I happened to spend some six weeks, from the end of July to the beginning of September 1886, at Arolla (a locality situated at the head of one of the southern offshoots of the valley of the Rhone near Sion), 6,400 feet above the level of the sea. During my wanderings about the woods and pastures which clothe the sides of the valley from this level to the snow-line, a couple of thousand feet higher, my attention was attracted by the characteristically alpine vegetation, and more especially by the Gentians, of which some species, such as Gentiana purpurea and G. campestris, were very abundant. G. verna and G. acaulis were also by no means uncommon; but the latter had almost ceased to flower.
It is well-nigh forty years since I occupied myself with systematic botany and I had no works of reference at hand except Gremli's 'Flore analytique,' which I happened to have bought at Lausanne, and Rapin's 'Guide des Botanistes dans le Canton de Vaud,' which a fellow-traveller was kind enough to lend me. But the extraordinary amount of variation which presented itself when I compared considerable suites of specimens with the diagnoses and descriptions in these works struck me so much that, all '"unanointed and unannealed" as I was in systematic work, I was tempted to see what 1 could make out of the facts for myself. In truth, the Gentians took hold of me rather than I of them and I have been more or less their bondservant ever  since. Beginning with Gentiana purpurea, I found that I could not understand that form without knowing something about the rest of the species of Gentiana; and, by a parity of reasoning, a knowledge of Gentiana involved that of the other genera of the Gentianeæ. So that since my return to England, I have been led to make a rapid survey of the whole Order; and it is the broad results of that survey that I wish to lay before the Linnean Society.
I have to thank the Director of Kew Gardens for the free use of the splendid herbarium under his charge. It affords the means of carrying out the investigation I have attempted much more thoroughly; and I am well aware of the incompleteness of the observations I venture to offer. But, at present, it is needful for me to turn my attention to other things; and if I venture to bring forward so imperfect a piece of work, it is in the hope that it may be taken up and finished by more competent hands.
Every botanist is aware that the Gentianeæ constitute one of the most natural and well-defined of the Orders of plants. The type of structure which runs through the five or six hundred species included in the Order undergoes but few and, for the most part, inconsiderable modifications. There are no trees and but few shrubs among them; a few are climbers; and there are a few aprophytes. The opposition, entirety, and palmati-venation of the leaves have but few exceptions; and it is very rarely that the flower departs from typical regularity. The chief distinctive characters of the groups into which the Order is at present divided lie in the flower, and indeed in the corolla; since the form and proportion of the calyx, the occasional synanthery, and the greater or less intrusion of the placentæ are characters which vary greatly from genus to genus.
Under these circumstances. I have confined myself almost entirely to the study of the structure of the flower. I find that, some seven or eight modifications of that structure are distinguishable; and that these again fall into two series, each of which is characterized by a peculiar disposition of the nectarial organs, and presents a gradation of forms of the corolla from the rotate, or rather stellate, condition through the campanulate to the extreme infundibulate character.
In one of these series the nectarial cells are situated on the inner surface of the cup, from the edge of which the lobes of the corolla proceed, and towards its basal end. They are aggregated in such a manner as to form either a single patch, bisected  by the vein which becomes the median vein of a corolla-lobe; or two patches, one on each side of that vein. I term the Gentianeæ of this series Perimelito.
In the other series there are no such patches of secreting-cells visible on the corolla; but in many members of the series there is a zone of such cells, which encircles the base of the ovary, and is therefore furnished by the outer faces of the carpellary phyllomes, which are often raised into tubercles. In others, I have not been able to make sure of the existence of nectarial cells, either on the surface of the ovary or on the stemono-carpellary internode, which forms the actual bottom of the flower-cup. But dried plants are very unfavourable subjects for the determination of points of this kind; and as many of these flowers, which have no apparent nectaries, are known to be visited by honey-sucking insects, I shall for the present assume that honey-secreting surfaces exist on the central parts of the flower. In contradistinction to the Perimelitæ, the Gentianeæ of this series may be termed Mesomelita.
In the series of the Perimelito four modifications of floral structure are discernible. To these 1 propose to give the names of Actinanthe, Keralanthe, Lephanthe, and Stephananthe.
The corolla is rotate, or, if it is more or less campanulate, the sinuses which separate the lobes are very deep. The nectarial arex are single or double, and often concave inwardly. There is no distinct gynophore; but the ovary is occasionally "stipitate"; that is to say, its basal moiety, which then contains no ovules and remains very narrow, constitutes an apparent stalk on which the dilated ovuligerous moiety is supported. The margins of the lobes of. the corolla may be produced into longer or shorter denticulations or laciniæ; but there are no filamentous appendages developed on the inner face of the corolla, The stigmatic surfaces are oblong-ovate.
 The species in which I have found this type of floral structure are:
In Exadenus the nectarial surfaces are hemispherically depressed and bulge outwards, so that their positions are marked externally by convexities of the corolla. Dr. Grisebach ('Genera et Species Gentianearum,' p. 322) says of each "fovea glandulifera," that it is "extus tantum conspicua, intus per petali substantiam clausa" but the real state of the case appears to me to be as I have stated it. Exadenzis, in fact, represents the first stage of a transition from Actinanthe to the next type.
This differs from the preceding in no essential respect, except in the deepening of the nectarial concavities, in such a manner that their external walls project as long horns or spurs which are sometimes directed downwards and sometimes upwards. Moreover, the lobes are relatively shorter, and the corolla is more or less campanulate. It is very possible, and indeed probable, that further inquiry may bring to light forms constituting a complete transition between Actinanthe and Keratanthe.
The genus Halenia is the only representative of this type.
 Thus, on purely morphological grounds and as a mere generalisation of the facts, without the introduction of any speculative considerations, the relations of the various types may be represented thus:
Or, to put the facts in another way, the several types in each series may be regarded as modifications of a common plan, of which the simplest exemplification is to be found in Actinanthe and Asteranthe respectively. If so, it is an easy step to the conclusion that both these are slightly diverse modifications of a still more simple, but, at present, purely hypothetical, common form, having the main features of Actnanthe and Asteranthe, but with the nectarial surfaces either feebly developed on both ovarian and corolline surfaces, or entirely absent. I will call this hypothetical "Ur-gentian," Haplanthe.
Thus far 1 have endeavoured to travel no hair's breadth beyond matters of observation and their obvious relations. It is plain that, even if all the five hundred and odd species of Gentianeæ had been created separately and raised in pots in the Garden of Eden, their morphological relations would have been exactly what they-are.
But, to a believer in evolution, the significance of these facts is unmistakable. With whatever caution one may be inclined to regard phylogenic speculations, it is hard, in such a case as this, to resist the force of the suggestion, that these morphological relations do really indicate the path which the evolution of the plants composing the Order has followed. At any rate, the evidence is strong enough to justify us in accepting this conception as a good working hypothesis. And there is the more justification  for doing so, that, if we regard the morphological facts by the light of Sprengel and Darwin's theory of the origin of flowers, they at once become intelligible.
I have little doubt that, with  larger knowledge, analogous causes will be found to be operative in all these cases. One of the great lessons which Darwin has taught us is faith in the doctrine of sufficient causes; and consequently hesitation in assuming that any structure, however slight. or unimportant in appearance, is devoid of significance in relation to either present or past conditions of existence, the chiefest of these being the struggle for existence with competitors, while climate and station probably occupy a very secondary place.
Even in respect of geographical distributionupon which, climate and station are usually assumed to exercise so great an influence,facts which have come under my notice in studying the Gentians have led me to be a little sceptical as to the extent of that influence.
At Arolla I never met with a specimen of Gentiana acaulis anywhere except in the region between the pine-woods and the snow-line. Yet this same species grows so freely in some parts of Southern England, as to be used for the borders of beds, in a kitchen-garden. The genus Erythroa, which is notorious for the slightness of the differences between its "species," is of world-wide distribution. It occurs all over Europe, in the Sinaitic Desert, in Egypt, in Hindostan, in the hottest parts of Australia, and in the moist temperate climate of New Zealand. Mr. Gunn, in a note appended to specimens of Gentiana montana in the Kew Herbarium, says that this species occurs everywhere, from the shore to the summit of the mountains. Gentiana campestris is said by Hooker and Arnott to be "abundant in Scotland, especially near the sea." It was no less abundant at Arolla from 6,400 feet to the snow-line.
In studying, with some care, the geographical distribution of various large and widespread groups of closely allied animals, such as the Canido, the Astacomorpha, and freshwater Fishes, I have been much impressed by the necessity of a most minute study of their morphology as a preliminary to any attempt to deal with the facts of distribution. I think there is no greater mistake than to suppose that distribution, or indeed any other large biological question, can be studied to good purpose by those who lack either the opportunity or the inclination to go through what they are pleased to term the drudgery of exhaustive anatomical, embryological, and physiological preparation.
Elaborate works on Distribution have been published which  are of little more value than catalogues of reference, because their authors have been unaware of this necessity. And 1 may point my remarks by showing that even such a brief and imperfect sketch of the minuter morphological characters of the Gentians as is here presented is fruitful of suggestions in regard to their Distribution.
My studies of Animal Distribution have led me to the belief that the division of the land-surface of the globe into large areas, which corresponds most nearly with the broader facts, may be stated as follows. Europe, Africa, Asia, and North America as far as Mexico form one great provinceARCTOGÆA; in which Ultra-Saharal Africa, Madagascar, Hindostan, and Indo-China are more or less distinctly characterized as subprovinces. For my present purpose it will suffice to speak collectively of the latter as the Southern Arctogæa, in contradistinction to the rest of the province as Northern Arctogæa. South America, with the Isthmus, as far as Mexico, constitutes a second great province, AUSTRO-COLUMBIA; Australia with the adjacent islands a third; while New Zealand and the neighbouring islets may be most conveniently regarded as a fourth.
In mentioning the species of the various types of Gentianeæ which 1 have examined, I have arranged them under the heads of A, for Northern ARCTOGÆA, B, AUSTRO-COLUMBIA; C, Southern ARCTOGÆA; D, AUSTRALASIA; E, NOVO-ZELANIA
Adopting this scheme of four great distributional provinces. one of which is subdivided into two regions, the following propositions appear to me to hold good of the Gentianeæ:
1. Species of the Order are found in all five regions. They flourish within the Arctic Circle and up to the limit of perpetual snow in mountain-ranges. They also abound in sundry tropical climates, both moist and dry. The Limnanthe type, represented exclusively by marsh- or water-plants, occurs in all the regions, and will riot be further mentioned.
2. The head-quarters of the Order (if we consider the number of types represented) are in the North Arctogæal and the Austro-Columbian regions, both of which contain representatives of all the types.
 The supposition that the distribution of the Gentianeæ in Pliocene or Miocene times was substantially similar to what it is now, is of course no solution to the problem of their distribution; it is simply driving the search for that solution farther back. Is it possible to fix an anterior limit to this retrogression?
 I suppose it would be, if one could fix the age of the first appearance of Diptera, Hymenoptera, and Lepidoptera provided with long haustra. For, upon Müller's hypothesis, the existence of the Keralanthe, Stepliananthe, Lissanthe, and Ptychanthe types presupposes that of such insects. Unfortunately, we are, as yet, hardly in a position to speak positively on this point. The most that can be said is that there is no evidence that they were abundant before the middle of the Mesozoic epoch, or that they existed in Palæozoic times. Free play, therefore, is left to speculation; and I do not think any good grounds. could be given for denying the existence of even the more specialized Gentianeæ in the Cretaceous epoch; while the "Ur-Gentian," the hypothetical anemophilous Haplanthe, may be dated back almost as much further as probabilities permit us to carry the existence of flowering plants. For it is obvious that a very slight further modification, in the direction of simplicity in Haplanthe, would bring about a form of flower which might serve as the starting-point for those of almost all the Orders of Dicotyledons. But speculation as to when or where the hypothetical Haplanthe may have originated is, for the present, idle. "Ignoramus" and, I fear, for a long while "ignorabimus."
Considering how slight the. morphological differences between the eight types really are, and that (according to the hypothesis) they have been brought about by the selective operation of agencies of the same order, it seems to me that it would be rash to deny that species belonging to the same type may have arisen in different localities. I do not think it probable that the process of modification and the materials it works upon would be so similar in widely different localities as to give rise to the close similarities which lead us to group individuals in the same species; but the polygeny of genera, and still more of larger groups, appears to me to be highly probable.
We are very much in the habit of tacitly assuming that because certain plants and certain animals exist only under certain climatal conditions, there is something in what we vaguely call the "constitution" of the plants or animals which binds them to these conditions, and renders it impossible for them to live elsewhere. I wish we could get rid of this word "constitution"; for I take it to be one of the many verbal anodynes by which the discomfort of ignorance is dulled. If it means anything definite,  it merely signifies that there is some morphological or physiological impediment to the existence of the plant or animal, outside the defined conditions; and our business is to find out what that impediment is. ...
Such are the observations I have to offer. I call them "Notes and Queries;" and I am afraid there are more queries than notes. My hope in offering them to the Society is to stimulate those who are better qualified than I am to carry through a serious botanical inquiry, and who have more time before them than 1 have to take up the subject. I believe that the systematic and exhaustive study of a single well-chosen Order, and of all the biological problems which it presents, would  inaugurate a new era in the progress of Botany. ...