyes, and, therefore, in the proper position for fertilising the flowers of this species, on an island in Derwentwater, at the distance of half a mile from any place where this plant grew: 'Nature' 1872 page 393.) But the most striking case which has been recorded is that by M. Godron, who shows by the nature of the hybrids produced that Primula grandiflora must have been crossed with pollen brought by bees from P. officinalis, growing at the distance of above two kilometres, or of about one English mile and a quarter. (10/14. 'Revue des Sc. Nat.' 1875 page 331.)

All those who have long attended to hybridisation, insist in the strongest terms on the liability of castrated flowers to be fertilised by pollen brought from distant plants of the same species. (10/15. See, for instance, the remarks by Herbert 'Amaryllidaceae' 1837 page 349. Also Gartner's strong expressions on this subject in his 'Bastarderzeugung' 1849 page 670 and 'Kenntniss der Befruchtung' 1844 pages 510, 573. Also Lecoq 'De la Fecondation' etc. 1845 page 27. Some statements have been published during late years of the extraordinary tendency of hybrid plants to revert to their parent forms; but as it is not said how the flowers were protected from insects, it may be suspected that they were often fertilised with pollen brought from a distance from the parent-species.) The following case shows this in the clearest manner: Gartner, before he had gained much experience, castrated and fertilised 520 flowers on various species with pollen of other genera or other species, but left them unprotected; for, as he says, he thought it a laughable idea that pollen should be brought from flowers of the same species, none of which grew nearer than between 500 and 600 yards. (10/16. 'Kenntniss der Befruchtung' pages 539, 550, 575, 576.) The result was that 289 of these 520 flowers yielded no seed, or none that germinated; the seed of 29 flowers produced hybrids, such as might have been expected from the nature of the pollen employed; and lastly, the seed of the remaining 202 flowers produced perfectly pure plants, so that these flowers must have been fertilised by pollen brought by insects from a distance of between 500 and 600 yards. (10/17. Henschel's experiments quoted by Gartner 'Kenntniss' etc. page 574, which are worthless in all other respects, likewise show how largely flowers are intercrossed by insects. He castrated many flowers on thirty-seven species, belonging to twenty-two genera, and put on their stigmas either no pollen, or pollen from distinct genera, yet they all seeded, and all the seedlings raised from them were of course pure.) It is of course possible that some of these 202 flowers might have been fertilised by pollen left accidentally in them when they were castrated; but to show how improbable this is, I may add that Gartner, during the next eighteen years, castrated no less than 8042 flowers and hybridised them in a closed room; and the seeds from only seventy of these, that is considerably less than 1 per cent, produced pure or unhybridised offspring. (10/18. 'Kenntniss' etc. pages 555, 576.)

From the various facts now given, it is evident that most flowers are adapted in an admirable manner for cross-fertilisation. Nevertheless, the greater number likewise present structures which are manifestly adapted, though not in so striking a manner, for self-fertilisation. The chief of these is their hermaphrodite condition; that is, their including within the same corolla both the male and female reproductive organs. These often stand close together and are mature at the same time; so that pollen from the same flower cannot fail to be deposited at the proper period on the stigma. There are also various details of structure adapted for self-fertilisation. (10/19. Hermann Muller 'Die Befruchtung' etc. page 448.) Such structures are best shown in those curious cases discovered by Hermann Muller, in which a species exists under two forms,—one bearing conspicuous flowers fitted for cross-fertilisation, and the other smaller flowers fitted for self-fertilisation, with many parts in the latter slightly modified for this special purpose. (10/20. 'Nature' 1873 pages 44, 433.)

As two objects in most respects opposed, namely, cross-fertilisation and self-fertilisation, have in many cases to be gained, we can understand the co-existence in so many flowers of structures which appear at first sight unnecessarily complex and of an opposed nature. We can thus understand the great contrast in structure between cleistogene flowers, which are adapted exclusively for self-fertilisation, and ordinary flowers on the same plant, which are adapted so as to allow of at least occasional cross-fertilisation. (10/21. Fritz Muller has discovered in the animal kingdom 'Jenaische Zeitschr.' B. 4 page 451, a case curiously analogous to that of the plants which bear cleistogene and perfect flowers. He finds in the nests of termites in Brazil, males and females with imperfect wings, which do not leave the nests and propagate the species in a cleistogene manner, but only if a fully-developed queen after swarming does not enter the old nest. The fully-developed males and females are winged, and individuals from distinct nests can hardly fail often to intercross. In the act of swarming they are destroyed in almost infinite numbers by a host of enemies, so that a queen may often fail to enter an old nest; and then the imperfectly developed males and females propagate and keep up the stock.) The former are always minute, completely closed, with their petals more or less rudimentary and never brightly coloured; they never secrete nectar, never are odoriferous, have very small anthers which produce only a few grains of pollen, and their stigmas are but little developed. Bearing in mind that some flowers are cross-fertilised by the wind (called anemophilous by Delpino), and others by insects (called entomophilous), we can further understand, as was pointed out by me several years ago, the great contrast in appearance between these two classes of flowers. (10/22. 'Journal of the Linnean Society' volume 7 Botany 1863 page 77.) Anemophilous flowers resemble in many respects cleistogene flowers, but differ widely in not being closed, in producing an extraordinary amount of pollen which is always incoherent, and in the stigma often being largely developed or plumose. We certainly owe the beauty and odour of our flowers and the storage of a large supply of honey to the existence of insects.

ON THE RELATION BETWEEN THE STRUCTURE AND CONSPICUOUSNESS OF FLOWERS, THE VISITS OF INSECTS, AND THE ADVANTAGES OF CROSS-FERTILISATION.

It has already been shown that there is no close relation between the number of seeds produced by flowers when crossed and self-fertilised, and the degree to which their offspring are aaffected by the two processes. I have also given reasons for believing that the inefficiency of a plant's own pollen is in most cases an incidental result, or has not been specially acquired for the sake of preventing self-fertilisation. On the other hand, there can hardly be a doubt that dichogamy, which prevails according to Hildebrand in the greater number of species (10/23. 'Die Geschlecter Vertheiling' etc. page 32.),—that the heterostyled condition of certain plants,—and that many mechanical structures—have all been acquired so as both to check self-fertilisation and to favour cross-fertilisation. The means for favouring cross-fertilisation must have been acquired before those which prevent self-fertilisation; as it would manifestly be injurious to a plant that its stigma should fail to receive its own pollen, unless it had already become well adapted for receiving pollen from another individual. It should be observed that many plants still possess a high power of self-fertilisation, although their flowers are excellently constructed for cross-fertilisation—for instance, those of many papilionaceous species.

It may be admitted as almost certain that some structures, such as a narrow elongated nectary, or a long tubular corolla, have been developed in order that certain kinds of insects alone should obtain the nectar. These insects would thus find a store of nectar preserved from the attacks of other insects; and they would thus be led to visit frequently such flowers and to carry pollen from one to the other. (10/24. See the interesting discussion on this subject by Hermann Muller, 'Die Befruchtung' etc. page 431.) It might perhaps have been expected that plants having their flowers thus peculiarly constructed would profit in a greater degree by being crossed, than ordinary or simple flowers; but this does not seem to hold good. Thus Tropaeolum minus has a long nectary and an irregular corolla, whilst Limnanthes douglasii has a regular flower and no proper nectary, yet the crossed seedlings of both species are to the self-fertilised in height as 100 to 79. Salvia coccinea has an irregular corolla, with a curious apparatus by which insects depress the stamens, while the flowers of Ipomoea are regular; and the crossed seedlings of the former are in height to the self-fertilised as 100 to 76, whilst those of the Ipomoea are as 100 to 77. Fagopyrum is dimorphic, and Anagallis collina is non-dimorphic, and the crossed seedlings of both are in height to the self-fertilised as 100 to 69.

With all European plants, excepting the comparatively rare anemophilous kinds, the possibility of distinct individuals intercrossing depends on the visits of insects; and Hermann Muller has proved by his valuable observations, that large conspicuous flowers are visited much more frequently and by many more kinds of insects, than are small inconspicuous flowers. He further remarks that the flowers which are rarely visited must be capable of self-fertilisation, otherwise they would quickly become extinct. (10/25. 'Die Befruchtung' etc. page 426. 'Nature' 1873 page 433.) There is, however, some liability to error in forming a judgment on this head, from the extreme difficulty of ascertaining whether flowers which are rarely or never visited during the day (as in the above given case of Fumaria capreolata) are not visited by small nocturnal Lepidoptera, which are known to be strongly attracted by sugar. (10/26. In answer to a question by me, the editor of an entomological journal writes—"The Depressariae, as is notorious to every collector of Noctuae, come very freely to sugar, and no doubt naturally visit flowers:" the 'Entomologists' Weekly Intelligencer' 1860 page 103.) The two lists given in the early part of this chapter support Muller's conclusion that small and inconspicuous flowers are completely self-fertile: for only eight or nine out of the 125 species in the two lists come under this head, and all of these were proved to be highly fertile when insects were excluded. The singularly inconspicuous flowers of the Fly Ophrys (O. muscifera), as I have elsewhere shown, are rarely visited by insects; and it is a strange instance of imperfection, in contradiction to the above rule, that these flowers are not self-fertile, so that a large proportion of them do not produce seeds. The converse of the rule that plants bearing small and inconspicuous flowers are self-fertile, namely, that plants with large and conspicuous flowers are self-sterile, is far from true, as may be seen in our second list of spontaneously self-fertile species; for this list includes such species as Ipomoea purpurea, Adonis aestivalis, Verbascum thapsus, Pisum sativum, Lathyrus odoratus, some species of Papaver and of Nymphaea, and others.

The rarity of the visits of insects to small flowers, does not depend altogether on their inconspicuousness, but likewise on the absence of some sufficient attraction; for the flowers of Trifolium arvense are extremely small, yet are incessantly visited by hive and humble-bees, as are the small and dingy flowers of the asparagus. The flowers of Linaria cymbalaria are small and not very conspicuous, yet at the proper time they are freely visited by hive-bees. I may add that, according to Mr. Bennett, there is another and quite distinct class of plants which cannot be much frequented by insects, as they flower either exclusively or often during the winter, and these seem adapted for self-fertilisation, as they shed their pollen before the flowers expand. (10/27. 'Nature' 1869 page 11.)

That many flowers have been rendered conspicuous for the sake of guiding insects to them is highly probable or almost certain; but it may be asked, have other flowers been rendered inconspicuous so that they may not be frequently visited, or have they merely retained a former and primitive condition? If a plant were much reduced in size, so probably would be the flowers through correlated growth, and this may possibly account for some cases; but the size and colour of the corolla are both extremely variable characters, and it can hardly be doubted that if large and brightly-coloured flowers were advantageous to any species, these could be acquired through natural selection within a moderate lapse of time, as indeed we see with most alpine plants. Papilionaceous flowers are manifestly constructed in relation to the visits of insects, and it seems improbable, from the usual character of the group, that the progenitors of the genera Vicia and Trifolium produced such minute and unattractive flowers as those of V. hirsuta and T. procumbens. We are thus led to infer that some plants either have not had their flowers increased in size, or have actually had them reduced and purposely rendered inconspicuous, so that they are now but little visited by insects. In either case they must also have acquired or retained a high degree of self-fertility.

If it became from any cause advantageous to a species to have its capacity for self-fertilisation increased, there is little difficulty in believing that this could readily be effected; for three cases of plants varying in such a manner as to be more fertile with their own pollen than they originally were, occurred in the course of my few experiments, namely, with Mimulus, Ipomoea, and Nicotiana. Nor is there any reason to doubt that many kinds of plants are capable under favourable circumstances of propagating themselves for very many generations by self-fertilisation. This is the case with the varieties of Pisum sativum and of Lathyrus odoratus which are cultivated in England, and with Ophrys apifera and some other plants in a state of nature. Nevertheless, most or all of these plants retain structures in an efficient state which cannot be of the least use excepting for cross-fertilisation. We have also seen reason to suspect that self-fertilisation is in some peculiar manner beneficial to certain plants; but if this be really the case, the benefit thus derived is far more than counter-balanced by a cross with a fresh stock or with a slightly different variety.

Notwithstanding the several considerations just advanced, it seems to me highly improbable that plants bearing small and inconspicuous flowers have been or should continue to be subjected to self-fertilisation for a long series of generations. I think so, not from the evil which manifestly follows from self-fertilisation, in many cases even in the first generation, as with Viola tricolor, Sarothamnus, Nemophila, Cyclamen, etc.; nor from the probability of the evil increasing after several generations, for on this latter head I have not sufficient evidence, owing to the manner in which my experiments were conducted. But if plants bearing small and inconspicuous flowers were not occasionally intercrossed, and did not profit by the process, all their flowers would probably have been rendered cleistogene, as they would thus have largely benefited by having to produce only a small quantity of safely-protected pollen. In coming to this conclusion, I have been guided by the frequency with which plants belonging to distinct orders have been rendered cleistogene. But I can hear of no instance of a species with all its flowers rendered permanently cleistogene. Leersia makes the nearest approach to this state; but as already stated, it has been known to produce perfect flowers in one part of Germany. Some other plants of the cleistogene class, for instance Aspicarpa, have failed to produce perfect flowers during several years in a hothouse; but it does not follow that they would fail to do so in their native country, any more than with Vandellia, which with me produced only cleistogene flowers during certain years. Plants belonging to this class commonly bear both kinds of flowers every season, and the perfect flowers of Viola canina yield fine capsules, but only when visited by bees. We have also seen that the seedlings of Ononis minutissima, raised from the perfect flowers fertilised with pollen from another plant, were finer than those from self-fertilised flowers; and this was likewise the case to a certain extent with Vandellia. As therefore no species which at one time bore small and inconspicuous flowers has had all its flowers rendered cleistogene, I must believe that plants now bearing small and inconspicuous flowers profit by their still remaining open, so as to be occasionally intercrossed by insects. It has been one of the greatest oversights in my work that I did not experimentise on such flowers, owing to the difficulty of fertilising them, and to my not having seen the importance of the subject. (10/28. Some of the species of Solanum would be good ones for such experiments, for they are said by Hermann Muller 'Befruchtung' page 434, to be unattractive to insects from not secreting nectar, not producing much pollen, and not being very conspicuous. Hence probably it is that, according to Verlot 'Production des Varieties' 1865 page 72, the varieties of "les aubergines et les tomates" (species of Solanum) do not intercross when they are cultivated near together; but it should be remembered that these are not endemic species. On the other hand, the flowers of the common potato (S. tuberosum), though they do not secrete nectar Kurr 'Bedeutung der Nektarien' 1833 page 40, yet cannot be considered as inconspicuous, and they are sometimes visited by diptera (Muller), and, as I have seen, by humble-bees. Tinzmann (as quoted in 'Gardeners' Chronicle' 1846 page 183, found that some of the varieties did not bear seed when fertilised with pollen from the same variety, but were fertile with that from another variety.)

It should be remembered that in two of the cases in which highly self-fertile varieties appeared amongst my experimental plants, namely, with Mimulus and Nicotiana, such varieties were greatly benefited by a cross with a fresh stock or with a slightly different variety; and this likewise was the case with the cultivated varieties of Pisum sativum and Lathyrus odoratus, which have been long propagated by self-fertilisation. Therefore until the contrary is distinctly proved, I must believe that as a general rule small and inconspicuous flowers are occasionally intercrossed by insects; and that after long-continued self-fertilisation, if they are crossed with pollen brought from a plant growing under somewhat different conditions, or descended from one thus growing, their offspring would profit greatly. It cannot be admitted, under our present state of knowledge, that self-fertilisation continued during many successive generations is ever the most beneficial method of reproduction.

THE MEANS WHICH FAVOUR OR ENSURE FLOWERS BEING FERTILISED WITH POLLEN FROM A DISTINCT PLANT.

We have seen in four cases that seedlings raised from a cross between flowers on the same plant, even on plants appearing distinct from having been propagated by stolons or cuttings, were not superior to seedlings from self-fertilised flowers; and in a fifth case (Digitalis) superior only in a slight degree. Therefore we might expect that with plants growing in a state of nature a cross between the flowers on distinct individuals, and not merely between the flowers on the same plant, would generally or often be effected by some means. The fact of bees and of some Diptera visiting the flowers of the same species as long as they can, instead of promiscuously visiting various species, favours the intercrossing of distinct plants. On the other hand, insects usually search a large number of flowers on the same plant before they fly to another, and this is opposed to cross-fertilisation. The extraordinary number of flowers which bees are able to search within a very short space of time, as will be shown in a future chapter, increases the chance of cross-fertilisation; as does the fact that they are not able to perceive without entering a flower whether other bees have exhausted the nectar. For instance, Hermann Muller found that four-fifths of the flowers of Lamium album which a humble-bee visited had been already exhausted of their nectar. (10/29. 'Die Befruchtung' etc. page 311.) In order that distinct plants should be intercrossed, it is of course indispensable that two or more individuals should grow near one another; and this is generally the case. Thus A. de Candolle remarks that in ascending a mountain the individuals of the same species do not commonly disappear near its upper limit quite gradually, but rather abruptly. This fact can hardly be explained by the nature of the conditions, as these graduate away in an insensible manner, and it probably depends in large part on vigorous seedlings being produced only as high up the mountain as many individuals can subsist together.

With respect to dioecious plants, distinct individuals must always fertilise each other. With monoecious plants, as pollen has to be carried from flower to flower, there will always be a good chance of its being carried from plant to plant. Delpino has also observed the curious fact that certain individuals of the monoecious walnut (Juglans regia) are proterandrous, and others proterogynous, and these will reciprocally fertilise each other. (10/30. 'Ult. Osservazioni' etc. part 2 fasc 2 page 337.) So it is with the common nut (Corylus avellana) (10/31. 'Nature' 1875 page 26.), and, what is more surprising, with some few hermaphrodite plants, as observed by Hermann Muller. (10/32. 'Die Befruchtung' etc. pages 285, 339.) These latter plants cannot fail to act on each other like dimorphic or trimorphic species, in which the union of two individuals is necessary for full and normal fertility. With ordinary hermaphrodite species, the expansion of only a few flowers at the same time is one of the simplest means for favouring the intercrossing of distinct individuals; but this would render the plants less conspicuous to insects, unless the flowers were of large size, as in the case of several bulbous plants. Kerner thinks that it is for this object that the Australian Villarsia parnassifolia produces daily only a single flower. (10/33. 'Die Schutzmittel' etc page 23.) Mr. Cheeseman also remarks, that as certain Orchids in New Zealand which require insect-aid for their fertilisation bear only a single flower, distinct plants cannot fail to intercross. (10/34. 'Transactions of the New Zealand Institute' volume 5 1873 page 356.)

Dichogamy, which prevails so extensively throughout the vegetable kingdom, much increases the chance of distinct individuals intercrossing. With proterandrous species, which are far more ccommon than proterogynous, the young flowers are exclusively male in function, and the older ones exclusively female; and as bees habitually alight low down on the spikes of flowers in order to crawl upwards, they get dusted with pollen from the uppermost flowers, which they carry to the stigmas of the lower and older flowers on the next spike which they visit. The degree to which distinct plants will thus be intercrossed depends on the number of spikes in full flower at the same time on the same plant. With proterogynous flowers and with depending racemes, the manner in which insects visit the flowers ought to be reversed in order that distinct plants should be intercrossed. But this whole subject requires further investigation, as the great importance of crosses between distinct individuals, instead of merely between distinct flowers, has hitherto been hardly recognised.

In some few cases the special movements of certain organs almost ensure pollen being carried from plant to plant. Thus with many orchids, the pollen-masses after becoming attached to the head or proboscis of an insect do not move into the proper position for striking the stigma, until ample time has elapsed for the insect to fly to another plant. With Spiranthes autumnalis, the pollen-masses cannot be applied to the stigma until the labellum and rostellum have moved apart, and this movement is very slow. (10/35. 'The Various Contrivances by which British and Foreign Orchids are fertilised' first edition page 128.) With Posoqueria fragrans (one of the Rubiaceae) the same end is gained by the movement of a specially constructed stamen, as described by Fritz Muller.

We now come to a far more general and therefore more important means by which the mutual fertilisation of distinct plants is effected, namely, the fertilising power of pollen from another variety or individual being greater than that of a plant's own pollen. The simplest and best known case of prepotent action in pollen, though it does not bear directly on our present subject, is that of a plant's own pollen over that from a distinct species. If pollen from a distinct species be placed on the stigma of a castrated flower, and then after the interval of several hours, pollen from the same species be placed on the stigma, the effects of the former are wholly obliterated, excepting in some rare cases. If two varieties are treated in the same manner, the result is analogous, though of directly opposite nature; for pollen from any other variety is often or generally prepotent over that from the same flower. I will give some instances: the pollen of Mimulus luteus regularly falls on the stigma of its own flower, for the plant is highly fertile when insects are excluded. Now several flowers on a remarkably constant whitish variety were fertilised without being castrated with pollen from a yellowish variety; and of the twenty-eight seedlings thus raised, every one bore yellowish flowers, so that the pollen of the yellow variety completely overwhelmed that of the mother-plant. Again, Iberis umbellata is spontaneously self-fertile, and I saw an abundance of pollen from their own flowers on the stigmas; nevertheless, of thirty seedlings raised from non-castrated fflowers of a crimson variety crossed with pollen from a pink variety, twenty-four bore pink flowers, like those of the male or pollen-bearing parent.

In these two cases flowers were fertilised with pollen from a distinct variety, and this was shown to be prepotent by the character of the offspring. Nearly similar results often follow when two or more self-fertile varieties are allowed to grow near one another and are visited by insects. The common cabbage produces a large number of flowers on the same stalk, and when insects are excluded these set many capsules, moderately rich in seeds. I planted a white Kohl-rabi, a purple Kohl-rabi, a Portsmouth broccoli, a Brussels sprout, and a Sugar-loaf cabbage near together and left them uncovered. Seeds collected from each kind were sown in separate beds; and the majority of the seedlings in all five beds were mongrelised in the most complicated manner, some taking more after one variety, and some after another. The effects of the Kohl-rabi were particularly plain in the enlarged stems of many of the seedlings. Altogether 233 plants were raised, of which 155 were mongrelised in the plainest manner, and of the remaining 78 not half were absolutely pure. I repeated the experiment by planting near together two varieties of cabbage with purple-green and white-green lacinated leaves; and of the 325 seedlings raised from the purple-green variety, 165 had white-green and 160 purple-green leaves. Of the 466 seedlings raised from the white-green variety, 220 had purple-green and 246 white-green leaves. These cases show how largely pollen from a neighbouring variety of the cabbage effaces the action of the plant's own pollen. We should bear in mind that pollen must be carried by the bees from flower to flower on the same large branching stem much more abundantly than from plant to plant; and in the case of plants the flowers of which are in some degree dichogamous, those on the same stem would be of different ages, and would thus be as ready for mutual fertilisation as the flowers on distinct plants, were it not for the prepotency of pollen from another variety. (10/36. A writer in the 'Gardeners' Chronicle' 1855 page 730, says that he planted a bed of turnips (Brassica rapa) and of rape (B. napus) close together, and sowed the seeds of the former. The result was that scarcely one seedling was true to its kind, and several closely resembled rape.)

Several varieties of the radish (Raphanus sativus), which is moderately self-fertile when insects are excluded, were in flower at the same time in my garden. Seed was collected from one of them, and out of twenty-two seedlings thus raised only twelve were true to their kind. (10/37. Duhamel as quoted by Godron 'De l'Espece' tome 2 page 50, makes an analogous statement with respect to this plant.)

The onion produces a large number of flowers, all crowded together into a large globular head, each flower having six stamens; so that the stigmas receive plenty of pollen from their own and the adjoining anthers. Consequently the plant is fairly self-fertile when protected from insects. A blood-red, silver, globe and Spanish onion were planted near together; and seedlings were raised from each kind in four separate beds. In all the beds mongrels of various kinds were numerous, except amongst the ten seedlings from the blood-red onion, which included only two. Altogether forty-six seedlings were raised, of which thirty-one had been plainly crossed.

A similar result is known to follow with the varieties of many other plants, if allowed to flower near together: I refer here only to species which are capable of fertilising themselves, for if this be not the case, they would of course be liable to be crossed by any other variety growing near. Horticulturists do not commonly distinguish between the effects of variability and intercrossing; but I have collected evidence on the natural crossing of varieties of the tulip, hyacinth, anemone, ranunculus, strawberry, Leptosiphon androsaceus, orange, rhododendron and rhubarb, all of which plants I believe to be self-fertile. (10/38. With respect to tulips and some other flowers, see Godron 'De l'Espece' tome 1 page 252. For anemones 'Gardeners' Chronicle' 1859 page 98. For strawberries see Herbert in 'Transactions of the Horticultural Society' volume 4 page 17. The same observer elsewhere speaks of the spontaneous crossing of rhododendrons. Gallesio makes the same statement with respect to oranges. I have myself known extensive crossing to occur with the common rhubarb. For Leptosiphon, Verlot 'Des Varieties' 1865 page 20. I have not included in my list the Carnation, Nemophila, or Antirrhinum, the varieties of which are known to cross freely, because these plants are not always self-fertile. I know nothing about the self-fertility of Trollius Lecoq 'De la Fecondation' 1862 page 93, Mahonia, and Crinum, in which genera the species intercross largely. With respect to Mahonia it is now scarcely possible to procure in this country pure specimens of M. aquifolium or repens; and the various species of Crinum sent by Herbert 'Amaryllidaceae' page 32, to Calcutta, crossed there so freely that pure seed could not be saved.) Much other indirect evidence could be given with respect to the extent to which varieties of the same species spontaneously intercross.

Gardeners who raise seed for sale are compelled by dearly bought experience to take extraordinary precautions against intercrossing. Thus Messrs. Sharp "have land engaged in the growth of seed in no less than eight parishes." The mere fact of a vast number of plants belonging to the same variety growing together is a considerable protection, as the chances are strong in favour of plants of the same variety intercrossing; and it is in chief part owing to this circumstance, that certain villages have become famous for pure seed of particular varieties. (10/39. With respect to Messrs. Sharp see 'Gardeners' Chronicle' 1856 page 823. Lindley's 'Theory of Horticulture' page 319.) Only two trials were made by me to ascertain after how long an interval of time, pollen from a distinct variety would obliterate more or less completely the action of a plant's own pollen. The stigmas in two lately expanded flowers on a variety of cabbage, called Ragged Jack, were well covered with pollen from the same plant. After an interval of twenty-three hours, pollen from the Early Barnes Cabbage growing at a distance was placed on both stigmas; and as the plant was left uncovered, pollen from other flowers on the Ragged Jack would certainly have been left by the bees during the next two or three days on the same two stigmas. Under these circumstances it seemed very unlikely that the pollen of the Barnes cabbage would produce any effect; but three out of the fifteen plants raised from the two capsules thus produced were plainly mongrelised: and I have no doubt that the twelve other plants were affected, for they grew much more vigorously than the self-fertilised seedlings from the Ragged Jack planted at the same time and under the same conditions. Secondly, I placed on several stigmas of a long-styled cowslip (Primula veris) plenty of pollen from the same plant, and after twenty-four hours added some from a short-styled dark-red Polyanthus, which is a variety of the cowslip. From the flowers thus treated thirty seedlings were raised, and all these without exception bore reddish flowers; so that the effect of the plant's own pollen, though placed on the stigmas twenty-four hours previously, was quite destroyed by that of the red variety. It should, however, be observed that these plants are dimorphic, and that the second union was a legitimate one, whilst the first was illegitimate; but flowers illegitimately fertilised with their own pollen yield a moderately fair supply of seeds.

We have hitherto considered only the prepotent fertilising power of pollen from a distinct variety over a plants' own pollen,—both kinds of pollen being placed on the same stigma. It is a much more remarkable fact that pollen from another individual of the same variety is prepotent over a plant's own pollen, as shown by the superiority of the seedlings raised from a cross of this kind over seedlings from self-fertilised flowers. Thus in Tables 7/A, B, and C, there are at least fifteen species which are self-fertile when insects are excluded; and this implies that their stigmas must receive their own pollen; nevertheless, most of the seedlings which were raised by fertilising the non-castrated flowers of these fifteen species with pollen from another plant were greatly superior, in height, weight, and fertility, to the self-fertilised offspring. (10/40. These fifteen species consist of Brassica oleracea, Reseda odorata and lutea, Limnanthes douglasii, Papaver vagum, Viscaria oculata, Beta vulgaris, Lupinus luteus, Ipomoea purpurea, Mimulus luteus, Calceolaria, Verbascum thapsus, Vandellia nummularifolia, Lactuca sativa, and Zea mays.) For instance, with Ipomoea purpurea every single intercrossed plant exceeded in height its self-fertilised opponent until the sixth generation; and so it was with Mimulus luteus until the fourth generation. Out of six pairs of crossed and self-fertilised cabbages, every one of the former was much heavier than the latter. With Papaver vagum, out of fifteen pairs, all but two of the crossed plants were taller than their self-fertilised opponents. Of eight pairs of Lupinus luteus, all but two of the crossed were taller; of eight pairs of Beta vulgaris all but one; and of fifteen pairs of Zea mays all but two were taller. Of fifteen pairs of Limnanthes douglasii, and of seven pairs of Lactuca sativa, every single crossed plant was taller than its self-fertilised opponent. It should also be observed that in these experiments no particular care was taken to cross-fertilise the flowers immediately after their expansion; it is therefore almost certain that in many of these cases some pollen from the same flower will have already fallen on and acted on the stigma.

There can hardly be a doubt that several other species of which the crossed seedlings are more vigorous than the self-fertilised, as shown in Tables 7/A, 7/B and 7/C, besides the above fifteen, must have received their own pollen and that from another plant at nearly the same time; and if so, the same remarks as those just given are applicable to them. Scarcely any result from my experiments has surprised me so much as this of the prepotency of pollen from a distinct individual over each plant's own pollen, as proved by the greater constitutional vigour of the crossed seedlings. The evidence of prepotency is here deduced from the comparative growth of the two lots of seedlings; but we have similar evidence in many cases from the much greater fertility of the non-castrated flowers on the mother-plant, when these received at the same time their own pollen and that from a distinct plant, in comparison with the flowers which received only their own pollen.

From the various facts now given on the spontaneous intercrossing of varieties growing near together, and on the effects of cross-fertilising flowers which are self-fertile and have not been castrated, we may conclude that pollen brought by insects or by the wind from a distinct plant will generally prevent the action of pollen from the same flower, even though it may have been applied some time before; and thus the intercrossing of plants in a state of nature will be greatly favoured or ensured.

The case of a great tree covered with innumerable hermaphrodite flowers seems at first sight strongly opposed to the belief in the frequency of intercrosses between distinct individuals. The flowers which grow on the opposite sides of such a tree will have been exposed to somewhat different conditions, and a cross between them may perhaps be in some degree beneficial; but it is not probable that it would be nearly so beneficial as a cross between flowers on distinct trees, as we may infer from the inefficiency of pollen taken from plants which have been propagated from the same stock, though growing on separate roots. The number of bees which frequent certain kinds of trees when in full flower is very great, and they may be seen flying from tree to tree more frequently than might have been expected. Nevertheless, if we consider how numerous are the flowers, for instance, on a horse-chestnut or lime-tree, an incomparably larger number of flowers must be fertilised by pollen brought from other flowers on the same tree, than from flowers on a distinct tree. But we should bear in mind that with the horse-chestnut, for instance, only one or two of the several flowers on the same peduncle produce a seed; and that this seed is the product of only one out of several ovules within the same ovarium. Now we know from the experiments of Herbert and others that if one flower is fertilised with pollen which is more efficient than that applied to the other flowers on the same peduncle, the latter often drop off (10/41. 'Variation under Domestication' chapter 17 2nd edition volume 2 page 120.); and it is probable that this would occur with many of the self-fertilised flowers on a large tree, if other and adjoining flowers were cross-fertilised. Of the flowers annually produced by a great tree, it is almost certain that a large number would be self-fertilised; and if we assume that the tree produced only 500 flowers, and that this number of seeds were requisite to keep up the stock, so that at least one seedling should hereafter struggle to maturity, then a large proportion of the seedlings would necessarily be derived from self-fertilised seeds. But if the tree annually produced 50,000 flowers, of which the self-fertilised dropped off without yielding seeds, then the cross-fertilised flowers might yield seeds in sufficient number to keep up the stock, and most of the seedlings would be vigorous from being the product of a cross between distinct individuals. In this manner the production of a vast number of flowers, besides serving to entice numerous insects and to compensate for the accidental destruction of many flowers by spring-frosts or otherwise, would be a very great advantage to the species; and when we behold our orchard-trees covered with a white sheet of bloom in the spring, we should not falsely accuse nature of wasteful expenditure, though comparatively little fruit is produced in the autumn.

ANEMOPHILOUS PLANTS.

The nature and relations of plants which are fertilised by the wind have been admirably discussed by Delpino and Hermann Muller; and I have already made some remarks on the structure of their flowers in contrast with those of entomophilous species. (10/42. Delpino 'Ult. Osservazioni sulla Dicogamia' part 2 fasc. 1 1870 and 'Studi sopra un Lignaggio anemofilo' etc. 1871. Hermann Muller 'Die Befruchtung' etc. pages 412, 442. Both these authors remark that plants must have been anemophilous before they were entomophilous. Hermann Muller further discusses in a very interesting manner the steps by which entomophilous flowers became nectariferous and gradually acquired their present structure through successive beneficial changes.) There is good reason to believe that the first plants which appeared on this earth were cryptogamic; and judging from what now occurs, the male fertilising element must either have possessed the power of spontaneous movement through the water or over damp surfaces, or have been carried by currents of water to the female organs. That some of the most ancient plants, such as ferns, possessed true sexual organs there can hardly be a doubt; and this shows, as Hildebrand remarks, at how early a period the sexes were separated. (10/43. 'Die Geschlechter-Vertheilung' 1867 pages 84-90.) As soon as plants became phanerogamic and grew on the dry ground, if they were ever to intercross, it would be indispensable that the male fertilising element should be transported by some means through the air; and the wind is the simplest means of transport. There must also have been a period when winged insects did not exist, and plants would not then have been rendered entomophilous. Even at a somewhat later period the more specialised orders of the Hymenoptera, Lepidoptera, and Diptera, which are now chiefly concerned with the transport of pollen, did not exist. Therefore the earliest terrestrial plants known to us, namely, the Coniferae and Cycadiae, no doubt were anemophilous, like the existing species of these same groups. A vestige of this early state of things is likewise shown by some other groups of plants which are anemophilous, as these on the whole stand lower in the scale than entomophilous species.

There is no great difficulty in understanding how an anemophilous plant might have been rendered entomophilous. Pollen is a nutritious substance, and would soon have been discovered and devoured by insects; and if any adhered to their bodies it would have been carried from the anthers to the stigma of the same flower, or from one flower to another. One of the chief characteristics of the pollen of anemophilous plants is its incoherence; but pollen in this state can adhere to the hairy bodies of insects, as we see with some Leguminosae, Ericaceae, and Melastomaceae. We have, however, better evidence of the possibility of a transition of the above kind in certain plants being now fertilised partly by the wind and partly by insects. The common rhubarb (Rheum rhaponticum) is so far in an intermediate condition, that I have seen many Diptera sucking the flowers, with much pollen adhering to their bodies; and yet the pollen is so incoherent, that clouds of it are emitted if the plant be gently shaken on a sunny day, some of which could hardly fail to fall on the large stigmas of the neighbouring flowers. According to Delpino and Hermann Muller, some species of Plantago are in a similar intermediate condition. (10/44. 'Die Befruchtung' etc. page 342.)

Although it is probable that pollen was aboriginally the sole attraction to insects, and although many plants now exist whose flowers are frequented exclusively by pollen-devouring insects, yet the great majority secrete nectar as the chief attraction. Many years ago I suggested that primarily the saccharine matter in nectar was excreted as a waste product of chemical changes in the sap; and that when the excretion happened to occur within the envelopes of a flower, it was utilised for the important object of cross-fertilisation, being subsequently much increased in quantity and stored in various ways. (10/45. Nectar was regarded by De Candolle and Dunal as an excretion, as stated by Martinet in 'Annal des Sc. Nat.' 1872 tome 14 page 211.) This view is rendered probable by the leaves of some trees excreting, under certain climatic conditions, without the aid of special glands, a saccharine fluid, often called honey-dew. This is the case with the leaves of the lime; for although some authors have disputed the fact, a most capable judge, Dr. Maxwell Masters, informs me that, after having heard the discussions on this subject before the Horticultural Society, he feels no doubt on this head. The leaves, as well as the cut stems, of the manna ash (Fraxinus ornus) secrete in a like manner saccharine matter. (10/46. 'Gardeners' Chronicle' 1876 page 242.) According to Treviranus, so do the upper surfaces of the leaves of Carduus arctioides during hot weather. Many analogous facts could be given. (10/47. Kurr 'Untersuchungen uber die Bedeutung der Nektarien' 1833 page 115.) There are, however, a considerable number of plants which bear small glands on their leaves, petioles, phyllodia, stipules, bracteae, or flower peduncles, or on the outside of their calyx, and these glands secrete minute drops of a sweet fluid, which is eagerly sought by sugar-loving insects, such as ants, hive-bees, and wasps. (10/48. A large number of cases are given by Delpino in the 'Bulletino Entomologico' Anno 6 1874. To these may be added those given in my text, as well as the excretion of saccharine matter from the calyx of two species of Iris, and from the bracteae of certain Orchideae: see Kurr 'Bedeutung der Nektarien' 1833 pages 25, 28. Belt 'Nicaragua' page 224, also refers to a similar excretion by many epiphytal orchids and passion-flowers. Mr. Rodgers has seen much nectar secreted from the bases of the flower-peduncles of Vanilla. Link says that the only example of a hypopetalous nectary known to him is externally at the base of the flowers of Chironia decussata: see 'Reports on Botany, Ray Society' 1846 page 355. An important memoir bearing on this subject has lately appeared by Reinke 'Gottingen Nachrichten' 1873 page 825, who shows that in many plants the tips of the serrations on the leaves in the bud bear glands which secrete only at a very early age, and which have the same morphological structure as true nectar-secreting glands. He further shows that the nectar-secreting glands on the petioles of Prunus avium are not developed at a very early age, yet wither away on the old leaves. They are homologous with those on the serrations of the blades of the same leaves, as shown by their structure and by transition-forms; for the lowest serrations on the blades of most of the leaves secrete nectar instead of resin (harz).) In the case of the glands on the stipules of Vicia sativa, the excretion manifestly depends on changes in the sap, consequent on the sun shining brightly; for I repeatedly observed that as soon as the sun was hidden behind clouds the secretion ceased, and the hive-bees left the field; but as soon as the sun broke out again, they returned to their feast. (10/49. I published a brief notice of this case in the 'Gardeners' Chronicle' 1855 July 21 page 487, and afterwards made further observations. Besides the hive-bee, another species of bee, a moth, ants, and two kinds of flies sucked the drops of fluid on the stipules. The larger drops tasted sweet. The hive-bees never even looked at the flowers which were open at the same time; whilst two species of humble-bees neglected the stipules and visited only the flowers.) I have observed an analogous fact with the secretion of true nectar in the flowers of Lobelia erinus.

Delpino, however, maintains that the power of secreting a sweet fluid by any extra-floral organ has been in every case specially gained, for the sake of attracting ants and wasps as defenders of the plant against their enemies; but I have never seen any reason to believe that this is so with the three species observed by me, namely, Prunus laurocerasus, Vicia sativa, and V. faba. No plant is so little attacked by enemies of any kind as the common bracken-fern (Pteris aquilina); and yet, as my son Francis has discovered, the large glands at the bases of the fronds, but only whilst young, excrete much sweetish fluid, which is eagerly sought by innumerable ants, chiefly belonging to Myrmica; and these ants certainly do not serve as a protection against any enemy. Delpino argues that such glands ought not to be considered as excretory, because if they were so, they would be present in every species; but I cannot see much force in this argument, as the leaves of some plants excrete sugar only during certain states of the weather. That in some cases the secretion serves to attract insects as defenders of the plant, and may have been developed to a high degree for this special purpose, I have not the least doubt, from the observations of Delpino, and more especially from those of Mr. Belt on Acacia sphaerocephala, and on passion-flowers. This acacia likewise produces, as an additional attraction to ants, small bodies containing much oil and protoplasm, and analogous bodies are developed by a Cecropia for the same purpose, as described by Fritz Muller. (10/50. Mr. Belt 'The Naturalist in Nicaragua' 1874 page 218, has given a most interesting account of the paramount importance of ants as defenders of the above Acacia. With respect to the Cecropia see 'Nature' 1876 page 304. My son Francis has described the microscopical structure and development of these wonderful food-bodies in a paper read before the Linnean Society.)

The excretion of a sweet fluid by glands seated outside of a flower is rarely utilised as a means for cross-fertilisation by the aid of insects; but this occurs with the bracteae of the Marcgraviaceae, as the late Dr. Cruger informed me from actual observation in the West Indies, and as Delpino infers with much acuteness from the relative position of the several parts of their flowers. (10/51. 'Ult. Osservaz. Dicogamia' 1868-69 page 188.) Mr. Farrer has also shown that the flowers of Coronilla are curiously modified, so that bees may fertilise them whilst sucking the fluid secreted from the outside of the calyx. (10/52. 'Nature' 1874 page 169.) It further appears probable from the observations of the Reverend W.A. Leighton, that the fluid so abundantly secreted by glands on the phyllodia of the Australian Acacia magnifica, which stand near the flowers, is connected with their fertilisation. (10/53. 'Annals and Magazine of Natural History' volume 16 1865 page 14. In my work on the 'Fertilisation of Orchids' and in a paper subsequently published in the 'Annals and Magazine of Natural History' it has been shown that although certain kinds of orchids possess a nectary, no nectar is actually secreted by it; but that insects penetrate the inner walls and suck the fluid contained in the intercellular spaces. I further suggested, in the case of some other orchids which do not secrete nectar, that insects gnawed the labellum; and this suggestion has since been proved true. Hermann Muller and Delpino have now shown that some other plants have thickened petals which are sucked or gnawed by insects, their fertilisation being thus aided. All the known facts on this head have been collected by Delpino in his 'Ult. Osserv.' part 2 fasc. 2 1875 pages 59-63.)

The amount of pollen produced by anemophilous plants, and the distance to which it is often transported by the wind, are both surprisingly great. Mr. Hassall found that the weight of pollen produced by a single plant of the Bulrush (Typha) was 144 grains. Bucketfuls of pollen, chiefly of Coniferae and Gramineae, have been swept off the decks of vessels near the North American shore; and Mr. Riley has seen the ground near St. Louis, in Missouri, covered with pollen, as if sprinkled with sulphur; and there was good reason to believe that this had been transported from the pine-forests at least 400 miles to the south. Kerner has seen the snow-fields on the higher Alps similarly dusted; and Mr. Blackley found numerous pollen-grains, in one instance 1200, adhering to sticky slides, which were sent up to a height of from 500 to 1000 feet by means of a kite, and then uncovered by a special mechanism. It is remarkable that in these experiments there were on an average nineteen times as many pollen-grains in the atmosphere at the higher than at the lower levels. (10/54. For Mr. Hassall's observations see 'Annals and Magazine of Natural History' volume 8 1842 page 108. In the 'North American Journal of Science' January 1842, there is an account of the pollen swept off the decks of a vessel. Riley 'Fifth Report on the Noxious Insects of Missouri' 1873 page 86. Kerner 'Die Schutzmittel des Pollens' 1873 page 6. This author has also seen a lake in the Tyrol so covered with pollen, that the water no longer appeared blue. Mr. Blackley 'Experimental Researches on Hay-fever' 1873 pages 132, 141-152.) Considering these facts, it is not so surprising as it at first appears that all, or nearly all, the stigmas of anemophilous plants should receive pollen brought to them by mere chance by the wind. During the early part of summer every object is thus dusted with pollen; for instance, I examined for another purpose the labella of a large number of flowers of the Fly Ophrys (which is rarely visited by insects), and found on all very many pollen-grains of other plants, which had been caught by their velvety surfaces.

The extraordinary quantity and lightness of the pollen of anemophilous plants are no doubt both necessary, as their pollen has generally to be carried to the stigmas of other and often distant flowers; for, as we shall soon see, most anemophilous plants have their sexes separated. The fertilisation of these plants is generally aided by the stigmas being of large size or plumose; and in the case of the Coniferae, by the naked ovules secreting a drop of fluid, as shown by Delpino. Although the number of anemophilous species is small, as the author just quoted remarks, the number of individuals is large in comparison with that of entomophilous species. This holds good especially in cold and temperate regions, where insects are not so numerous as under a warmer climate, and where consequently entomophilous plants are less favourably situated. We see this in our forests of Coniferae and other trees, such as oaks, beeches, birches, ashes, etc.; and in the Gramineae, Cyperaceae, and Juncaceae, which clothe our meadows and swamps; all these trees and plants being fertilised by the wind. As a large quantity of pollen is wasted by anemophilous plants, it is surprising that so many vigorous species of this kind abounding with individuals should still exist in any part of the world; for if they had been rendered entomophilous, their pollen would have been transported by the aid of the senses and appetites of insects with incomparably greater safety than by the wind. That such a conversion is possible can hardly be doubted, from the remarks lately made on the existence of intermediate forms; and apparently it has been effected in the group of willows, as we may infer from the nature of their nearest allies. (10/55. Hermann Muller 'Die Befruchtung' etc. page 149.)

It seems at first sight a still more surprising fact that plants, after having been once rendered entomophilous, should ever again have become anemophilous; but this has occasionally though rarely occurred, for instance, with the common Poterium sanguisorba, as may be inferred from its belonging to the Rosaceae. Such cases are, however, intelligible, as almost all plants require to be occasionally intercrossed; and if any entomiphilous species ceased to be visited by insects, it would probably perish unless it were rendered anemophilous. A plant would be neglected by insects if nectar failed to be secreted, unless indeed a large supply of attractive pollen was present; and from what we have seen of the excretion of saccharine fluid from leaves and glands being largely governed in several cases by climatic influences, and from some few flowers which do not now secrete nectar still retaining coloured guiding-marks, the failure of the secretion cannot be considered as a very improbable event. The same result would follow to a certainty, if winged insects ceased to exist in any district, or became very rare. Now there is only a single plant in the great order of the Cruciferae, namely, Pringlea, which is anemophilous, and this plant is an inhabitant of Kerguelen Land, where there are hardly any winged insects, owing probably, as was suggested by me in the case of Madeira, to the risk which they run of being blown out to sea and destroyed. (10/56. The Reverend A.E. Eaton in 'Proceedings of the Royal Society' volume 23 1875 page 351.)

A remarkable fact with respect to anemophilous plants is that they are often diclinous, that is, they are either monoecious with their sexes separated on the same plant, or dioecious with their sexes on distinct plants. In the class Monoecia of Linnaeus, Delpino shows that the species of twenty-eight genera are anemophilous, and of seventeen genera entomophilous. (10/57. 'Studi sopra un Lignaggio anemofilo delle Compositae' 1871.) The larger proportion of entomophilous genera in this latter class is probably the indirect result of insects having the power of carrying pollen to another and sometimes distant plant much more securely than the wind. In the above two classes taken together there are thirty-eight anemophilous and thirty-six entomophilous genera; whereas in the great mass of hermaphrodite plants the proportion of anemophilous to entomophilous genera is extremely small. The cause of this remarkable difference may be attributed to anemophilous plants having retained in a greater degree than the entomophilous a primordial condition, in which the sexes were separated and their mutual fertilisation effected by means of the wind. That the earliest and lowest members of the vegetable kingdom had their sexes separated, as is still the case to a large extent, is the opinion of a high authority, Nageli. (10/58. 'Entstehung und Begriff der Naturhist. Art' 1865 page 22.) It is indeed difficult to avoid this conclusion, if we admit the view, which seems highly probable, that the conjugation of the Algae and of some of the simplest animals is the first step towards sexual reproduction; and if we further bear in mind that a greater and greater degree of differentiation between the cells which conjugate can be traced, thus leading apparently to the development of the two sexual forms. (10/59. See the interesting discussion on this whole subject by O. Butschli in his 'Studien uber die ersten Entwickelungsvorgange der Eizelle; etc. 1876 pages 207-219. Also Engelmann "Ueber Entwickelung von Infusorien" 'Morphol. Jahrbuch' B. 1 page 573. Also Dr. A. Dodel "Die Kraushaar-Algae" 'Pringsheims Jahrbuch f. Wiss. Bot.' B. 10.) We have also seen that as plants became more highly developed and affixed to the ground, they would be compelled to be anemophilous in order to intercross. Therefore all plants which have not since been greatly modified, would tend still to be both diclinous and anemophilous; and we can thus understand the connection between these two states, although they appear at first sight quite disconnected. If this view is correct, plants must have been rendered hermaphrodites at a later though still very early period, and entomophilous at a yet later period, namely, after the development of winged insects. So that the relationship between hermaphroditism and fertilisation by means of insects is likewise to a certain extent intelligible.

Why the descendants of plants which were originally dioecious, and which therefore profited by always intercrossing with another individual, should have been converted into hermaphrodites, may perhaps be explained by the risk which they ran, especially as long as they were anemophilous, of not being always fertilised, and consequently of not leaving offspring. This latter evil, the greatest of all to any organism, would have been much lessened by their becoming hermaphrodites, though with the contingent disadvantage of frequent self-fertilisation. By what graduated steps an hermaphrodite condition was acquired we do not know. But we can see that if a lowly organised form, in which the two sexes were represented by somewhat different individuals, were to increase by budding either before or after conjugation, the two incipient sexes would be capable of appearing by buds on the same stock, as occasionally occurs with various characters at the present day. The organism would then be in a monoecious condition, and this is probably the first step towards hermaphroditism; for if very simple male and female flowers on the same stock, each consisting of a single stamen or pistil, were brought close together and surrounded by a common envelope, in nearly the same manner as with the florets of the Compositae, we should have an hermaphrodite flower.

There seems to be no limit to the changes which organisms undergo under changing conditions of life; and some hermaphrodite plants, descended as we must believe from aboriginally diclinous plants, have had their sexes again separated. That this has occurred, we may infer from the presence of rudimentary stamens in the flowers of some individuals, and of rudimentary pistils in the flowers of other individuals, for example in Lychnis dioica. But a conversion of this kind will not have occurred unless cross-fertilisation was already assured, generally by the agency of insects; but why the production of male and female flowers on distinct plants should have been advantageous to the species, cross-fertilisation having been previously assured, is far from obvious. A plant might indeed produce twice as many seeds as were necessary to keep up its numbers under new or changed conditions of life; and if it did not vary by bearing fewer flowers, and did vary in the state of its reproductive organs (as often occurs under cultivation), a wasteful expenditure of seeds and pollen would be saved by the flowers becoming diclinous.

A related point is worth notice. I remarked in my Origin of Species that in Britain a much larger proportion of trees and bushes than of herbaceous plants have their sexes separated; and so it is, according to Asa Gray and Hooker, in North America and New Zealand. (10/60. I find in the 'London Catalogue of British Plants' that there are thirty-two indigenous trees and bushes in Great Britain, classed under nine families; but to err on the safe side, I have counted only six species of willows. Of the thirty-two trees and bushes, nineteen, or more than half, have their sexes separated; and this is an enormous proportion compared with other British plants. New Zealand abounds with diclinous plants and trees; and Dr. Hooker calculates that out of about 756 phanerogamic plants inhabiting the islands, no less than 108 are trees, belonging to thirty-five families. Of these 108 trees, fifty-two, or very nearly half, have their sexes more or less separated. Of bushes there are 149, of which sixty-one have their sexes in the same state; whilst of the remaining 500 herbaceous plants only 121, or less than a fourth, have their sexes separated. Lastly, Professor Asa Gray informs me that in the United States there are 132 native trees (belonging to twenty-five families) of which ninety-five (belonging to seventeen families) "have their sexes more or less separated, for the greater part decidedly separated.") It is, however, doubtful how far this rule holds good generally, and it certainly does not do so in Australia. But I have been assured that the flowers of the prevailing Australian trees, namely, the Myrtaceae, swarm with insects, and if they are dichogamous they would be practically diclinous. (10/61. With respect to the Proteaceae of Australia, Mr. Bentham 'Journal of the Linnean Society Botany' volume 13 1871 pages 58, 64, remarks on the various contrivances by which the stigma in the several genera is screened from the action of the pollen from the same flower. For instance, in Synaphea "the stigma is held by the eunuch (i.e., one of the stamens which is barren) safe from all pollution from her brother anthers, and is preserved intact for any pollen that may be inserted by insects and other agencies.") As far as anemophilous plants are concerned, we know that they are apt to have their sexes separated, and we can see that it would be an unfavourable circumstance for them to bear their flowers very close to the ground, as their pollen is liable to be blown high up in the air (10/62. Kerner 'Schutzmittel des Pollens' 1873 page 4.); but as the culms of grasses give sufficient elevation, we cannot thus account for so many trees and bushes being diclinous. We may infer from our previous discussion that a tree bearing numerous hermaphrodite flowers would rarely intercross with another tree, except by means of the pollen of a distinct individual being prepotent over the plants' own pollen. Now the separation of the sexes, whether the plant were anemophilous are entomophilous, would most effectually bar self-fertilisation, and this may be the cause of so many trees and bushes being diclinous. Or to put the case in another way, a plant would be better fitted for development into a tree, if the sexes were separated, than if it were hermaphrodite; for in the former case its numerous flowers would be less liable to continued self-fertilisation. But it should also be observed that the long life of a tree or bush permits of the separation of the sexes, with much less risk of evil from impregnation occasionally failing and seeds not being produced, than in the case of short-lived plants. Hence it probably is, as Lecoq has remarked, that annual plants are rarely dioecious.

Finally, we have seen reason to believe that the higher plants are descended from extremely low forms which conjugated, and that the conjugating individuals differed somewhat from one another,—the one representing the male and the other the female—so that plants were aboriginally dioecious. At a very early period such lowly organised dioecious plants probably gave rise by budding to monoecious plants with the two sexes borne by the same individual; and by a still closer union of the sexes to hermaphrodite plants, which are now much the commonest form. (10/63. There is a considerable amount of evidence that all the higher animals are the descendants of hermaphrodites; and it is a curious problem whether such hermaphroditism may not have been the result of the conjugation of two slightly different individuals, which represented the two incipient sexes. On this view, the higher animals may now owe their bilateral structure, with all their organs double at an early embryonic period, to the fusion or conjugation of two primordial individuals.) As soon as plants became affixed to the ground, their pollen must have been carried by some means from flower to flower, at first almost certainly by the wind, then by pollen-devouring, and afterwards by nectar-seeking insects. During subsequent ages some few entomophilous plants have been again rendered anemophilous, and some hermaphrodite plants have had their sexes again separated; and we can vaguely see the advantages of such recurrent changes under certain conditions.

Dioecious plants, however fertilised, have a great advantage over other plants in their cross-fertilisation being assured. But this advantage is gained in the case of anemophilous species at the expense of the production of an enormous superfluity of pollen, with some risk to them and to entomophilous species of their fertilisation occasionally failing. Half the individuals, moreover, namely, the males, produce no seed, and this might possibly be a disadvantage. Delpino remarks that dioecious plants cannot spread so easily as monoecious and hermaphrodite species, for a single individual which happened to reach some new site could not propagate its kind; but it may be doubted whether this is a serious evil. Monoecious plants can hardly fail to be to a large extent dioecious in function, owing to the lightness of their pollen and to the wind blowing laterally, with the great additional advantage of occasionally or often producing some self-fertilised seeds. When they are also dichogamous, they are necessarily dioecious in function. Lastly, hermaphrodite plants can generally produce at least some self-fertilised seeds, and they are at the same time capable, through the various means specified in this chapter, of cross-fertilisation. When their structure absolutely prevents self-fertilisation, they are in the same relative position to one another as monoecious and dioecious plants, with what may be an advantage, namely, that every flower is capable of yielding seeds.



CHAPTER XI.

THE HABITS OF INSECTS IN RELATION TO THE FERTILISATION OF FLOWERS.

Insects visit the flowers of the same species as long as they can. Cause of this habit. Means by which bees recognise the flowers of the same species. Sudden secretion of nectar. Nectar of certain flowers unattractive to certain insects. Industry of bees, and the number of flowers visited within a short time. Perforation of the corolla by bees. Skill shown in the operation. Hive-bees profit by the holes made by humble-bees. Effects of habit. The motive for perforating flowers to save time. Flowers growing in crowded masses chiefly perforated.

Bees and various other insects must be directed by instinct to search flowers for nectar and pollen, as they act in this manner without instruction as soon as they emerge from the pupa state. Their instincts, however, are not of a specialised nature, for they visit many exotic flowers as readily as the endemic kinds, and they often search for nectar in flowers which do not secrete any; and they may be seen attempting to suck it out of nectaries of such length that it cannot be reached by them. (11/1. See, on this subject Hermann Muller 'Befruchtung' etc. page 427; and Sir J. Lubbock's 'British Wild Flowers' etc. page 20. Muller 'Bienen Zeitung' June 1876 page 119, assigns good reasons for his belief that bees and many other Hymenoptera have inherited from some early nectar-sucking progenitor greater skill in robbing flowers than that which is displayed by insects belonging to the other Orders.) All kinds of bees and certain other insects usually visit the flowers of the same species as long as they can, before going to another species. This fact was observed by Aristotle with respect to the hive-bee more than 2000 years ago, and was noticed by Dobbs in a paper published in 1736 in the Philosophical Transactions. It may be observed by any one, both with hive and humble-bees, in every flower-garden; not that the habit is invariably followed. Mr. Bennett watched for several hours many plants of Lamium album, L. purpureum, and another Labiate plant, Nepeta glechoma, all growing mingled together on a bank near some hives; and he found that each bee confined its visits to the same species. (11/2. 'Nature' 1874 June 4 page 92.) The pollen of these three plants differs in colour, so that he was able to test his observations by examining that which adhered to the bodies of the captured bees, and he found one kind on each bee.

Humble and hive-bees are good botanists, for they know that varieties may differ widely in the colour of their flowers and yet belong to the same species. I have repeatedly seen humble-bees flying straight from a plant of the ordinary red Dictamnus fraxinella to a white variety; from one to another very differently coloured variety of Delphinium consolida and of Primula veris; from a dark purple to a bright yellow variety of Viola tricolor; and with two species of Papaver, from one variety to another which differed much in colour; but in this latter case some of the bees flew indifferently to either species, although passing by other genera, and thus acted as if the two species were merely varieties. Hermann Muller also has seen hive-bees flying from flower to flower of Ranunculus bulbosus and arvensis, and of Trifolium fragiferum and repens; and even from blue hyacinths to blue violets. (11/3. 'Bienen Zeitung' July 1876 page 183.)

Some species of Diptera or flies keep to the flowers of the same species with almost as much regularity as do bees; and when captured they are found covered with pollen. I have seen Rhingia rostrata acting in this manner with the flowers of Lychnis dioica, Ajuga reptans, and Vici sepium. Volucella plumosa and Empis cheiroptera flew straight from flower to flower of Myosotis sylvatica. Dolichopus nigripennis behaved in the same manner with Potentilla tormentilla; and other Diptera with Stellaria holostea, Helianthemum vulgare, Bellis perennis, Veronica hederaefolia and chamoedrys; but some flies visited indifferently the flowers of these two latter species. I have seen more than once a minute Thrips, with pollen adhering to its body, fly from one flower to another of the same kind; and one was observed by me crawling about within a convolvulus with four grains of pollen adhering to its head, which were deposited on the stigma.

Fabricius and Sprengel state that when flies have once entered the flowers of Aristolochia they never escape,—a statement which I could not believe, as in this case the insects would not aid in the cross-fertilisation of the plant; and this statement has now been shown by Hildebrand to be erroneous. As the spathes of Arum maculatum are furnished with filaments apparently adapted to prevent the exit of insects, they resemble in this respect the flowers of Aristolochia; and on examining several spathes, from thirty to sixty minute Diptera belonging to three species were found in some of them; and many of these insects were lying dead at the bottom, as if they had been permanently entrapped. In order to discover whether the living ones could escape and carry pollen to another plant, I tied in the spring of 1842 a fine muslin bag tightly round a spathe; and on returning in an hour's time several little flies were crawling about on the inner surface of the bag. I then gathered a spathe and breathed hard into it; several flies soon crawled out, and all without exception were dusted with arum pollen. These flies quickly flew away, and I distinctly saw three of them fly to another plant about a yard off; they alighted on the inner or concave surface of the spathe, and suddenly flew down into the flower. I then opened this flower, and although not a single anther had burst, several grains of pollen were lying at the bottom, which must have been brought from another plant by one of these flies or by some other insect. In another flower little flies were crawling about, and I saw them leave pollen on the stigmas.