"On the natural order of plants called Proteaceae, also published as "On the Proteaceae of Jussieu", was a paper written by Robert Brown on the taxonomy of the plant family Proteaceae. It was read to the Linnean Society of London in the first quarter of 1809, and published in March 1810. It is significant for its contribution to the systematics of Proteaceae, and to the floristics of Australia, and also for its application of palynology to systematics."Warning: template has been deprecated.— Excerpted from On the natural order of plants called Proteaceae on Wikipedia, the free encyclopedia.
The Errata for this volume lists the following corrections to this page:
The Linnean system of botany, though confessedly artificial, has not only contributed more than all others to facilitate the knowledge of species, but, by constantly directing the attention to those essential parts of the flower on which it is founded, has made us acquainted with more of their important modifications that we probably should have known, had it not been generally adopted, and has thus laid a more solid foundation for the establishment of a natural arrangement, the superior importance of which no one has been more fully impressed with than Linnæus himself.
There are still, however, certain circumstances respecting the stamina and pistilla, which appear to me to have been much less attended to than they deserve, both by Linnæus and succeeding botanists. What I chiefly allude to is the state of these organs before the expansion of the flower. The utility of ascertaining the internal condition of the ovarium before fœcundation will hardly be called in question, now that the immortal works of Gærtner and Jussieu have demonstrated the necessity of minutely studying the fruits of plants in attempting to arrange them according to the sum of their affinities, as in many cases the true nature of the ripe fruit, especially with respect to the placentation of the seeds, can only be determined by this means. Its importance is indeed expressly inculcated by many botanists,
who, however, have frequently neglected it in practice: nor do I find any one who has steadily kept it in view, except Aubert Du Petit-Thouars in his excellent work on the plants of Madagascar and the Isles of France and Bourbon.
The bursting of the antheræ has, it is true, been generally observed, and many of its most unusual modes have been introduced into the characters of genera; but the examination of these organs, at a still earlier period, has been universally neglected; and hence the very imperfect knowledge which, even now, is possessed of their real nature in two of the most remarkable families of plants, the Orchideæ and Asclepiadeæ.
Examples of the great advantage of observing the antheræ in this early stage will hereafter be given in my general remarks on the order which is the proper subject of this essay. But I trust I shall be pardoned for here introducing some account of their structure in Asclepiadeæ, as it will enable me not only to bring forward the most striking proof of the importance of this consideration with which I am acquainted, but also, as I apprehend, to decide a question which has long occupied, and continues to divide, the most celebrated botanists.
The point in dispute is whether this order, comprehending Asclepias, Cynanchum, Pergularia, Stapelia, and several genera, at present confounded with these, ought to be referred to Pentandria or Gynandria, and, if to the latter, whether the antheræ are to be considered as five or ten; all of which opinions have had advocates of the greatest name in the science.
Linnæus has assigned no reason for his opinion, which, however, it appears he retained after he became acquainted with the observations of Jacquin and Rottboell; but it is probable he
was induced to adopt it more from the consideration of the close analogy these plants have with the manifestly pentandrous Apocineæ, than from regarding them as strictly referable to this class; for, in his natural generic characters of Asclepias and Pergularia, he very clearly describes both these genera as gynandrous.
Jussieu has entered more fully into the subject, but seems also to have been chiefly guided by this analogy and the observations of others; as he concludes by expressing his doubts, respecting both the origin and use of the parts.
Richard, whose description of these organs I find in Persoon's Synopsis, has indeed come nearer to the solution to the question; his account, however, of the origin of the lateral processes hereafter mentioned, proves that this description was not altogether formed on actual observation.
Jacquin, the first botanist that submitted these plants to minute examination, and whose figures well enough illustrate most points of their structure, has adopted a very different opinion, referring them to Gynandria, in which he is followed by Koelreuter, Rottboell and Cavanilles, all of whom likwise agree with him in considering them as decandrous; while Dr. Smith, in his late valuable Introduction to Botany, who conceives that "no plants can be more truly gynandrous," regards them as having only five antheræ. And lastly Desfontaines supposes the five glands of the stigma to be the true antheræ, considering the attached masses of pollen as mere appendages to these.
All the authors who thus refer them to Gynandria seem quite confident in the justness of their views; and yet the inspection of a single flower bud overturns, as it appears to me, with irresistible evidence, the conclusion they had formed from premises apparently so satisfactory.
My attention, while in New Holland, having been much
engaged by the plants of this family, the species in that continent being both numerous and with difficulty reducible to established genera: I there observed the following facts concerning them, all of which I have, since my return to England, confirmed by the examination of different species of the same tribe.
The observations of Jacquin on this subject being generally known, it must be unnecessary to enter into a minute description of those organs which are well exhibited by his figures in every respect, except as to the origin of the supposed antheræ.
If a flower bud of any plant of this family, while scarcely half the size it attains immediately before expansion, be carefully examined, it will be found that the polleniferous sacs, as they are termed by Jacquin and his followers, in which they suppose the antheræ to be merely immersed, are really the organs by which the fœcundating matter is secreted: for at this period they are perfectly closed, and consequently all communication cut off between the stigma and their contents now consisting of a turbid fluid or pulpy mass. If the stigma be at the same time observed, the gland-like bodies which originate in its grooved angles are already visible; but, instead of having the cartilaginous or horny texture which they are length acquire, are as yet semi-fluid, and of hardly a determinate form. Near the base of each side of these grooves a more superficial depression is observable, which, though in some cases extremely short, is in others of considerable length, and generally forms a right angle with the corresponding groove. In these depressions, the processes by which, at a more advanced stage, the contents of the antheræ are connected with the stigma, are immersed, and at this period they are found to be semi-fluid. By degrees the glands, as well as their lateral processes, acquire a firmer consistence, and the inferior or outer extremity of each of the processes, being extended beyond its
depression of furrow, on the bursting of the opposite cell of the corresponding anthera, firmly attaches itself to its contents, now become a regular mass of a waxy consistence.
If the accuracy of this statement be admitted, it will probably be allowed that the Asclepiadeæ cannot be regarded as gynandrous, especially in the sense in which they are so considered by botanists; but lest it should not be thought completely satisfactory, it may be added, that in a still earlier stage of the flower bud I have found the fœcundating matter already secreted in the cells of the antheræ, while the glands of the stigma, as well as their processes, were absolutely invisible.
As to the question of their being pentandrous or decandrous, every analogy must lead us to refer them to the former class; nor indeed have they, when not considered as gynandrous, been ever supposed to belong to Decandria.
An œconomy, in many respects similar to that now described, obtains also in Orchideæ, in which, however, the processes connecting the antheræ with the stigma, where they exist, are in many cases derived from the masses of pollen themselves; but in others they as certainly originate from the stigma, or its glandular appendage.
The result of my examination of these two interesting orders of plants, I hope hereafter to submit to the Society; and I now proceed to the proper subject of the present paper.
The natural order of Proteæ, or, as it is less exceptionally called, Proteaceæ, was first established in the Genera Plantarum of the celebrated Jussieu; and the description there prefixed to it will, with a few alterations, still apply to the order, now that it has received so many additions, not only in species, but in very distinct genera, several of which were first published by
The general description and definition of the order will be most advantageously placed at the head of its systematic arrangement; before entering upon which, I shall offer some remarks on its geographical distribution, and likewise on such modifications of structure in the different organs as appear to be of the greatest importance in indicating or characterizing genera.
The geography of plants being as yet in its infancy, the smallest addition to our knowledge of a subject which promises to become of considerable importance, will probably be received with indulgence; and in this persuasion I venture to make the following observations on the order before us. In the first place, it is remarkable that the Proteaceæ are almost entirely confined to the southern hemisphere. This observation originated with Mr. Dryander, and the few exceptions hitherto known to it, occur considerably within the tropic. The fact is the more deserving of notice, as their diffusion is very extensive in the southern hemisphere, not merely in latitude and longitude, but also in elevation; for they are not only found to exist in all the great southern continents, but seem to be generally, though very unequally, spread over their different regions: they have been observed also in the larger islands of New Zealand and New Caledonia; but hitherto neither in any of the lesser ones, nor in Madagascar. As in America, they have been found in Terra del Fuego, in Chili, Peru, and even Guiana, it is reasonable to conclude that the intermediate regions are not entirely destitute of them. But with respect to this continent, it may be observed, that the number of species seems to be comparatively small, their organization but little varied; and further, that they have a
much greater affinity with those of New Holland than of Africa.
Of the botany of South Africa, scarce any thing is known, except that of the Cape of Good Hope, where this family occurs in the greatest abundance and variety; but even from the single fact of a genuine species of Protea having been found in Abyssinia by Bruce, it may be presumed, that in some degree they are also spread over this continent.
With the shores, at least, of New Holland, under which I include Van Diemen's Island, we are now somewhat better acquainted, and in every known part of these, Proteaceæ have been met with.
But it appears that, both in Africa and New Holland, the great mass of the order exists about the latitude of the Cape of Good Hope; in which parallel it forms a striking feature in the vegetation of both continents.
What I am about to advance repecting the probable distribution of this family in New Holland, must be very cautiously received; as it is in fact chiefly deduced from the remarks I have myself made in captain Flinders's Voyage, and subsequently during my short stay in the settlements of New South Wales and Van Diemen's Island, aided by what was long ago ascertained by Sir Joseph Banks, and by a very transitory inspection of an herbarium collected on the west coast, chiefly in the neighbourhood of Shark's Bay, by the botanists attached to the expedition of captain Baudin.
From knowledge so acquired I am inclined to hazard the following observations.
The mass of the order, though extending through the whole of the parallel already mentioned, is by no means equal in every part of it; but on the south-west coast from a more decided
feature in the vegetation of the country, and contains a far greater number of species than on the east:—and in that part of the south coast, which was first examined by captain Flinders, it seems to be more scanty than at either of the extremes.
On the west coast also, the species upon the whole are more similar to those of Africa than on the east, where they bear a somewhat greater resembland to the American portion of the order.
From the parallel of the mass, the order diminishes in both directions; but the diminution towards the north is probably more rapid on the east than on the west coast.
Within the tropic, on the east coast, no genera have hitherto been observed, which are not also found beyond it; unless that section of Grevillea, which I have called Cyclopteræ, be considered as a genus. Whereas at the southern limit of the order several general make their appearance, which do not occur in its chief parallel.
The most numerous genera are also the most widely diffused. Thus Grevillea, Hakea, Banksia, and Persoonia, extensive in species in the order in which they are here mentioned, are spread nearly in the same proportion; and they are likewise the only genera that have as yet been observed within the tropic.
Of such of the remaining genera, as consist of several species, some, as Isopogon, Petrophila, Conospermum, and Lambertia, are found in every part of the principal parallel, but hardly exist beyond it. Others, as Josephia and Synaphea, equally limited to this parallel, have been observed only towards its western extremity; while Embothrium (comprehending for the present under this name all the many-seeded plants of the order), which is chiefly found on the east coast, and makes very little progress towards the west, advances to the utmost limit of south latitude, and there ascends to the summits of the highest mountains.
Genera consisting of one or very few species, and which exhibit generally the most remarkable deviations from the usual structure of the order, are the most local, and are found either in the principal parallel, or in the highest latitude.
The range of species in the whole of the order seems to be very limited; and the few cases which may be considered as exceptions to this, occur in the most extensive genera, and in such of their species as are most strictly natives of the shores. Thus Banksia integrifolia, which grows more within the influence of the sea than any plant of the order, is probably also the most widely extended, at least in one direction, being found within the tropic, and in as high a latitude as 40°. It is remarkable, however, that with so considerable a range in latitude, its extension in longitude is comparatively small: and it is still more worthy of notice, that no species of this family has been found common to the eastern and western shores of New Holland.
The celebrated traveller Humboldt is the first who has expressly pointed out a remarkable difference in the distribution of the species of plants.
He observes that, while the greater number grow irregularly scattered and mixed with each other, there are some which form considerable masses, or even extensive tracts, to the nearly absolute exclusion of other species. Of plants growing this in society, the greater number occur in the temperate zones; and of these, the most decided instances will readily present themselves to every botanist. I venture to add, that such as exist within the tropic, are found, either at considerable heights or on the sea-shores.
To this class very few of the Proteaceæ can be said to belong. Protea argentea of Linnæus is the most striking example among
the African species; and my friend Mr. Ferdinand Bauer has observed a similar tendency in Protea mellifera.
Among the New Holland species, Banksia speciosa is the sole instance, and even that only in certain circumstances, of this manner of growth.
The favourite station of Proteaceæ is in dry stony exposed places, especially near the shores, where they occur also, though more rarely, in loose sand. Scarcely any of them require shelter, and none a good soil. A few are found in wet bogs, or even in shallow pools of fresh water; and one, the Embothrium ferrugineum of Cavanilles, grows, according to him, in salt marshes.
Respecting the height to which plants of this order ascend, a few facts are already known. The authors of the Flora Peruviana mention, in general terms, several species as being alpine; and Humboldt, in his valuable Chart of Æquinoctial Botany, has given the mean height of Embothrium emarginatum about 9300 feet, assigning it a range of only 300 feet. On the summits of the mountains of Van Diemen's Island, in about 43° south latitude, at the computed height of about 4000 feet, I have found species of Embothrium, as well as other genera hitherto observed in no other situation. Embothrium, however, as it is the most southern genus of any extent, so it is also, as might have been presumed, the most alpine of the family.
Two genera only of this order are found in more than one continent: Rhopala, the most northern genus, which, though chiefly occurring in America, is to be met with also in Cochin-china and in the Malay Archipelago; and Embothrium, the most southern genus of any extent, is common to New Holland and America.
From this account of the geographical distribution of the Proceaceæ, I proceed to make some general remarks on the structure and modifications of their different parts. The order, which consists of shrubs of the most rigid nature, or of trees of moderate size, contains also one herbaceous plant, my Symphionema paludosum, which however, except in this respect and in the union of the tops of its filaments, does not remarkably differ from the usual structure of the family.
The pubescence, which is very general in the order, consists either of a short and in many cases nearly impalpable tomentum, or of soft hairs which are either spreading, close pressed, or somewhat crisped, generally simple, but in some genera fixed by the middle, and in a very few cases glandular.
The existence or absence of pubescence in the adult leaves cannot always be depended upon in distinguishing species; but the short tomentum, especially of their under surface, is of greater consequence than the spreading hairs. In the bracteæ more reliance may be placed on it, and in the different parts of the flower I have never hesitated to employ it in my specific characters. In the calyx I have even derived the greatest advantage in some difficult genera, especially Serruria, from attending to its differences in direction.
Mr. Salisbury has introduced the pubescence of fruit into several of his generic characters, and in some I think with evident advantage, but in such only as where from its abundance and length it performs a function of manifest importance in assisting dissemination: hence I conceive it may be safely admitted into the characters of Protea and Isopogon; but I can perceive no advantage whatever in employing it in those of Serruria and Spatalla. For this reason too it ought not to be used in the capsular or drupaceous genera, in which indeed experience
proves it to be of no further moment than in distinguishing species.
Dr. Smith has given it as his opinion, that from the disposition of leaves in New Holland plants no conclusion can safely be drawn as to their genera. This remark however appears to me only applicable to certain families, or rather genera; for in many tribes the plants of that country are altogether as constant in their leaves as in any other part of the world. In proof of this, it may be sufficient to mention the order Rubiaceae; and there are many others in which I find nothing at all remarkable in this respect.
As to Proteaceae, it must be acknowledged that in Banksia both verticillated and scattered leaves occur; but the leaves constantly in threes in Lambertia seems to me a circumstance of even greater importance than the number of flowers in the involucrum; and the opposite leaves of Xylomelum distinguish it at once both from Rhopala and Hakea.
Although the form and divisions of leaves in the order are variable in no common degree, yet there are certain genera, both among those of Africa and New Holland, which the leaves even in these respects assist in indicating. Thus, in that genus to which I have applied the name of Protea (the Erodendrum of Mr. Salisbury), and I believe also in my Leucadendron, there is no instance of a divided or toothed leaf; thus also the leaves of Spatalla are filiform and undivided, and those of Serruria filiform and almost always pinnatifid. Their dichotomous divisions in Simsia and Franklandia are still more characteristic; and their division and remarkable reticulation readily distinguish Synaphea from Conospermum.
The inflorescence in Proteaceæ, whatever use botanists may think proper to make of it in their generic characters, is of
undoubted importance in determining genera, and even in the primary division of the order it appears to be of nearly equal consequence with the fruit itself; for, in dividing the order into two sections from the structure of the ovarium, it will be found that while all the single-seeded genera have each flower subtended by a proper bractea, or more rarely are without one, those with two or more seeds have, with very few exceptions, the flowers of their spikes of clusters disposed in pairs, each pair being furnished with only one bractea common to both flowers: it may also be observed that all the American and two thirds of the New Holland species have this mode of inflorescence, while only one instance of it occurs in Africa.
The single envelope of the stamina and pistillum in Proteaceæ I have, with Jussieu, denominated calyx, chiefly because the stamina, of equal number with its laciniæ, are constantly opposite to them, and from the close analogy subsisting between this family and that of Thymeleæ, in which I believe the greater number of botanists will allow that this envelope is really calyx: and as this latter argument may be considered as the stronger, I shall endeavour to establish the identity of this organ in these two families. In several of the Thymeleæ, especially in Pimelea, the lower part of the tube of the calyx is, as it were, jointed with the upper; after the falling off of which, it remains surrounding the fruit: this is also the case in several genera of Proteaceæ, as in Adenanthos of Labillardiere, in Isopogon, in Grevillea Chrysodendron, and still more remarkably in Franklandia, in which the persistent tube becomes indurated and even nearly woody, a change surely not likely to take place in a genuine corolla. But though I have thus adopted the language of Jussieu, I am decidedly of opinion that, in all families having a single
envelope, it will be still better to call it perianthium or perigonium, which latter term was proposed by Ehrhart, and is adopted by Decandolle.
A circumstance meriting the attention of the theoretical botanist, respecting the calyx in this order, is its invariable division into four leaves or segment; for the single exception noted by Linnæus in his description of the male flowers of Brabejum, he himself seems afterwards to have distrusted, from the manner in which he has introduced it into the amended generic character given in the Mantissa; and I may add, that in nearly 400 species of the order, which I have examined, I have not met with a single exception to this rule.
With this uncommon constancy in point of number, it is remarkable that there is, in the whole order, a strong tendency to irregularity in form, the various kinds of which are of great importance in characterizing genera.
Before the expansion of the calyx the margins of its segments are applied to each other; and from the unequal degrees of cohesion in many cases subsisting among them after expansion, several kinds of irregularity arise. I am not sure that any term has been contrived for this manner of æstivation, except it be the æstivatio valvata of Linnæus; but as he has not defined it, and as his commentator Reuss has given the very different æstivation of grasses as an example, I have, in introducing this circumstance into the general description of the order, specified it at length.
From the colour of the calyx, many genera of Proteaceæ are indicated with tolerable certainty. Thus Synaphea is distinguished from Conospermum by its yellow flowers; and no instance of yellow flowers has been met with in the numerous genera Serruria and Spatalla, nor any of purple in Leucadendron. In some
genera however, as in Banksia and Isopogon, it is evidently of very little importance.
The fleshy or scale-like bodies, which surround the ovarium in the greater number of plants of this family, are in many case so manifestly secreting organs, that it is surprising Mr. Salisbury should hesitate in considering them as nectaria, and denominate them calli; a term which excludes the idea of secretion. But whatever their functions may be, great assistance may certainly be derived from their various modifications, in distinguishing genera. Their importance however in this respect, like that of all other parts, not only in this, but, as I apprehend, in every natural family, is very unequal, and in some cases seems to be entirely lost. Thus, in the genus Leucadendron as it is here constituted, they are wanting in several species, and in some I am inclined to think exist only in the males.
In most of the regular-flowered genera they are four in number, and alternate with the leaves or laciniæ of the calyx. In these genera they are also generally in the form of succulent scales, distinct, or more rarely cohering at their base, and in a very few instances adhering to the calyx; but in Persoonia they are nearly round and fleshy, and in Bellendena, Symphionema, Simsia, Agastachya, Petrophila, and Isopogon, they are entirely wanting.
In the irregular-flowered genera with two or many seeds their number is less than four, in most cases only one exists, in a few others three, and in some none.
Varieties in the structure or apparent origin of the stamina, afford, as might be expected, important generic characters. Their usual insertion in the order is in the concave tops of the laciniæ of the calyx; all considerable deviations from which may safely be employed in characterizing genera. In this way
Rhopala, Xylomelum, and Lambertia are readily distinguished from Embothrium, Grevillea, and Hakea; and thus also Persoonia and Brabejum remarkably differ from Gevuina; while Bellendena differs from all others in having its stamina distinct from the calyx, affording however an indication of the real origin of these organs in the whole family.
The deviations from the usual structure of antheræ in this order are not many; but some of them are so singular a nature as to constitute the essential characters of the genera in which they take place. These genera are Simsia, Conospermum, and Synaphea, all of which are most truly syngenesious; for not only do their antheræ firmly cohere together, but the corresponding lobes of these being, when considered separately, entirely open, are so applied to each other as to form but one cell, without a trace of any intermediate membrane. In Simsia the four antheræ are perfect, each consisting, as in the rest of the order, of two lobes, and therefore the whole before bursting constitute four cells. Whereas in Conospermum and Synaphea one filament is entirely barren, the two lateral ones have each a single-lobed anthera, and the fourth alone is perfect: hence before bursting the whole form only two cells.
This remarkable structure, which can only be ascertained before the opening of the calyx, necessarily escaped Dr. Smith in describing his Conospermum, for I conclude he had only the expanded flower before him, and the appearance of the antheræ in this state after their separation justifies him in referring the genus to Tetrandria: but according to the view now given of its structure, it can have no other pretension to a place in this class than its belonging to Proteaceæ; and the order Syngenesia Monogamia being abolished, it must be referred to Triandria.
The only remaining anomaly in these parts occurs in
Franklandia, and consists in the anthera, or rather that portion of the filament on which it is fixed, adhering to the calyx through its whole length.
The figure of the pollen has been attended to by a few theoretical, but by hardly any practical botanists; yet I am inclined to think, not only from its consideration in this family, but in many others, that it may be consulted with advantage in fixing our notions of the limits of genera: and though its minuteness may perhaps always exclude it from a place in generic characters, yet it well deserves, to use the words of Linnæus when speaking of habit, to be "occulte consulendum."
Its usual figure in the order is triangular with secreting angles, a beautiful contrivance for insuring impregnation in a tribe, in which, from the very scanty, or in many cases apparent want of secretion by the stigma, it must otherwise have been very uncertain; for by this form and secretion, as well as by the singular œconomy of the calyx, it remains so long in contact with the stigma, as probably to compensate for the somewhat defective structure of that organ.
From this figure the principal deviation is in the extensive genera Banksia and Josephia, in all of which it is elliptical or oblong, and either straight or bent into a semilunar form; and in Franklandia and Aulax, where it is spherical. The only remaining exception with which I am acquainted is the original Embothrium of Forster, his E. coccineum, in which, as in Banksia, it is oblong; a circumstance that, together with the more important character of a regular club-shaped stigma, and some other differences, has determined me to separate it from all the other species of Embothrium, except E. lanceolatum of Flora Peruviana, whose pollen however remains to be examined.
The external modification of the ovarium must be very
cautiously used in the generic characters of this family; even its being sessile or pedicellated is not always of sufficient importance, though I think Mr. Salisbury has done well in introducing it into his characters of Serruria and Spatalla, in both which genera I had overlooked it before the publication of his Essay.
Its internal structure, which ought always to be ascertained, will be found of the greatest importance in most cases, but fails in Persoonia, the species of which differ in having one or two seeds: it would seem however, in this case, that an irregularity in a point of such importance could not take place unaccompanied with other anomalies in the same organ, and accordingly such are found to exist in this genus, and will be mentioned when treating of the fruit.
Besides number, the insertion of the ovula is also to be attended to; for though this may generally be presumed from the situation of the radicula in the ripe seed, yet to this criterion there are several exceptions, even in the present order: thus, while the radicula constantly points downward in the whole of the order, the insertion of the ovulum is in many cases at the top or side of the cell of the ovarium. My observations on this subject are as yet incomplete; but, from those that I have made, I am inclined to think such differences will be connected with genera, or rather perhaps with particular kinds of fruit. Thus I conjecture, in Leucospermum, Mimetes, Nivenia, and Spatalla, the insertion to be uniformly lateral.
The style, though not subject to much variety in this family, will be found in a few cases to furnish generic characters. Thus in Protea, strictly so called, the persistent subulate style forms an important part of its character: and the persistency of the whole of the style in the greater number of species of Grevillea will probably be used by future botanists in distinguishing
them from that remarkable section of the genus, which I have at present united with them and called Cyclopteræ. Its length also, when compared with that of the calyx, seems in some cases to be of importance, as in distinguishing Adenanthos from Spatalla; but in general this circumstance can hardly be had recourse to except in specific characters.
The form of the stigma is in many cases of considerable importance in characterizing genera, a fact which could not escape the penetration of Dr. Smith when establishing his new genera of this order: thus its conical papilla in the Conchium (the Hakea of Schrader) will in many, though certainly not in all cases, distinguish it from Grevillea: but its form in both these genera will readily serve to separate them from Xylomelum and Rhopala; and thus also Spatalla remarkably differs from Adenanthos. Upon the whole, however, it seems that its obliquity is of greater importance than its form; for this, when existing in any great degree, is generally accompanied with a corresponding irregularity in the calyx: but as this irregularity is produced for the purpose of bringing all the antheræ into contact with the stigma, so its obliquity in the dioicous genera Leucadendron and Aulax is not attended with so great a degree of irregularity, which would here serve no end, impregnation depending on the pollen of different individuals, to insure which the surface of the stigma in these genera is rough with papulæ; a circumstance that, together with its form, readily distinguishes them from all others of the order.
In Synaphea, the stigma or summit of the style inosculates with the divisions of the barren filament, which in some species appear beyond it in horn-like processes, but in others are entirely lost in its substance. I am acquainted with nothing like this in the whole vegetable kingdom; and such a singularity
alone, when occurring in several species, would have determined me to separate these plants from Conospermum: but being also accompanied by other remarkable differences, both of structure and appearance, no genus, I apprehend, can be better founded than this.
That the opinion of Christian Knaut and Vaillant respecting the non-existence of naked seeds is correct when anatomically considered, there can be no doubt; but the practical utility of deviating in this subject from the common language of botanists may still be questioned: and accordingly Gærtner, who was fully aware of the truth of their position, has nevertheless continued to describe the seeds of many plants as naked. I confess however I am inclined to adopt the opposite decision of the French botanists, at the head of whom is Richard, who has also proposed terms for distinguishing the various species hitherto confounded under the name of naked seeds. The fruit of the monospermous genera of Proteaceæ might probably be with advantage referred to that which he has termed Ahena; but as I am unwilling in the present paper to adopt any term not more generally sanctioned and understood than this, I shall content myself with calling those nuces, which are either not at all or but slightly compressed and not bordered; and apply the term samara to such as are either very much compressed, or with a less remarkable compression are surrounded or terminated by a membranaceous border: that I regard these distinctions however as in some cases of very little importance, may be inferred from this, that my genus Leucadendron includes both these kinds of fruit.
The first observation I have to offer on the fruits of Proteaceæ is, that there is no really bivalvular capsule in the order; a truth which was not perceived by Gærtner in describing his
Banksia dactyloides (the Conchium dactyloides of Dr. Smith), and which has equally escaped Cavanilles and Labillardiere in their characters of Hakea. Dr. Smith has more cautiously omitted this consideration in his character of that genus, and Professor Schrader has accurately described the suture as only existing on one side: such fruits then are as truly folliculi as those of Grevillea, Rhopala, or Embothrium; and that the existence of a distinct placenta is by no means necessary to constitute this kind of fruit, is proved even by some genera of Apocineæ, to which family this term was first applied.
A circumstance occurs in some species of Persoonia to which I have met with nothing similar in any other plant: the ovarium in this genus, whether it contains one or two ovula, has never more than one cell; but in several of the two-seeded species a cellular substance is after fœcundation interposed between the ovula; and this gradually indurating acquires in the ripe fruit the same consistence as the putamen itself, from whose substance it cannot be distinguished; and thus a fruit originally of one cell becomes bilocular: the cells however are not parallel, as in all those cases where they exist in the unimpregnated ovarium, but diverge more or less upwards.
In all the seeds of this order there is a very manifest chalaza, which, whatever may be the point of insertion of the seed, is always situated at the upper extremity; and I have not been able to observe any fasciculus of vessels connecting it with the umbilicus in cases where this latter is placed in a different part of the seed.
I am not aware of any function being ascribed to the chalaza of seeds, except the nutrition of their proper membrane: but it appears to me too remarkable a part to be destined for this purpose only; and some observations I have made induce
me to suppose that it is the organ secreting the liquor amnios. This opinion I was first led to form by observing in some species of Persoonia, in which the inspissated remains of this fluid are visible in the ripe fruit, that it evidently originate in the chalaza and continued to adhere to it: nothing has hitherto occurred to invalidate this opinion, which is here however hazarded merely as a conjecture, requiring for its confirmation more numerous and decisive facts than I can at present adduce.
That the albumen of seeds is merely that condensed portion of the liquor amnios which remains unabsorbed by the embryo, seems to me very satisfactorily established; and as this fluid is in the early stage never wanting, all seeds may in one sense be said to have albumen: but while in some tribes this unabsorbed part in the ripe seed many times exceeds the size of the embryo, so there are others in which not a vestige of it remains; and such has hitherto been supposed to be the case with Proteaceæ: nor are the few exceptions with which I am at present acquainted of so decisive a nature as to invalidate this character of the order; for they occur only in some species of Persoonia, where the semi-fluid remains of this substance are observable between the cotyledons; and in Bellendena, in which it continues to form a thin fleshy coat on the inner surface of the proper membrane of the seed. From such instances however we may expect to find plants with a more copious albumen, which nevertheless it maybe necessary from the whole of their organization to refer to this family.
The radicula pointing towards the base of the fruit in all Proteaceæ is a circumstance of the greatest importance in distinguishing the order from the most nearly related tribes; and its constancy is more remarkable, as it is not accompanied by the usual position or even uniformity in the situation of the external umbilicus.
If Gærtner had not described the plumula of Protea argentea, I should not have hesitated to assert that it was inconspicuous in the whole order.
The number of cotyledons when more than two is a circumstance of little importance. In Persoonia, the only genus of the order in which a plurality of cotyledons has been observed, I am not even certain that their number is constant in those species in which this anomaly occurs.
In the following part of this essay it may be observed, that the genera into which I have subdivided the great African family Protea, are in most cases similar to those already proposed by Mr. Salisbury in the Paradisus Londinensis: from that essay however they are certainly not derived, but before its publication were formed and submitted to the judgment of Mr. Dryander, at whose suggestion they are now offered to the Society. That the results of an examination conducted by two observers wholly independent of each other, are so similar, will probably be considered as some proof of their correctness.
As Mr. Salisbury's generic names have the unquestionable right of priority of publication, I have in most cases adopted them, though I wish some of them had been differently constructed. But as I cannot accede to his application of the Linnæan names Protea and Leucadendron, I shall here, that I may not disturb the following arrangement, assign my reasons for differing from him in this respect; and as in so doing I am obliged to trace the progress of Linnæus's knowledge of the family, I persuade myself that this will in some degree compensate for the otherwise unwarrantable length of the discussion.
1735; no generic characters are there given, but from the references to Boerhaave's figures it is evident that the genus is to be understood in the same extensive sense which he at length gave it in the second Mantissa. In 1737 appeared the Genera Plantarum, and in it for the first time the natural generic character of Protea: as in this work he only cites Lepidocarpodendron and Hypophyllocarpodendron of Boerhaave, it follows that here the genus is more limited, though its character is not peculiarly applicable to either of Boerhaave's genera referred to; and the description of antheræ and germen is not reconcilable to any plant whatever of the family. In the same year Hortus Cliffortianus was published, in which he resumes his first opinion of Protea, reducing to it all Boerhaave's genera, but referring to the character given in his own Genera Plantarum. It does not appear on what ground this change of opinion was formed; for in Clifford's garden, according to Viridarium Cliffortianum, there had only been two species, Protea argentea and saligna, neither of which had flowered, and the former was already lost; while in his Herbarium, now in the collection of Sir Joseph Banks, the specimens of all the three species given in the body of the work are without fructification, and of Protea racemosa added in the appendix there is no specimen whatever.
If Linnæus is to be considered in a great degree the author of the Prodromus Floræ Leydensis, published by A. Van Royen in 1740, as has been asserted by some of his pupils, and may be inferred from a passage in his Diary published by Dr. Maton, it must be noticed as his next work in the order of time; for from the same Diary it appears that he could only have been employed in its composition in 1738. In this work the genus Protea is given in the same extensive sense as in Hortus Cliffortianus, and no fewer than 21 species are characterized, of which
however only two were in the Leyden garden, the rest being described from specimens in Van Royen's Herbarium.
In 1738 he also published his Classes Plantarum, in which, notwithstanding he appears to have composed it while engaged in the arrangement of Van Royen's collection, another fluctuation of opinion occurs, Protea being limited as in the first edition of the Genera Plantarum, and to Leucadendros, which here for the first time occurs, he refers the Conocarpodendron of Boerhaave.
In 1740 he published the second edition of Systema Naturæ, where the names Protea and Leucadendron are both given; but the references to Boerhaave are reversed, Protea being confined to his Conocarpodendron, and Leucadendron comprehending his other two genera. In this sense they also appear in the second edition of the Genera Plantarum published in 1742, in which the character of Leucadendron is first given, some of whose species he must, from the annexed asterisk, have seen recent: his description of corolla and pistillum is only applicable to Lepidocarpodendron.
In 1745 Linnæus received the Herbarium of Herman, from which he composed his Flora Zeylanica: the fourth volume of this collection containing a mixture of Ceylon and African plants, the latter are not noticed in this work; but from an inspection of the Herbarium itself, now in the Banksian collection, it appears that he had added generic names to most of them: of Proteæ only three species exist in the volume, of which Protea conocarpa is one: of this there are on the same page to specimens, whose heads of flowers are separately pasted; under one of these specimens he has written Leucadendron, and under the second Protea; to a specimen of Protea Serraria on a different
page he has given the name of Santolina. These facts are mentioned to prove, that at this period his knowledge of the family must have been chiefly derived from Boerhaave's figures, and perhaps from specimens which he had casually seen.
In 1748 the sixth edition of Systema Naturæ appeared, where the essential characters of Protea and Leucadendron first occur, both of them evidently derived from the natural characters previously given.
In 1753 the Species Plantarum, the most accurate of all his works, was given to the world; both genera are found in it, their species characterized, and trivial names for the first time applied to them: of Protea there are only two species, P. argentea and fusca; to the former however he referred as varieties P. saligna, conifera, and three others; to the whole adding the following observation, which may be supposed to contain his chief reason for applying his name Protea to this genus rather than to that for which in his Classes Plantarum he had first intended it. "Planta naturalis in patria argentea excellit fronde inter arbores nitidissima omnium; at culta et captiva extra patriam exuit decus; variat dein etiam domi mille modis verè Protea."
At this time he had in his Herbarium a specimen without fructification of Protea argentea properly so called; but of its supposed varieties or of P. fusca none whatever. Of his genus Leucadendron he had only one species, L. proteoides, afterwards called Protea purpurea, a plant differing in many respects from the tribe to which he had, though not without hesitation, referred it.
In 1754 the fifth edition of Genera Plantarum appeared, in which the characters of both genera remain exactly as in the second.
In 1759 was publishd the tenth edition of Systema Naturæ,
where the essential generic characters are nearly the same as in the sixth, and the specific characters are copied from the Species Plantarum.
Of this latter work the second edition appeared in 1762: it contains two additional species of Leucadendron described from Burmannus's Collection and Plantæ Africanæ: Protea argentea of the first edition is here divided into two species; the first Protea argentea now so called, the second comprehending P. saligna, conifera, and three other nearly related species: to this latter the greater part of the observation added to P. argentea of the first edition is annexed, though evidently less applicable to the species thus divided.
In the sixth edition of Genera Plantarum printed in 1764 no alterations are made in the characters of these two genera.
In Mantissa prima published in 1767, two new species of Leucadendron are described: neither of these, however, he had in his Herbarium: the first, Leucadendron speciosum, he had probably accidentally seen, the antheræ if which are described as filaments, and their callous apices alone as true antheræ: the description of the second, L. pinifolium, is by Van Royen.
In the twelfth edition of Systema Naturæ published in the same year, the species of Leucadendron are arranged in a different, and, as the author intended, a more natural order; from which it may be concluded that at this time considerable additions had been made to his Herbarium: but L. glomeratum is unaccountably omitted. Protea here receives again P. Levisanus, the P. fusca of the first edition of the Species Plantarum, which in the second had been referred to Brunia.
In Mantissa altera published in 1771, the two genera are united under the name of Protea; new characters are given to
the species, and most of them are described from specimens then in his Herbarium; five species are added which had already been published by the accurate Bergius; and three, P. totta, strobilina and parviflora, are here first met with: in his description of the last, he seems to suspect it to be a male plant, which we now certainly know to be the case. P. glomerata is here again taken up; but Protea acaulis, cancellata and conocarpa are omitted; and Protea conifera of the second edition of the Species Plantarum is subdivided into three species, P. conifera, pallens and saligna.
In the thirteenth edition of the Systema Vegetabilium published in 1774, the essential character of the genus is adapted to its present state, and no alteration occurs among the species, except that P. speciosa is considered as a variety of P. Lepidocarpodendron.
From this statement it appears, that Linnæus in his earlier works had not sufficient materials for obtaining an accurate notion of this family; and hence that perpetual fluctuation of opinion concerning it, which has been now pointed out, and may in few words be recapitulated.
1st, He gave the genus Protea the same extent which he at length assigned to it in the Mantissa.
2dly, He limited it, leaving unnoticed that part to which at a latter period he exclusively applied the name.
3dly, He resumed his first opinion.
4thly, He subdivided it into two genera, giving them the same names which are adopted in the present paper.
5thly, He continued the subdivision but reversed the names, and for a reason, as it would seem, which is now known to be founded in error.
And lastly, Having acquired more perfect materials and perceiving the insufficiency of his characters, he united them together, thus ending exactly where he commenced.
But, as in this he has been universally followed for nearly forty years, Protea can no longer be considered as more strongly associated with any one species of the genus than another; and therefore this name so familiar to botanists, if the necessity of again subdividing the genus be allowed; ought certainly to be given to that part which is best known, and which contains the greatest number of published species, especially if the name be at least as applicable to this as to any other subdivision: now this part unquestionably is the Lepidocarpodendron of Boerhaave, the Protea of the first edition of the Genera Plantarum and Classes Plantarum, and of the present Essay.
The question respecting the application of the name Leucadendron is reducible to a smaller compass. Mr. Salisbury is aware that the Linnæan character of the genus is only applicable to Lepidocarpodentron of Boerhaave; and therefore, consistently with the reasons which determined him in his application of the name Protea, Leucadendron ought to have been retained for that which he has called Erodendrum in Paradisus Londinensis; and this it seems he would have done, had it not been differently used by Plukenet, whom he professes to follow in this respect. But as rejecting Linnæan names when accompanied by characters, for those of Plukenet who never published a single character, is somewhat unusual, it must be supposed to have arisen from the latter author's more appropriate use of this significant name, while it may also be presumed that Linnæus's application of it is wholly unsuitable; and it is at least to be expected that in his own application he is consistent with Plukenet, whom he means to follow.
To determine how far this is the case, I have examined the figures published by Plukenet under the name of Leucadendros, and also his Herbarium, which forms part of the Sloanean collection in the British Museum. Of his three species so named the first is Protea argentea, his "Leucadendros africana arbor tota argentea sericea foliis integric, Atlas Tree, D. Herman." of which the figure represents a branch without fructification, and a separate fruit possibly of the same plant, but rather, as I suspect, belonging to a different species of the same genus.
On the same plate is figured a single leaf, in all probability belonging to P. conocarpa, with the following name, "Leucadendro similis africana arbor argentea folio summo crenaturis florida, an Leucadendros africana foliis serratis D. Herman.?" The separate fruit accompanying this probably does not belong to it, but to some species of that division of Leucadendron which Mr. Salisbury has called Euryspermum.
The third species, his "Leucadendros africana, seu Scolymocephalus angustiori folio apicibus tridentatis," is a good figure of a flowering branch of Protea cucullata.
It could not certainly from his publications alone be understood why the name Leucadentros is applied to these three plants so little alike, while different names are given to species much more nearly related to some of them that they are to each other: of this however the solution is to be found in his Herbarium; on consulting which I find, that after the publication of Protea argentea, with whose flowers he was unacquainted, he had acquired flowering specimens of Protea hirta, and had supposed these two species to be the same, pasting between two leaves of argentea four loose head of hirta, and under the whole copying in his own hand the name Leucadendros, &c. at full length from his Phytographia. This satisfactorily explains where he referred
P. cucullata to Leucadendros, its flowers being very similar to those of Protea hirta. As to his application of this name to P. conocarpa, it could only proceed from his total ignorance of its fructification; for, as he has figured a nearly related species, P. hypophylla, under the very different name of Thymelea, &c., it is reasonable to conclude, that had he seen the flower of P. conocarpa he would have given it the same generic name. This P. conocarpa however, of which it may truly be said he know nothing, and concerning which at least no information is to be derived from his works, is the only species of the three which belongs to Mr. Salisbury's genus Leucadendron.
But the original Leucadendros of Herman, of Plukenet, and of Linnæus himself, is Protea argentea, the only plant of the family to which the name can properly be applied; to this therefore I have assigned it in the following arrangement.
Before proceeding to this arrangement, I am happy in having an opportunity of acknowledging that assistance which has so liberally been afforded me.
To the invaluable Herbarium and Library of Sir Joseph Banks I have on this, as on all other occasions, enjoyed the freest access; an advantage which has been greatly enhanced by the opportunity it has given me of consulting my friend Mr. Dryander, both as to the formation of genera and respecting synonyms, on which point his sound judgment and unrivalled erudition so well enable him to decide.
To Dr. Smith I am indebted for the permission of inspecting the Linnæan Collection, and for the most friendly and satisfactory answers to the queries on this subject which he allowed me to put to him.
Mr. Lambert, whose Herbarium in this tribe is only surpassed
by that of Sir Joseph Banks, has, with his accustomed liberality, submitted it without reserve to my examination.
, who for many years possessed the most extensive collection of living Proteas that has ever been formed, and who also received from his intelligent collector Mr. Niven a valuable Herbarium of native specimens, most obligingly permitted me to examine these, and even to dissect such as were new. For the like privilege I am indebted to the friendship of Mr. Aiton of Kew, who sent me his whole collection, peculiarly valuable as containing many of the original specimens of Mr. Masson: and lastly, I have to acknowledge the great assistance I have derived from the extensive collection presented to this Society by my friend Dr. Roxburgh, who during his short residence at the Cape appears to have paid particular attention to this tribe of plants, and who, besides the many new species discovered by him, has given a greater value to his Herbarium by numerous observations on the sexes, the size, and places of growth, which I have every where inserted on his authority.
Calyx tetraphyllus v. quadrifidus, æstivatione valvatâ.
Stamina quatuor, (altero nunc sterili,) laciniis calycis opposita.
Folia sparsa, nunc verticillata v. opposita, persistentia, exstipulata, indivisa v. variè dentata, seu incisa profundiùsve laciniata, rarissimè verè composita.
Inflorescentia subspicata, modò laxius, in racemum v. corymbum floribus sæpè geminatis, nunc densiùs congesta in capitulum, vel aggregata supra receptaculum planiusculum, involucro persistenti, sæpiùs imbricato, subtensum: in quibusdam quasi abortione, uniflorum, indicante involucro calyculum tunc æmulante. Bracteæ dum flores geminati singulis paribus communes; in capitatis persistentes sæpiùsque auctæ et induratæ, rarò connatæ; in aggregatis nanæ, plerumque deciduæ, quandoque nullæ.
Flores in plerisque hermaphroditi perfecti, nunc organorum vitio diclines.
Calyx tetraphyllus, foliolis distinctis v. sæpiùs plùs minùs arctè cohærentibus tubulosus; limbo quadrifido, æquali, laciniis subspathulatis; nunc irregulari sive ex earum cohæsione rariùsve inæqualitate: coloratus, subcoriaceus, avenius, extùs sæpè pubescens, intùs glaber rariùsve barbâ utplurimùm partiali instructus, valvatìm aperiens, ante expansionem marginibus subtrunctatis mutuò cohærentibus: deciduus v. marcescens, dum tubulosus sæpiùs a basi tandem quadrifidâ abscedens, quandoque basi integrâ diutiùs persistente.
Stamina quatuor, (altero nunc sterili,) foliolis calycis opposita, iisdemque sæpissimè inserta, in plerisque juxta apicem, quandoque prope medium v. basin; rarò hypogyna; calycem nunquam superantia.
To render this essay as complete as I am able, I proceed to notice such plants, as either belong or have been referred to Proteaceæ, but from my imperfect acquaintance with which, or from the unsatisfactory accounts hitherto given of them, could not with certainty be referred to any of the genera described, or, if referable to any of them, I could not with confidence propose as distinct species; and shall conclude with the addition of a few synonyms to the species described from Ray's Historia Plantarum, which had escaped me when the paper was first read to the Society.
Protea linifolia. Jacq. Hort. Schœnb. 1. p. 11. t. 26.
Obs. There can be no doubt of the genus of this plant, or of the individual figured by Jacquin being a male. From the same figure, by which alone I am acquainted with it, it seems to be very nearly related to Leucadendron tortum, from which it differs in having the male heads sessile, and in the laminæ of the calyx being quite smooth.