Trans. Roy. Soc. Edin. 37: 203-286 (1892)
XIV.—A Comparison of the Minute Structure of Plant Hybrids with that of their  Parents,
and its Bearing on Biological Problems.
By J. MUIRHEAD MACFARLANE, D.Sc., F.R.S.E. (Plates I.-VIII.)

(Read 4th May and 15th June 1891.)

V. GENERAL SUMMARY OF RESULTS ON SEED HYBRIDS.

We may now briefly recapitulate some of the more evident or naked-eye characters of hybrids, and gradually pass to finer details. It has been demonstrated that in hair production, if the parents possess one or more kinds that are fundamentally similar, but which differ in size, number, and position, the hybrid reproduces these in an intermediate way. Illustrations of this were presented by Geum intermedium, Erica Watsoni, Cypripedium Leeanum, and Masdevallia Chelsoni. But if only one parent possess hairs over a given region the hybrid usually inherits these to half the extent, as in the petals of Dianthus barbatus and some floral parts of Bryanthus erectus. If the hairs of two parents are pretty dissimilar, instead of blending of these in one, the hybrid reproduces each, though reduced in size and number by half. The gland hairs of Saxifraga Andrewsii, the simple and gland hairs of Ribes Culverwellii, and those on the vegetative organs of Bryanthus erectus are examples. The peculiar case of hair distribution in [271] relation to colour formation noticed in the sepal of Cypripedium Leeanum may also be noted here.

In the formation of nectaries as traced in Philageria, Dianthus, Saxifraga, Ribes, &c., the above principles also hold.

The distribution of stomata over any epidermal area has been proved to be a mean between the extremes of the parents, if the stomata of the parents occur over one surface or both, and if the leaves are similar in consistence, but, as in Hedychium Sadlerianum, and to a less degree in Saxifraga Andrewsii, if the stomatic distribution and leaf consistence differ in the parents, this may give rise to correspondingly different results in the hybrid.

In amount of cuticular deposit, and arrangement of it into ridges or other localized growths, hybrids have been proved intermediate between the parents. We may merely recall here the case of Philageria stem, which inherited cuticular ridges from Lapageria, though reduced to half the size, since the Philesia parent was devoid of them.

As WICHURA has already proved for the vegetative leaves of hybrid willows, the venation of hybrid leaves is very uniformly intermediate between those of the parents. Figures are given with this paper of the vegetative leaves of Philageria and Saxifraga, and of the, petals of Dianthus and Geum. The relation of the bundles to special terminations, as in the water stomata of Saxifraga, is in conformity with the venation.

But the growth of tissue in a hybrid which is to determine the outline or angular position which any organ or part of one will assume is intermediate between those of the parents when the latter show traceable differences. Thus the sepals and petals, as also the styles and style-arms, of Geum intermedium, the floral parts as a whole of Saxifraga Andrewsii and Ribes Culverwellii, the frilling of some of the floral parts of Bryanthus and Cypripedium Leeanum are pronounced cases, while minor ones have been referred to.

Turning to minuter anatomical details, every hybrid has yielded a large series of examples which prove that the size, outline, amount of thickening, and localization of growth of cell walls, is, as a rule, intermediate between those of the parents. We have repeatedly stated that as the outcome of growth localization, intercellular spaces of a hybrid are modified in size and shape as are the cells which surround them. Now this clearly demonstrates that the living protoplasm which has formed the cells is so organized in its molecular or micellar constitution that in every cell and over every infinitesimally minute area on its surface where cellulose is to be laid down the balanced effect of both parents is felt.

Equally in the laying down of secondary wall thickenings, whether of a cuticularized, lignified, or colloid nature, numerous citations have been made where the amount and mode of deposition is evenly between the extremes of the parents. Perhaps the most striking case is that of the bundle-sheath cells of Philageria and its parents, where usually five lignified lamellae are traceable in each cell of Lapageria, eleven or twelve in Philesia, and eight or nine in Philageria.[272]

In summarizing as to protoplasm and its modifications as plastids, where considerable differences can be traced in the plastids of two parents the hybrid gives excellent results. Only in a few parent plants have these differences been sufficiently marked to allow of comparison with the hybrid. The leucoplasts in the epidermal cells of the parents of Dianthus Lindsayi are very different in size, while most of the leucoplasts in the hybrid are exactly intermediate, but from careful measurement of lantern projection images of these it has been found that some very nearly resemble those of the female parent. The chromoplasts of the petal cells in Geum intermedium and of the sepal cells in Masdevallia Chelsoni are additional illustrations. Those of the former are very variable in size and number, but this is probably to be explained from its inheriting half of its hereditary features from Geum rivale, which is equally variable as a species. Leaves of corresponding age and position from Saxifraga Andrewsii and its parents have furnished chloroplasts of small size and dark green colour in one parent, of large size and soft emerald green colour in the other, and an intermediate type in the hybrid, though some diverge towards the " Geum" parent in having large chloroplasts.

But the average size, shape, and lamellar deposition in starches of Hedychium hybrids are perhaps the most interesting cases adduced. When we remember that these are bodies formed temporarily as reserve food, and that they are built up by addition of successive micellae through the agency of minute protoplasmic masses or leucoplasts, we have a direct proof that these leucoplasts are themselves fundamentally modified. Their activity in the cells of the hybrid is evinced by the building up of starch grains which, though only of temporary duration in the history of the plant, are so accurately constructed as to be an exact combination in appearance of a half corpuscle of each parent.

Finally, we may recall the facts advanced as to colour, flowering period, chemical combinations, and growth vigour, which, though scanty and fragmentary in their nature, all point to the conclusion that hybrids are intermediate between their parents in general life phenomena.

VI. THE BEARING OF HYBRIDITY ON BIOLOGICAL PROBLEMS.

A wide and attractive field for the biologist is still open in the investigation of plant and animal hybrids. Though much work of a laborious kind has been expended on the plant side, we must regard it merely as the small beginning to an inquiry that will yield results of great value. On the animal side it may truly be said that all the results are in the future. Such being the case, we can scarcely hope to do more at this stage of the inquiry than indicate shortly what seem to be lights cast on certain hitherto doubtful or intricate problems, from a minute study of plant hybrids.

(a) Relative Potency of the Male and Female Sex Elements in the Formation of an Organism.—This problem has greatly occupied the minds of biologists during the last decade, and a solution has only been attempted hitherto from consideration of the [273] behaviour of the sex elements at, and immediately subsequent to, the period of fertilisation. The tissues of hybrids shed a very exact light on the subject. No matter what tissue or set of tissues are chosen, if the cells composing such are tolerably diverse in the parents, one can trace with ease the modifying action which both sex elements have had on them, while these clearly prove to us that each sex element, after union with its complementary sex element, represents potentially half its former individuality, or retains half its former hereditary properties. If one select, for example, a few adjoining cells from the leaf epidermis of Dianthus Grievei and its parents, as figured in Plate IV. figs. 1-3, and compare these, one sees that the average cell of the hybrid is an exact mean between the cells of the parents. On comparing further the epidermal tissue of a dozen hybrids, if one were to be guided alone by the number of epidermal cells and of stomata over a given area, a like conclusion would be reached. In such special cases as the sepaline gland of Philageria and Lapageria, we deal with cells resulting from repeated division of a set of mother cells. To effect upbuilding of the hybrid gland, therefore, proliferation of a set of cells takes place, each of which has a Philesia heredity towards arrest of growth and a Lapageria heredity towards luxuriant cell proliferation, the resultant being a gland built up of half as many cells as that of the one parent.

But VAN BENEDEN* went further than most of his zoological co-workers were prepared to go when he asserted that each cell in an organism is a hermaphrodite structure. To this thesis many subsequently took exception, and with some show of reason perhaps, seeing that no direct proof in individual cell life was forthcoming. But one is forced to accept its absolute correctness from study even of one hybrid. It is this hermaphroditism of the entire hybrid organism which not only impresses on it the structure that the naked eye and microscope reveal, but which causes it to have a life cycle whose successive steps are intermediate between the parent extremes. Thus sufficient facts are in our possession, and will, we hope, be greatly supplemented ere long, to prove that the period of bud-bursting, of leaf-expansion, of flower production, fruit ripening, and other vital phenomena in hybrids are all dependent to a wonderfully exact degree on hereditary inheritance. Naturally, when we make this statement, we wish it to be clearly understood that secondary causes may modify or obscure the exactness of phenomena. Thus every one who is practically conversant with plant life knows the powerful influence which soil, moisture, situation, &c., have in altering the even tenor of a plant's way. It is on this account that we would earnestly desire to have continuous and exhaustive experiments carried out, where every possible care might be taken to eliminate disturbing factors.

(b) Unisexual Heredity.—By this term we would designate the outcome of those observations now recorded which prove that structures found only in one parent, and with no corresponding counterpart in the other, are handed down, though reduced by half. The sepaline honey-gland of Philageria, the small circumstomatic cellular knobs of Saxifraga Andrewsii, the colour patches on the sepal of Cypripedium Leeanum, and the [274] spiral or spiro-reticulate thickenings on cells of Masdevallia Chelsoni, are observed cases.

For the evolutionist these have some value. Whether one adopt the view that environmental surroundings are the main agents in conferring acquired characters, or that these wholly arise by accidental variation, we have strong grounds for believing that these acquired characters are handed clown, though weakened in intensity by half. Nevertheless, if these are of advantage, sexual union of the progeny, coupled with possible further variation along the same line, may retain or even intensify the new character. But the case of Cypripedium Leeanum is of more than ordinary interest, for not only are the colour spots that are present in C. insigne and absent in C. Spicerianum inherited though in less intensity of tint, by the hybrid, but we find a complete absence of hairs where, under ordinary heredity transmission from C. Spicerianum, they should have been formed. Now it has been repeatedly noticed that when a species varies from the normal it seldom does so in one point or structural detail, but a certain variation-wave, so to speak, travels through the entire organism, giving it that combined set of characters which make it rank as a sub-species. The relation between colour production and hair distribution already described not only shows how new characters may be imported into a line of organisms, but how these may even be powerful enough to minimize the normal action of the other parent. A somewhat similar case is that of Saxifraga Andrewsii, in which the circumstomatic knobs inherited from the "Aizoon" parent are to all appearance correlated as a morphological character, with lime secretion by the stomata as a physiological one.

(c) Bisexual Heredity.—The cases of such that have been noticed are few, and do not probably possess great interest apart from hybrid study. We include under this head such an example as Ribes Culverwellii, in which the simple hairs of R. Grossularia and the oil-secreting peltate hairs of R. nigrum are both separately reproduced, though about half as large* as those of the parents. Saxifraga Andrewsii and Carduus Carolorum† likewise have distinct types of hair inherited from both parents. No cases are known to me where internal elements or tissue masses are thus separately reproduced. All the hybrids in which the above has been observed are derived from parents considerably removed in systematic relationship, and the incompatibility of blending the diverse types of hair probably explains their appearance as separate growths.

But the general principle here illustrated on an exaggerated scale is that the offspring of two parents may inherit from each diverse peculiarities which, instead of blending evenly, retain their separate individuality. Future experiment and observation alone will decide for us whether these can be passed down through two, three, or more generations, and till we have the evidence it would be impossible to generalize. [275]

(d) On the Divergence of some Hybrids, or Parts of Hybrids, towards One Parent.—It is undoubted that not a few hybrids show a decided leaning towards one parent, though I consider from examination of several that have thus be described that the number has been considerably over-estimated. With undoubted cases we will now concern ourselves. Regarding these many have asserted that the male parent or male element predominates, and sets up one-sided variation changes. We would readily grant both from perusal of FOCKE'S "Pflanzen-mischlinge" and from direct observation that this is frequently true. But no one can deny that there are many artificial hybrids which do take more after the female parent. Professor BROOKES,* recognising the strength of the former position, has formulated a theory of variation and heredity alike ingenious and plausible. In the present state of our knowledge we would not reject it, but we may still inquire whether some other and simpler explanation cannot be given.

If we view the male and female sex elements of any plant as aggregations of a purely physical but very complex set of substances, it must necessarily follow that if the relative amount, or weight, or combination proportions, in each male or female element varies, variation will result after conjugation, and it will only be where the amount in each conjugating cell is an exact average of the producing organism that a new organism will develop which will show throughout an average combination of the characters of both parents. Now in the struggle for existence which holds among pollen cells and egg cells it will seldom happen that exactly the same amount of sex substance will be formed in exactly similar combinations in each. But of the two, which, we may ask, will vary most? Direct observation proves that it is the male or pollen cell, and the reason for this is obvious. The great majority of flowering plants mature their ovules with the contained egg cells inside cavities where space for growth, and elaborate means for nutrition and protection up to the time of fertilization exists. But the opposite is the case with pollen cells, which are crowded together in the anther cavity, and often obtain nourishment by approximate sustenance through each other. Those therefore nearest the sustentative source will have the advantage, unless of course they are strongly pressed against by some firm bounding wall.

I have been greatly surprised both with the average constancy in size of the egg cell and with the greater variability of the pollen cells in such plants as Lilium, Scilla, and Digitalis. But we now know that the nuclear substance is specially concerned in fertilization, and STRASBURGER has formulated a hypothesis† to account for diversities in hybrids by supposing that the two parents have a different average amount of chromatin substance in their sperm and egg nuclei. We would extend the hypothesis further, and regard the amount as a varying one even in the same parent. Hitherto I have not been able to measure the nuclei of sex cells, but in many vegetative cells there is clear evidence that the variability in size of the nucleus Is very great, and further that there are considerable differences in the size of the nuclei and nucleoli even in adjoining cells. So much is this the case that I have felt quite safe only in comparing the [276] nuclei of Cytisus Adami with those of its parents. Now if this be true of vegetative cell nuclei, there is strong probability that it will equally hold with reproductive nuclei, and accordingly the greater resemblance of any resultant embryo to one parent over another would be satisfactorily explained on a physical basis.

But this does not on first look enable us to explain those cases where local divergences toward either parent occur in a hybrid which otherwise is very evenly balanced. Many examples of this have cropped up in the course of investigation and description. But an application of the same hypothesis in its minuter bearings will clear away most difficulties. If we view a fertilized egg of any plant which is about to segment to form an embryo as being not merely a chemically complex nucleated mass of protoplasm, but as a microcosm in which the orderly-arranged molecules of the conjugated male element have so exactly fitted into and become united with corresponding molecules of the female element that after conjugation co-ordinated groups of molecules are set apart as stem-producers, root-producers, leaf-producers, and hair-producers, we will have done much to clear away obstacles. But physically there is no reason why we may not assume that each cell of the future plant has representative molecules in the apparently simple egg. Now if such be the case it may not unfrequently happen that corresponding groups of molecules from male and female cells do not unite exactly owing to incomplete nutrition or other defect in the maturing of one group. Thus one of the two may in part break down or become weakened and the complemental sexual part thereby give to the resulting tissue a one-sided character. Since the nuclear substance of the male or pollen cell is the one most liable to variable development through over- or under-nutrition, or through advantageous or disadvantageous position, it follows that variation will oftenest have its expression from the male side.

(e) Mechanical or Physiological Obstacles to Fertilization as an Explanation of infertility in some Hybrids.—Great importance has been attached by many to the fact that some parent species which appear even to be nearly related refuse to cross, or only do so on one side, reciprocal crosses being apparently impossible. But as STRASBURGER has well emphasized* a very simple mechanical explanation like that advanced on the animal side by PFLUGER in the case of Amphibians may explain the difficulty. Thus it is possible that the sperm nucleus, or pollen tube containing such, of some plant species may be too large for the receptive area of the egg cell or ovular surface, though the opposite application might prove quite fertile. Similarly the relative length and shape of style, size of pollen grain, strength of pollen coat, amount of mucilage secreted by the stigma, time of ripening of stamens and stigma, must all be studied before we pronounce any attempted hybrid union impossible. Equally simple physiological obstacles connected with colour or some, special chemical production may help in explaining partial or entire sterility; When treating of Cytisus Adami we noted a great abundance of tannin material in the "purpureus" parent, a relatively small quantity in the "laburnum" parent, and an intermediate amount in the hybrid. This extended even [277] to the pollen grains, and gave to each a certain and recognizable tint. Now we know that repeated attempts to cross C. purpureus and C. Laburnum have failed, and it is quite possible that the act action of the abundant tannin material on the stigma or egg cell of C. Laburnum, and vice versa the small amount of it for C. purpureus, may largely explain the failures. Many of the negative hybridization experiments of the past therefore may have less depth of significance than one is inclined on first view to attach to them.

(f) On the Relative Fertility of Hybrids in Relation to Heredity.—The concensus of opinion among the older hybridizers was that very few hybrids were fertile, and that those which were, gradually returned to one of the parent types. During the last twenty or twenty-five years the opinion has been freely criticised, and rightly so, since horticulturists in that period have carried forward experimental hybridization by leaps and bounds, and have imported through intelligent collectors not a few wild plants which they regarded as, and in some instances have proved to be, natural hybrids.

But even though the subject of pollination, with all the marvellous floral adaptations for it, were unknown to such experimenters as KÖLREUTER, KNIGHT, GÄRTNER, WICHURA, and others, the main outcome of their researches can scarcely be set aside, though we may have to give a more liberal interpretation to it in the future. To sum up present-day experiences, it may be said that crosses between species that are nearly related in structure and habit can readily be effected, and the offspring may be largely fertile, at least among certain genera. Crosses between species that differ considerably in form, flower-colour, and habit are more difficult to perform, and the hybrids are largely sterile, while crosses between such divergent species or genera as Dianthus alpinus and barbatus, Saxifraga Geum and Aizoon, Lapageria arid Philesia are almost or wholly sterile.

Now when the pollen and egg cells from each of these three roughly classified groups are examined one finds that a few of those from the first are shrivelled-looking and badly formed; from the second a considerable percentage are thus affected; while from the third it is difficult to get one good pollen grain, and rather difficult to get one well-formed egg-cell, though these do not appear to be so much affected as the pollen grains.

We have repeatedly referred to, and in Plate V. fig. 10 b have illustrated a bad pollen sample. The cells are always smaller, often greatly smaller, than in either parent; the protoplasm is devoid of rich nutritive granules and is scanty in quantity, so that it does not fill the cell cavity, the wall is irregular in outline and imperfectly formed. Shortly, therefore, it may be said that while the vegetative cells of a hybrid can develop gradually into organs that are a blended reproduction of those of the parents, the generative cells fail to receive or to form appropriate protoplasmic material.

Consideration of this causes us to look at the theories that have been advanced to account for heredity. DARWIN'S theory of pangenesis has been put aside as cumbrous and difficult to conceive of in practice, though it explained phenomena of heredity all along the line better than any previously existing view. More applicable, however, does NAGELI'S idioplasmatic theory appear, in spite of gratuitous assumptions that have been [278] urged against it. The fundamental idea animating the pangenetic theory is that the sex-cell's are the cumulative expression of all, the actions and reactions, the integrations and disintegrations which have been associated with the protoplasms up to the time when these sex-cells have been fully formed. NAGELI in expressing the same fundamental idea brought it more into line with modern cell' discovery by assuming that the, nucleoplasm was a continuous network. WEISMANN has objected to NAGELI'S  hypothesis* as follows: "The idioplasm does not form a directly continuous, network throughout the entire body," and "it is perfectly certain that the idioplasm cannot form a continuous network throughout the whole organism if it is seated in the nucleus and not in the cell-body."

But it may well be asked, How do we know that idioplasm, nuclear substance, nucleolar substance, or chromatic substance, is not connected into one network? A dozen years have not passed since the majority of biologists would have rejected the idea of an intercellular network. Our minds should be open to receive fairly any hypothesis, or facts favouring a hypothesis, that may be presented without dogmatising that such cannot be.

We repeat it, then, as an observed fact, that the reproductive cells of hybrids are to a greater or less extent small, imperfect, and badly formed, and that the more divergent the. parent types the more numerous do the imperfect cells become. If with WEISMANN we view each of these as descendants from the germ-plasm of the hybrid egg, why do these fail to mature and continue the hybrid progeny? It may be replied that the susceptible germ-plasms refuse to blend, or blend so imperfectly, that while the blended somatoplasms develop the vegetative part of the hybrid, the germ-plasms break down. But this compels us to assume a greatly more cumbrous state of matters than does the pangenesis theory, for we must suppose that these imperfectly blended masses of germ-plasms are carried up with the growth of the stem, and finally appear at the floral extremities in an aborted state, and that this continues year after year in a hybrid shrub or tree  and we must further assume in the case of Cytisus Adami, that the same is effected by vegetative union of the parts of two parents, without the intervention of sex-cells. It may be urged in the latter case that some germ-plasm cells were mixed up amongst the apparently pure vegetative or somatoplasmic cells, but even if this be granted, it still proves that a hybrid growth can develop apart from sexual union. We believe that a simple and more natural explanation can be given, a short summary of which has already appeared in Nature (vol. 44, 1891).

(g) Vegetable Cell Structure in Relation to Hybridity.—Observations made by me, alike on resting and dividing cells, during the last few years, and preparations which Mr MANN made and kindly showed me, caused me to adhere to my already published views on cell-life, viz., that in the ordinary resting state of an active cell, i.e., one capable of, and at times showing, division, a nucleus with nucleolus and endo-nucleolus are integral parts, and that after division of the cell has ceased proliferation of the inner parts may still go on leading to a multi-endonucleolar, then to a multi-nucleolar, and finally to a multi-nuclear [279] state.* In. certain rare cases, e.g., endosperm cells, proliferation of the nucleolus may produce a temporary multi-nucleolar state, while. the nucleus and cell can divide at. a later period. Behaviour of the endo-nucleolus and nucleolus during division causes me to regard. these as the special cell-centres, and this is well illustrated in species of Spirogyra, where; the whole role of nuclear division is subsidiary to the nucleolus, and is only initiated, subsequent to indications of commencing division in it.† We regard this as the true explanation of division processes in other plant cells. Now in Spirogyra one can readily see that the nucleolar material not only forms the main mass of chromatic substance, but that it is connected by an extremely fine network system with the nuclear membrane, which is. also, chromatic, and during division breaks down to fuse with the radiating threads from the. nucleolus. In re-formation of the daughter nuclei also round the daughter nucleoli, the nuclear membrane gradually reappears, first on the outer poles or nuclear faces, but some of the nucleolar threads can be traced to radiate out through and beyond the nuclear membrane and across the cell-cavity to the pyrenoid centres. Now, in my earlier studies of Spirogyra‡ I was puzzled to understand how these radiating threads that were originally continuous in the nuclear spindle seemed to separate as deposition of the cell-septum took place between. Recent careful study with high powers reveals that from the pyrenoid centres of some bands extremely fine chromatic threads stretch across to, and connect, the pyrenoid centres of other bands. A connection of these from one cell to another has. not as yet been traced, but, apart from observations on other plant-cells which favour it, the strong probability is that such exists, for this network is quite continuous during division up to time of deposition of the cell-partition, and as the latter is laid down by union of micellae from the peripheral protoplasm and from the spindle threads, these may retain delicate continuations of their substance through the formed partition.

We would consider, then, that the nucleolus is the special chromatic and cell centre; that it sends out fine radiating processes—the intranuclear network—which partially fuse externally to constitute the nuclear membrane, the interspaces of the network being occupied by nucleoplasm concerned in metabolic change; that radiating continuations of the chromatic substance pass out beyond the nuclear membrane, and form a network in the protoplasm, while we would suggest for future proof or disproof that they further may be continued through wall pores to form an intercellular chromatic connection.

Not only in Spirogyra but in leaf cells of Dionaea and of Masdevallia radiating chromatic threads have been traced.

The question now arises as to the nature and origin of the chromatic substance. This is pretty generally viewed now by biologists as sexual substance par-excellence, and as being the bearer of hereditary characters. To explain its distribution in each cell, we may consider with most biologists that the simplest plant and animal cells, have no [280] nuclear differentiation, and consist of an apparently simple protoplasmic mass. But the power of movement, of digesting and assimilating food-particles, of retreating from centres of disturbance or irritation, &c., would cause us to inquire whether the apparently undifferentiated mass is not traversed by a fine protoplasmic reticulum of a neuromuscular kind. Such is the view that many have held and still hold.

But unicellular forms that show sexuality show also a nucleus, nucleolus, and endonucleolus, the two last being often and carefully figured by BUTSCHLI, HUXLEY, and others, while we consider their occurrence as universal in all cells of sexual plants and animals. We have, however, already asserted our conviction that the nucleolus is the important cell centre, and we have further proved by hybrid investigation that every cell of an organism is hermaphrodite. Let us suppose, for attempted explanation, that the nucleolus with its radiating chromatic threads is purely sexual, and is made up of the fused chromatic constituents of male and female cells. Let us suppose further that the nucleolar or sexual substance is gathered, round a central differentiation or aggregation of the protoplasmic reticulum, which might he the endonucleolus, and that it sends out radiating chromatic processes along these threads which in part anastomose into a fine chromatic layer—the nuclear membrane—so as to enclose in the interstices of the meshwork system a quantity of nutritive protoplasm which is at once a bed for the nucleolus and a feeder of it. Other radiating chromatic threads continued from the nucleolus and passing beyond the nuclear membrane would ramify minutely through the protoplasm along the threads of the reticulum, giving such appearances as we have traced in Spirogyra, Masdevallia Veitchiana, Ornithogalum pyramidale, and Dionaea.

The hypothesis would enable us to explain much that is at present involved and obscure, while it would also enable us to dispense with the need for germ plasms. It would permit us to entertain the possibility of a comparatively rapid intercommunication of particles, and an even more rapid propagation of external stimuli, from cell to cell, accompanied by change in every molecule reached by these stimuli. The sum-total of these would be expressed in the sex-cells, which are the slowest to mature.

We would thus view a plant as a group of connected hermaphrodite cells, descended from a fertilized egg-cell, and bound together by a fine chromatic ramification, the centre of which in each cell, is the nucleolus. This chromatic system, intimately in contact with the general protoplasm, would receive stimuli and nourishment from it, while the combined action of these and other agents would tell not on one cell or cell-group, but be shared to a greater or less extent, by all.

The above view does not compel us to suppose that the older cells in which the nuclei' are carried round in the protoplasmic current are thus connected, for these have passed the stage of active division, and have their permanent life functions already expressed.

If we apply the above views to explain the frequent sterility of hybrids, a possible, or we may venture to say, a probable hypothesis can be framed. If each reproductive cell of an organism is specialised as an epitome of the, individual which produces it (and in [281] spite of arguments advanced by WEISMANN and his school, we adhere to DARWIN'S widely collected facts and reasonings on them as entirely favouring this), and gathers up the features of that individual in its development and maturation owing to the constant action and reaction between its chromatic substance and that of co-organismal cells, it follows that for the accomplishment of this there must be a certain co-ordination or rhythmic harmony in the motion of the molecules, and an appropriate attraction—chemical or otherwise—in the combining molecules. If otherwise, then instead of integration of molecule to molecule, disintegration or at least an incapacity for union will bold.

But it should here be emphasized that reproductive cells are greatly more concentrated in their history than ordinary vegetative cells, and only attain their full maturity after the active stage has been passed in the last, or, as Professor RYDER has well put it in his suggestive paper on the subject, "Sexuality begins when growth ends."* This does not, however, interfere with the fact that sex-cells are often cut off at a very early period from the vegetative ones, for the former may then undergo, as we know them in many cases to do, a slow maturing process, and be greatly acted on or modified by the latter. Now we know that the most impressionable time in the history both of plants and animals is that of growth—not of maturity—and therefore the experiments which may have been instituted on animals, and such arguments as those bearing on exercierknochen, are practically worthless, because the individuals practised on have not in most cases been treated from the earliest impressionable period, when the substance of the sexual cells is in process of formation. It should be noted also that in the human subject and other mammals the eggs are observable in the Graffian follicles at birth, and yet are not matured and shed till years of slow upbuilding and moulding action have affected them.

If we return now to hybrid production of the more extreme types, though in virtue of the attraction which exists between sexual elements, the original male and female cells from parents of different species—in the absence of cells from the same species—may be capable of uniting, and, in the process, of overcoming the repulsion due to dissimilar co-relative molecules in each, when the attempt is made by all the hermaphrodite cells of the resulting hybrid organism to concentrate representative hermaphrodite groups of molecules, many cases will occur in which these will blend imperfectly, owing to differences in the composition and amount of chemical substances present, or interference and cancelling effects due to unequal propagation of waves of motion between the molecules. Thus many groups of molecules will break down or fail to reach their destination, so that gaps or vacancies will occur in the organic completeness of the pollen or egg cell. It will then have the shrivelled half-empty look so characteristic of hybrid sex-cells that are sterile. In hybrids from more nearly related species the interfering or cancelling effects will be reduced in proportion, and a larger number of sex-cells will have a chance to mature.

(h) Value of Microscopic Characters in the Future Verification of Doubtful Hybrids.—[282] We have advanced reasons, drawn from microscopic study as well as from other points of view, that Bryanthus erectus is a true hybrid, and that its reputed parentage is correct. In the progress of horticulture, forms are continually appearing which are asserted to be hybrids, and similarly as reputed wild species or varieties are being more carefully scrutinised their hybrid nature is at times suggested. The great difficulty in safely determining whether this is so has been the absence of sufficiently marked naked-eye characters in the parents and hybrid. In a valuable contribution to hybridity by Mr MEEHAN* many plants are mentioned which LINNAEUS looked upon as hybrids between species, but which he nevertheless described as species since they freely reproduced themselves. From a rather hasty study of some of these we should be inclined to question LINNAEUS' verdict in their case, but such forms as Trifolium hybridum, present an apparently strong case for the systematist. Armed now with an increased range of characters for comparison, it should be possible to decide whether some at least have not an undoubted relation to the supposed original parents. In such cases, nevertheless, it must be kept in mind that if their origin dates back over a long period such changes may subsequently have been effected in them by variation and selection that the comparison can only be approximate, unless indeed one were to produce the hybrid artificially, and find close microscopic resemblances between the natural and artificial types. In any case we consider it as undoubted that recognition of hybrids from careful microscopic study should now be possible in the great majority of cases.

(i) The Possible Origin of Species from Hybrids.—When the literature of hybridity perused from the historical standpoint one cannot fail to be impressed with the more liberal spirit in which the subject is treated, and with the increasing belief in hybrids that are tolerably, or even very fertile. Specially is this so on the botanical side, but a paper by the late FRANCIS DAY,† from the zoologists' standpoint, proves that great interest will centre round the subject at no distant date. Hitherto it may be said that authorities, with few exceptions, have declared wholly against the view that hybrids may be sufficiently fertile, and their progeny sufficiently strong and adaptable to be fitted for survival, not to say increase, in the struggle for existence. The admirable experiments conducted by WICHURA on willows go far to prove, one would think, that by the fourth or fifth generation enfeeblement and decay become so marked that continued production fails. But against this is to be placed the fact that many of our horticulturists are ardent believers in the continued fertility of hybrids, as witness the article by Professor MEEHAN already cited, though we believe that an over-sanguine expectation is sometimes entertained under this head.

When one finds the undoubted hybrid between Geum rivale and G. urbanum frequently described by systematists as a species, and that in many places the hybrid is nearly or quite as abundant as either parent, that it freely produces good seeds, and further that it has, as. we have already indicated, many points of superiority as a combined [283] organism which neither parent possesses separately, we have good reason for the exercise of caution before pronouncing decisively against species production from hybrids. Still it must be confessed that our experimental statistics are so meagre and unsatisfactory that no final opinion can be given. In saying this we do not in the least under-estimate the conclusions arrived at by KÖLREUTER, GÄRTNER, and their successors, but the wonderful effects of altered conditions of soil, climate, and situation in giving relative fertility to hybrids that were formerly regarded as sterile were not fully recognized in their day, and are only now being to some degree appreciated.

Strong reasons can be urged for the prosecution of careful and prolonged investigations on the subject in our botanic gardens, experimental stations, and private gardens.

Though a partial investigation has already been made, no account has been taken in this paper of second or third hybrids, or of hybrids in which, by reciprocal crossing, different results are got. These will be treated of in a subsequent paper.

The author gladly acknowledges the help received from various quarters since commencement of this investigation. Valuable supplies of material have been received through the kindness of the Directors and Curators of Kew, Glasnevin, Edinburgh and Glasgow Botanic Gardens, and from many private sources. The constant aid extended by Professor BALFOUR and Mr LINDSAY deserve special mention, while Mr RICHARDSON and Mr FORGAN have given valuable help and advice on micro-photographic details. Through the generosity of the Botanical Committee of the Royal Society, a grant was given for purchase of material and illustration of the paper.

October 1891.

INDEX