Observations of a Naturalist in the Pacific Between 1896 and 1899, Volume 2 / Plant-Dispersal

by H. B. Guppy

The Mangrove Zone (the Coasts of Ecuador and Colombia)

We come now to the mangrove zone which comprises, with the remarkable exception of a long stretch of arid sea-border to the north of the Gulf of Guayaquil, the whole remaining western sea-border of South America, namely, the Ecuadorian and Colombian coasts. My own acquaintance with this region is limited to the estuary of the Guayas or the Guayaquil River and to the southern shore of the Gulf of Guayaquil; but I am able to avail myself of the researches of Baron von Eggers, which cover the entire Ecuadorian coast; and with Ecuador, therefore, I will bring this brief sketch of the littoral flora of one side of a large continent to a conclusion.

The Ecuadorian coast, lying, as Baron von Eggers observes, between the rainless and desert coasts of Peru and the “ewig grüne” coasts of Colombia, may be regarded as a transition-area presenting very varied and complicated conditions. With the cause of the remarkable contrasts exhibited by the strand-flora, not only on the coast of Ecuador, but along the whole west coast of South America through some forty-five degrees of latitude from Patagonia to Colombia, I will presently deal. Here it may be remarked in passing that the Humboldt Current has played the determining part in producing the abnormal climatic conditions to which these remarkable contrasts in the strand-flora of this coast of the continent are mainly due.

The mangrove zone, marking a more or less abrupt transition from a region of drought and semi-sterility to one of humidity and rank tropical vegetation, begins about lat. 3° 30ʹ S., that is, in the vicinity of Tumbez, or perhaps nearer the boundary-line between Ecuador and Peru in lat. 3° 20ʹ (see [Note 72]). Occupying the southern shore of the Gulf of Guayaquil it extends up the Guayas estuary to Guayaquil and rather beyond. But when we follow the coast of Ecuador northward from the island of Puna towards Santa Elena Point, we come upon one of the most remarkable phenomena presented on the west coast of South America. The dry region begins again and the mangroves disappear; and these conditions continue through about 2¹⁄₂ degrees of latitude until we reach the equator, when the mangrove zone soon recommences, and, as I infer, continues northward without a break to the coast of Central America.

Dealing first with the mangrove districts of the south side of the Gulf of Guayaquil and of the Guayas or Guayaquil estuary, we may observe that probably in few localities of the globe have the forces of nature worked more in unison to produce the conditions favouring the growth of the mangrove. The reason why this particular locality has been thus favoured will be discussed later on in this chapter. I may here observe that Baron von Eggers was so struck with the exceptional features of the mangrove-growth in this region that he was inclined to look for the American centre of the genus Rhizophora, the prevailing mangrove, in the estuary of the Guayas River.

I will not enter into a detailed description of the mangrove-formation of this coast, which has indeed been given by the German botanist; but I will merely refer to the leading features such as they presented themselves to me. In the first place, reference will be made to the sea-border of the province of Eloro, where I spent nine or ten days, making Puerto Bolivar, the port of Machala, my headquarters—a locality about thirty miles east of Tumbez. Except in the Guayas estuary I have never seen such a magnificent growth of mangrove.

By following the line of light railway that runs about six kilometres inland from Puerto Bolivar to Machala, the capital of the province, we obtain a good section of the mangrove-belt, which may be thus described. The mangrove-swamp proper extends about three kilometres inland. Whilst the small variety of Rhizophora mangle (mangle chico) immediately fronts the sea, Laguncularia grows on the islets close to the seaward margin of the swamp. When we enter one of the numerous broad creeks that intersect the border of the mangrove-belt we soon find ourselves in the true mangrove forest, where prevail tall trees of Rhizophora mangle (mangle grande) that rise to a height of 70 or 80 feet or more. Gloomy as the depths of the swamp are, they acquire quite a funereal aspect, the branches of the trees being draped with pendent Tillandsias. These long, hair-like, tangled growths hang vertically from the branches of the trees and may be 20 or 30 feet in length. In the rear of the zone of tall mangroves we come upon a more open district of the swamp. The forest proper gives place to a tract occupied by small trees of Rhizophora, Laguncularia, and Avicennia, with here and there whole acres occupied only by the shrubby Salicornia peruviana which attains the height of a man.

[To face page [484].

Emery Walker sc.
Rough Plan of the Gulf of Guayaquil.

(The main stream of the Humboldt current, as indicated by the arrows, turns off to the north-west at Cape Blanco; whilst a small branch crosses the mouth of the Gulf of Guayaquil and flows along the Ecuador coast north of Santa Elena Point.)

Here terminates the mangrove-swamp proper, and about three kilometres from the sea it passes gradually into a region of extensive bare mud-flats which are penetrated by salt-water creeks, two or three yards across and a foot or two in depth, that are bordered by low and shrubby Avicennias, the Salicornia bushes above noted, and dwarfed trees of Rhizophora mangle only four or five feet high. These flats, which are evidently only overflowed by the sea at the higher spring tides, were at the time of my visits much sun-cracked and in some parts incrusted with salt; but the mud was rather soft, and in places Sesuvium portulacastrum and Batis maritima flourished in quantity on it. These mud-flats, about two kilometres across, pass by degrees into the low-lying level district known as the Machala plains, on which the capital of the province is built. Here the soil is dryish, and, notwithstanding that it displays on its surface when exposed to the sun a white saline efflorescence, a dry jungle type of vegetation of the xerophilous character here thrives. I noticed casually the Algaroba (Prosopis), a yellow-flowered Cordia, cacti of the Opuntia and Cereus kinds, besides several small trees and shrubs often thorny.

These Machala plains, on account of the fine saline incrustation above mentioned, are of much interest, since at a distance of six kilometres from the coast they thus display on their surface the effect of sea-water infiltration, their level above the sea being only a few feet. We have seen the three stages of this infiltration landward of sea-water: first, the mangrove-swamps daily overflowed by the tide; second, the mud-flats behind them which are only overflowed by the fortnightly spring-tides; third, the vegetated plains behind all, which are sufficiently raised to be above the reach of the tides, but which are nevertheless soaked with sea-water that displays its presence in the salt left by evaporation on the surface of the soil.

But another interesting point is here raised. At the back of the mangrove-belt, in most parts of the world, we usually find a particularly rank and luxuriant vegetation where the Scitamineæ often take a leading part; whereas on the sea-border of this part of Ecuador the mangrove-swamps pass gradually into arid saliniferous plains. With this singular fact is to be associated the circumstance that we see here in operation, some four or five miles from the coast, a process by which great quantities of sea-salts are accumulating below the surface. This may possibly be concerned with the origin of the great saline deposits of Northern Chile. However this may be, there is some reason for believing—and I understand that this is the opinion of Dr. Wolff, the historio-geographer of Ecuador—that in the course of ages the tendency will be towards an extension of the dry, sterile regions of Northern Peru into Ecuador. This subject is referred to again in a later page of this chapter.

Whilst in this neighbourhood I made the ascent for some fifteen miles of the Santa Rosa River, which opens into the sea near Puerto Bolivar. It is a tidal estuary that has no proportion in size to the small river that enters it. In its lower third we passed at first between long mangrove-islands formed almost entirely, as viewed from the boat, of the tall Rhizophora trees draped with Tillandsias, and presenting really a magnificent spectacle. In the middle third we were penetrating into the rear of the mangrove-belt. The giant swamp-fern (Chrysodium aureum) abounded, and here and there we passed by a patch entirely held by the large shrubs of Salicornia peruviana. The tall Rhizophora trees were replaced by the short variety, the “mangle chico,” which ceased altogether about ten miles from the mouth of the estuary, but probably only about five miles from the nearest part of the coast. The water at the place where the Rhizophora trees ceased was evidently quite fresh during nine out of the twelve hours, being only salt in the latter part of the rising tide. Above the mangroves, in the upper third of the ascent, Hibiscus tiliaceus, with Chrysodium aureum, flourished on the banks. The shallows at the margins were occupied by a considerable variety of semi-aquatic and other plants, such as Pontederia (two species); one of the Alismaceæ, with the flower and fruit of Sagittaria and the leaves of Alisma; Typha, Polygonum, and an Amaryllid like Crinum. Plants of Pistia and Pontederia floated in the stream.

I have said enough to give a general idea of the composition of the mangrove-belt of the Ecuador littoral, and will refer but briefly to the mangroves and other river-side plants in the neighbourhood of the city of Guayaquil, some forty miles up the Guayas estuary. As I have remarked in [Note 38], the water of the river off the city is usually quite fresh except at high water; but the sea has much freer access to the channels at the back of Guayaquil, where at high water the density was 1·014. In these channels are displayed the typical mangrove formation, trees of Rhizophora mangle bordering the water, whilst behind they are mingled with Avicennia tomentosa and Laguncularia. On the banks of the main river, where they are overflowed at high water, Anona paludosa was the most frequent tree, being associated with the Rhizophora, Hibiscus tiliaceus, and other trees. Above the city, Polygonum glabrum was growing in dense masses at the river’s edge, whilst Pontederia and Pistia flourished on the low muddy banks and floated in quantities in the river.

Before quitting the subject of the mangrove-formation of Ecuador, I will refer shortly to the two varieties of Rhizophora mangle that here occur. Baron von Eggers received the impression that the common type of this species, a low tree bordering the coast, did not exist in Ecuador, such a type as he says is characteristic of the West Indies and of Central America, and, I may add, also of Fiji. The species he regards as acquiring a new facies in Ecuador, where it exists as tall forest-trees, branchless for half their height, and exhibiting other divergent characters. However, I found that the common type of the species occurs normally on the coast in the vicinity of Puerto Bolivar, thirty miles east of Tumbez, a district above described.

There are two distinct forms of Rhizophora mangle exhibited in the mangrove-belt of the coasts around Puerto Bolivar. One of them, which the indigenes name “mangle chico,” is a small tree, 10 to 15 feet high, with useless timber, that immediately borders the sea, and, in fact, largely forms the margin of the swamp, not only on its seaward side, but also on the land side, where it passes into drier ground. The other, the “mangle grande,” a tall tree reaching to 60 or 80 and sometimes perhaps to 100 feet in height, composes the interior, and indeed the bulk, of the mangrove-belt, and possesses a hard and durable timber much employed in the district.

Distinct as these two types are, it is not difficult to find intermediate forms, and, in truth, in some localities they prevail. But the interesting point is that this peculiar Ecuadorian type of the species, a type that attracted the attention of the eminent German botanist, comes near the “Selala,” the mysterious seedless Rhizophora of the Fijian swamps—a subject fully discussed in [Chapter XXX.], where I have compared the Fijian and Ecuadorian Rhizophoras. Both the “Selala” of Fiji and the “mangle grande” of Ecuador are intermediate between the American Rhizophora mangle and the Asiatic R. mucronata, resembling the last in their inflorescence, but in other points approaching the American species. The “Selala,” however, comes nearer to the Asiatic tree, whilst the “mangle grande” comes nearer to the American tree. Unlike the Fijian tree, that of Ecuador is not sterile, but matures its fruit; and it displays no evidence of the vegetative reproduction so characteristic of the “Selala.”

Sandy beaches are not common on the mangrove-fronted shores of the south side of the Gulf of Guayaquil. However, on the seaward side of the long low mangrove island of Jambeli, on which the lighthouse is placed off Puerto Bolivar, there is a long stretch of beach of whitish, mainly non-calcareous, sand. The Coco palms behind the beach give the coast quite the aspect of a Pacific island strand. Ipomœa pes capræ flourishes on the sand nearest to the sea; and immediately behind, the beach is more or less occupied by a Cyperus 2 to 3 feet high, and by Canavalia obtusifolia. Further back grows a small Acacia tree, and behind it the yellow-flowered Cordia tree of the district; and in the rear of all lie extensive mud-flats, partly occupied by stunted bushes of Avicennia tomentosa and by Sesuvium portulacastrum, which in their turn pass into the mangrove-swamps.

On account of the enormous amount of drift of all descriptions that is carried to the sea by the Guayas or Guayaquil River, floating vegetable materials are abundant in the Gulf of Guayaquil, and are thrown up in quantity on the coasts of Ecuador. One of the most interesting spectacles at Guayaquil is presented by the floating river-drift. Huge tree-trunks and floating islets, the last-named ranging from 3 or 4 to 30 or 40 feet or more across, were, at the time of my visit in February, being carried to and fro unceasingly in front of the city by the tide, gradually making their way down the river, and ultimately reaching the open waters of the gulf. Floating plants of Pistia were in abundance; and their fate when they reached the sea must have been tragical. The islets were exceedingly interesting; they were evidently formed of materials lifted up bodily from the shallows at the margin of the river, and then carried off in the stream. They were mainly composed of two species of Pontederia and of Polygonum glabrum in the position of growth; the first often in flower. Pistia and a variety of smaller plants nestled among them, such as Salvinia, portions of Azolla, Lemna, &c.; and in one islet I noticed, oddly enough, the growing rhizome of a sensitive plant (Mimosa pudica). A great quantity of floating seeds collect amongst the roots and stems of the plants composing the islets, and here I obtained much of the smaller seed-drift.

Most frequent in the floating drift of the river at Guayaquil were the seeds of Anona paludosa, often in a germinating condition. The seeds are liberated by the decay of the floating fruit, which was also common in the drift. Amongst the larger materials were the seeds of Entada scandens and of Mucuna; the empty seeds of the vegetable-ivory palm (Phytelephas macrocarpa), the sound seed possessing no floating power; the “stones” of Spondias lutea, L., as identified by Mr. Holland, of the Kew Museum; the empty small nuts of several palms, including, apparently, Oreodoxa, &c. Amongst miscellaneous materials were small gourds, which are referred to in [Note 47], and an occasional empty cacao fruit. Smaller seeds were also abundant, and included those of Hibiscus tiliaceus, Erythrina, Vigna, Ipomœa, and others. Carried into the river from the neighbouring mangrove-creeks, where they abound, there were floating seedlings of Rhizophora and Avicennia, fruits of Laguncularia often germinating, and the seeded joints of Salicornia peruviana.

There was of course, in addition, much that was strange in the floating drift of the Guayas River, which received its contributions not only from the river-side vegetation and the neighbouring mangrove-swamps, but also from the interior mountain ranges culminating in Chimborazo, the slopes of which are drained by its tributaries. I had several opportunities of meeting the drift of the Guayas River in the open waters of the Gulf of Guayaquil. Much of it is carried along the south side of the gulf; and I picked up at sea, ten to twenty miles from the mouth of the estuary, many of the things above enumerated, such as Erythrina and Mucuna seeds, seeds of Hibiscus tiliaceus, the empty vegetable-ivory seeds, the seedlings of Rhizophora and Avicennia, and the germinating fruits of Laguncularia and Salicornia peruviana. Much of these materials mingled with local drift is stranded on the long beach of Jambeli Island, thirty miles from the mouth of the estuary. Here, besides the seeds of Canavalia obtusifolia and Ipomœa pes capræ derived from the locality, I found the seedlings of Rhizophora and Avicennia, and the fruits of Laguncularia and Salicornia peruviana, that might have been in part derived from the adjacent swamps, as well as much of the drift of the Guayas River, such as the seeds of Anona paludosa, Entada scandens, Erythrina, and Mucuna, the small gourds, the same small palm-nuts, the empty seeds of Phytelephas, the “stones” of Spondias lutea, and much other material previously familiar to me, but nowhere a sign of the floating Pistias and of the flowering Pontederia islets of that estuary.

The Stretch of Dry Coast from the Vicinity of Puna Island to the Equator.— This remarkable piece of sea-border, covering nearly three degrees of latitude, and in its aridity and general character recalling the sterile sea-coast of Peru, is placed between the humid mangrove-fronted coast of the Guayas estuary and the similarly humid and mangrove-fronted coasts of Northern Ecuador and Colombia. The mangrove seems to be almost absent from this stretch of dry coast. Mr. F. P. Walker, of the Santa Elena Cable Station, tells me that some time ago a little mangrove-growth existed near the Point, but that it has disappeared; and Baron von Eggers implies the absence of mangroves from the whole coast. The first-named speaks of the dry character of the coast district from Santa Elena Point to within half a degree of the equator; and the last-named, in his description of the coast, mentions cacti and thorny plants as typical of the vegetation. Since this region represents a typical locality where the direct influence of the Humboldt current on the climate of almost the whole west coast of South America can be put to the proof, I will refer to its peculiar climatic conditions below in my discussion of the general question, and will here content myself with saying that on this dry portion of the coast of Ecuador we have reproduced, but in a less pronounced degree, the climatic conditions of the coast of Peru.

The Humboldt or Peruvian Current and the Climate of the West Coast of South America.—The question we will now briefly consider is one that is concerned with the determining causes of the singular distribution of coast-plants on the west coast of South America. The reader will have already seen that the matter is an affair of climate; but it is an affair of climate in which (although it affects forty or more degrees of latitude), latitude, in a general sense, scarcely counts. All the naturalists, from Humboldt onward, who have sojourned in this region of the globe have displayed a deep interest in this subject; and I suppose there can be no region of the globe where there are so many climatic anomalies as interesting to the meteorologist. Here, for instance, might be obtained materials for solving the irritating mystery of a London fog; and if the suggestion of Baron von Eggers, before alluded to, is carried out, and a station is established by the Meteorological Societies of Europe and America at some suitable locality like Santa Elena on the coast of Ecuador, we might obtain, among other results, another line of investigating the causes of the fogs of our metropolis, a subject about which Captain Carpenter has recently made an important preliminary inquiry.

I will assume that my readers are already acquainted with the nature of the problem to be discussed relating to the climate of the west coast of South America, and that they are familiar with the view generally held that the aridity of this extensive coast region, stretching from the thirtieth parallel of south latitude to the equator, arises from the loss by the trade-winds of all their moisture in the interior of the continent before reaching the western countries of Chile and Peru. Mr. Ball, in his book on South America, opposed this view, though from reasons only partially valid, since he instanced the Ecuador coast as being, contrary to the theory, a wet coast, whereas we know that a large stretch of it is arid and not unlike Peru. The parting of the ways in this discussion lies in the answer to the query, Why should the south-east trade carry so much moisture to the east side of South America, whilst the south-west winds, that are equally prevalent on the west coast of the continent, are drying winds which convert the sea-border into a desert, as in Northern Chile, or into a region of semi-sterility, as in the instance of Peru? Other things being equal, we should expect both sea-borders of the continent in these latitudes to be well watered. In the answer to the question why the south-east trade should be a wet wind and the south-west wind a dry one lies a fatal objection to the prevailing view.

When Professor Davis, in his article on North America in the Encyclopædia Britannica (vol. 25), observes in connection with the arid coast regions on the west side of the continent that the southerly flow of the winds along the Pacific coast gives them a drying quality, thus causing the extension to the coast in South California and in North Mexico of the arid regions of the interior, he seems to imply that these winds acquire their drying capacity in flowing from cooler to warmer latitudes. On this view all trade-winds should be drying winds, whereas the reverse would appear to be the case.

There is some condition, present on one coast of the South American continent and absent on the other, which determines why a southerly wind, blowing landward, is in the one case moist and in the other dry. According to my own view the winds of the arid coast regions of western North America cross the cool waters of the Californian current, and thus acquire their drying quality on striking a sea-border more highly heated than the winds themselves. On the tropical west coast of South America the winds also become drying winds by passing over the cold waters of the Peruvian or Humboldt current, where mists are in consequence of frequent occurrence; and on striking the more highly heated land-surface at the sea-border the moving air does not part with any more moisture until an altitude of some thousands of feet above the sea is reached, when the cloud-belt forms. On the mountains bordering the coast of the Antofagasta province, in January, the clouds gathered at an elevation of 4,000 to 5,000 feet. Perhaps the best way to contrast the east and west coasts of tropical South America in this respect would be to say that whilst the wind blows landward in both regions, the land is the condenser on the east side, and the sea, owing to the interposition of the cold Humboldt current, is the condenser on the west side.

During a fortnight spent in February at Ancon, about twenty miles north of Callao, I noticed that with the prevailing cool south-westerly wind the coast was clear, but it was misty at sea. On the few days when there were warm westerly and north-westerly breezes, the weather was thick at sea; and if this condition was pronounced, the whole coast was enveloped in mist; but more usually the coast-line was fairly clear except at the promontories, along the sides of which clouds blown in from the sea rolled in lines inland, not generally attaining an elevation over 300 or 400 feet, but sometimes reaching 900 or 1,000 feet, and gradually disappearing a mile or two inside the coast-line. These sea-born clouds thus vanished as they traversed the more highly heated land-surface; and the air-current continuing its inland course mounted the slopes of the adjacent mountain ranges of the Andes, some three or four miles from the coast, until at an altitude of some 5,000 or 6,000 feet condensation again occurred and the cloud-belt was formed at those cooler levels.

From the summit of a range rising to a height of about 2,500 feet to the north-west of Lima I had presented to me a splendid spectacle, on February 12th, in the formation of the coast-belt of clouds. The forenoon was clear, but about 2 p.m. the sea-born clouds began to roll inland, concealing the lower two-thirds of the island of San Lorenzo, which has an elevation of almost 1,400 feet, and completely covering up Callao and the low country bordering the sea, but extending only a mile or two from the coast-line. The dense cloud that covered Callao appeared, as I looked down upon it from my mountain-peak, like a billowy field of snow sparkling in the sun, with the summit of San Lorenzo standing out like some bare alpine summit from amidst the snows. Yet beneath that dazzling covering Callao lay all in gloom; whilst only six miles up the broad valley of the Rimac the city of Lima stood in a blaze of sunlight, its domes and towers reflecting back the light as I looked at the strange contrast it presented with the buried city of the coast. The mystery of a London fog seemed to lie unfolded at my feet, ready for the man who can read the signs aright.

That the mere presence of a cold current on a coast with the winds blowing off the land (as in the case of the Labrador current, which extends down the Atlantic coast of North America to Cape Hatteras and beyond) produces no sterilising effect on the vegetation of the sea-border of a continent is well brought out in the beautifully executed maps in Prof. Russell’s recent work on North America. The essential condition for producing sterility on the sea-border of a continent is not only that the waters of a cold current should wash its coasts, but that the regular winds should blow landward across its cool surface. These are what we find on the west coast of South America.

Not with the hope of adding anything new to our knowledge of the climatology of this region, but with the purpose of becoming personally acquainted with the problem involved, I paid considerable attention to this subject during the three months passed on the west coast of South America between Port Valdivia and Guayaquil. It was not until I had dropped my thermometer into the cool water of the Humboldt current and had watched the formation of the fogs on the sterile coast of Peru that the real nature of the problem presented itself. From the pages of a work like Tschudi’s Travels in Peru one acquires an excellent idea of the extraordinary climatic conditions of this region, and the same may be said of the narratives of Darwin and other travellers; but it is necessary to be brought into personal contact with these conditions before one can appreciate their significance.

As is well known, says Baron von Eggers, the Humboldt current explains the anomalous climate of the coast of Peru, and one may add of North Chile and Ecuador. The current, which represents the extension northwards of the west wind-drift of the Roaring Forties (see Dickson in Encycl. Brit., xxxi. 404; and Admiralty Current Charts of the Pacific), begins on the coast between the 33rd and 40th parallels of south latitude, according to the season. North of Valdivia, as we approach Valparaiso, in lat. 33°S., the effect of its presence is at once seen in the increasing dryness of the climate and in the alteration in the character of the vegetation. It has, however, been shown that the current needs the co-operation of the prevailing southerly and westerly winds as they blow landward over its cool waters. On the coast of Peru these moist winds often generate fog and mist as they cross the current. They reach the coast as drying winds, having a temperature much cooler than the lower coast regions; and the air-currents do not precipitate any moisture on the land until an elevation of 4,000 to 6,000 feet is attained where the cloud-belt is formed.

In order to establish this theory it is, however, necessary to show that when the Humboldt current leaves the coast normal conditions of humidity occur, to which the vegetation responds, and that when the current strikes the coast again the conditions of aridity reappear. In its course northward the current divides off Cape Blanco, the principal mass of its waters making towards the Galapagos Group, whilst the remainder, after crossing the Gulf of Guayaquil, flow along the coast of Ecuador between Santa Elena and the equator. Now, it is along this stretch of the Ecuador coast that the conditions of aridity reappear and that the climate of the Peruvian sea-border is in a modified form reproduced. In the interior of the Gulf of Guayaquil, on the other hand, where the sea-border is no longer subjected to the influence of the cold waters of the Humboldt current, the genius of the tropics, repressed through so many degrees of latitude, bursts its bonds, and presents us with a spectacle of littoral vegetation that, so far as mangrove-growth is concerned, is probably unrivalled on our globe.

This contrast is well shown in the mean annual temperatures on the opposite sides of the Gulf of Guayaquil. Baron von Eggers, quoting Dr. Wolff, states that whilst the mean for the year at Puna is about 75° F., and at Santa Elena about 73°, on the south side of the gulf at Balao it is several degrees warmer and is evidently not under 80°. The mean temperature for the second week of March during my sojourn at Puerto Bolivar, which is near the beginning of the mangrove region on the south side of the gulf, was 79°, the mean daily range being 74° to 83·5°. This stretch of dry coast reaching north from Puna to the equator is evidently regarded by Baron von Eggers and others who have studied the climatology of Ecuador as the critical area required to confirm the theory connecting the aridity of the west coast of South America with the Humboldt current. Here the sea for the greater part of the year has a temperature (according to the British Admiralty chart of surface-temperatures) of 70° to 75°; the mean temperature of the air is 73° to 75°; the rainy season, instead of covering a period of six months and over, as in the humid regions north and south of this coast, has a duration of only two or three months; the prevailing wind is south-west; whilst the direct influence of the cool waters of the current is shown in the general cloudiness that prevails during the last half of the year and in the drizzling mists that are frequent from June to October. Reference has already been made to the manner in which the vegetation on this dry coast of Ecuador responds to the arid conditions, as, for instance, in the absence of mangroves and in the prevailing character of the plants of the sea-border, cacti, thorny plants, and such like. For my information on this exceedingly interesting tract of coast, which is the test-ground of the Humboldt current theory, I am indebted to the papers of Baron von Eggers (see end of this volume) and to Mr. F. P. Walker, of the Central and South American Telegraph Company’s Station at Santa Elena, who very kindly communicated with me by letter. Some additional remarks are given in [Note 73], and my own observations on the temperature of the Humboldt current from Antofagasta northward are summarised in [Note 74].

Before quitting the Ecuador coast a word may be said relating to the prediction of Villavicencio that the climate of this sea-border will assimilate itself to that of the rainless coasts of Peru. This is, I believe, also the opinion expressed by Dr. Wolff in his Geografia y Geologia del Ecuador (Leipzig, 1892); and it is referred to by Mr. Webster in his article on Ecuador in the seventh volume of the Encyclopædia Britannica. There is a prevailing impression amongst the more observing residents that I met in the Ecuadorian province of Eloro, on the Peruvian border, that the country is drying up. A few pages back I have described how in the Machala district of this province the mangrove-belt passes landward into an arid region suggestive of the sea-border of Peru. This transition is startling to one who expects to find behind the mangrove-belt, as he would find in most parts of the world, a humid region where Nature revels in the rank luxuriance of plant-growth. This is, however, not always the case, since on the lee or dry sides of the large islands of Fiji the mangrove-belt is backed by extensive arid plains, for an explanation of which, as I have shown in [Note 22], we have to appeal rather to the hygrometer than to the rain-gauge. This is true also of Ecuador; but whilst the reason is intelligible enough in Fiji, it only carries us a step farther back in the case of the Machala plains in Ecuador. These plains are continuous with similar districts across the Peruvian border where they reach the coast; and if the reader will refer again to my description of the section of the mangrove-belt and the plains in its rear from Puerto Bolivar to Machala, he will incline to the view that the desiccation of the sea-border of Ecuador is now in progress.

Evidence of a more direct nature could doubtless be supplied by those who have long resided on the coast of Ecuador, and in illustration I will give an extract from one of Mr. Walker’s letters dated May, 1904, from Santa Elena.—“The rainfall here might for the last ten years be put down at two showers per year. It is said that the last good rainy season was in 1891. The inhabitants say that formerly it always rained enough to make the grass grow every year, but during the eleven years I have been here there appears to be a marked falling off of the rainfall.”

It has been only possible to touch the fringe of this interesting question here; but from the standpoint of the study of the littoral flora of the west coast of South America it is of some importance. Immediately behind the epoch of the present marine molluscan fauna of this coast there lies an age when, as we learn from Philippi, the shells of Chile were more akin to those of the Atlantic and Mediterranean faunas than to those now found on the Chilian coasts. The transition is a sudden one; and amongst other explanations of this strange transformation Suess suggests the sealing up of a communication through the Panama isthmus by volcanic eruptions and the appearance of the Humboldt current (Das Antlitz der Erde, French edit. by Margerie, ii. 825). May it not be, my readers may ask, that the west coast of South America is still in the age of progressive sterility; and that before this age began Peru possessed a normal tropical strand-flora? It has been remarked in [Chapter VIII.] that the same species of mangroves occur on both the Atlantic and Pacific coasts of America, and that at all events their present distribution belongs to an age when the Gulf of Mexico was in communication with the Pacific Ocean. May we not, again, suppose that in that age the mangroves extended far south on the coast of Peru, just as they do now on the coast of Brazil?

Coral reefs are stated not to exist in tropical latitudes on the west coast of South America in our own day; but we might almost expect that at the close of the Tertiary period, and perhaps before the appearance of the Humboldt current, they existed with the mangroves on the coast of Peru. As bearing on the subject of a change of climate on that coast in times geologically not remote, I may allude to the circumstance, which is discussed more in detail in [Note 75], that I found, sometimes in fair quantity, blocks of massive coral, long since dead, much pierced by boring shells, and in places undergoing a chemical change, at Arica (lat. 18° 25ʹ S.), at Callao (12° 3ʹ S.), and at Ancon (11° 45ʹ S.) on the coast of Peru.

These masses, which varied from a few inches to two or three feet in size, gave me the impression of having been torn off the bottom, in some cases in recent times, in others perhaps centuries ago, by the huge sea-waves that from time to time overwhelm this coast. At Ancon, where they were sufficiently abundant to be used for bordering the flower beds in the hotel garden, they were most numerous in the vicinity of a rocky spur of andesite that protruded from the beach between the tide levels and was more or less covered at high water. A few paces inland from the beach some of these coral masses, evidently stranded long ago, were undergoing that queer process of disintegration which everything calcareous seems to undergo on the beaches and plains of this almost rainless coast. Like the bones of the Incas lying bleaching on the neighbouring plains, like the sea-shells and bones of bird and beast cast up long ago on the beach, they were falling to powder where they lay, and the coral fragment lay often in the midst of its own débris. The blocks on the beach proper were for the most part still hard and compact, and the same may be said of those observed on the beaches of Callao and Arica.

The corals were quite different from those with which I was familiar in the reefs of the Pacific islands, and, bearing in mind the known distribution of coral reefs, I was a little dubious about them. Accordingly I sent some specimens to the British Museum, and Mr. Jeffrey Bell has kindly informed me that they seem to be decayed and much injured perforated examples of Porites. When powdered they effervesce in an acid, but the bulk of the material remains undissolved.

No more eloquent testimony could be afforded of the rainless climate than these corals crumbling on the Ancon plains when washed a few paces inland from the beach. They could be noticed in all stages of disintegration from the block surrounded by a little line of disintegrated material, representing the initial products of its own decay, to the crumbling mass, almost friable in the fingers, that was lying in the midst of its own dust and loose polyp-tubes, and finally to the little mound of débris that alone remained. Mr. Darwin, in his Journal of Researches (chap. xvi.), refers to a similar process of decay in the elevated shell-beds of San Lorenzo, off the coast of Callao. On the higher terraces a layer of saline powder, consisting of sulphates and muriates of lime and soda but with very little carbonate of lime, was the sole indication of the shell-beds. Dry climatic conditions at the sea-border evidently favour, as he observes, the early decay of exposed calcareous remains.

The Shore-plants and Stranded Seed-drift of the Panama Isthmus.

I spent two days at Panama and two days at Colon in examining the neighbouring beaches and estuaries of the Pacific and Atlantic coasts of the isthmus. On the Panama side the mangrove-belt was formed on the seaward border of “mangle chico” (the small prevailing type of Rhizophora mangle), Laguncularia, and Avicennia; whilst behind it passed into extensive swampy tracts occupied by the Swamp Fern (Chrysodium aureum), Hibiscus tiliaceus, and other plants. On the Colon or Atlantic side the mangrove-belt had precisely the same composition and presented the same species, Rhizophora and Avicennia usually forming the outposts on the reef-flat, whilst Laguncularia was abundant in the rear. In the estuary of the Rio Chagres, Rhizophora and Laguncularia were abundant near the mouth, and Chrysodium aureum and Hibiscus tiliaceus by the waterside higher up. Dr. Seemann, in his volume on the botany of the voyage of H.M.S. Herald, observes that the species of Laguncularia common on both the Atlantic and Pacific coasts of the Panama isthmus is L. racemosa. This species differs in the form of its fruit from the Ecuador tree. Laguncularia racemosa, Rhizophora mangle, and I may add Anona paludosa and Conocarpus erecta, are all plants of the mangrove-formation that occur not only on the Pacific and Atlantic coasts of America but also on the west coast of Africa. It is likely, I may add, that the “mangle grande,” the Ecuadorian type of Rhizophora mangle, exists in the Panama isthmus, since in the higher part of the estuary of the Chagres I found trees approaching it in characters.

Amongst the plants growing on the Panama beaches I noticed Canavalia obtusifolia, Hibiscus tiliaceus, and Ipomœa pes capræ, all of which occur also on the Atlantic side of the isthmus. The Manchineel (Hippomane mancinella), found also on the Atlantic side of the continent, grows on the Panama beaches. Its fruits, which look like crab-apples, lose their outer fleshy covering when drying on the sand. Not being familiar with this poisonous tree, I allowed some of the milky sap of the fruits to touch the skin, and suffered great pain for five or six hours. The fruit possesses an inner coat of air-bearing cork-like tissue; and the stone, if I may so term it, thus acquires great floating power. I kept some afloat in sea-water for five weeks, and no doubt they will float for months.

The seed-drift to be observed stranded on the beaches and floating in the estuaries on both sides of the isthmus is, generally speaking, the same—a circumstance of great importance in plant-distribution, since we can here see rivers bringing down the same seeds from the same “divide” to the shores of the Pacific and Atlantic oceans. In the case of a plant like Entada scandens, which grows in the interior, this is a matter of much interest, as it thus possesses here a centre of dispersal from which its seeds can be carried by the currents eastward to the West African coast and westward across the Pacific to Malaya and (given time) around the shores of the Indian Ocean to the East African coast. In describing the possible routes of dispersion from this centre I have described the distribution of the species.

I am indebted to Mr. Holland, of the Kew Museum, for the identification of some of the drift-seeds and fruits collected by me on the isthmus, those identified by him being followed by the letter H. On the beaches and floating in the estuaries on both sides of the isthmus I found Rhizophora seedlings; seeds of Entada scandens and Mucuna urens (medic.), H.; seedvessels of Spondias lutea (Linn.), H.; Prioria copaifera (Griseb.), H., with decayed seed; and the empty nuts, 1¹⁄₂ to 2 inches in size, of more than one species of Astrocaryum, H. Although in the case of the two last-named genera the seedvessels were useless for dispersal, being evidently brought down from the interior by the rivers, they serve to illustrate the important principle that the rivers bring down the same seed-drift on both the Atlantic and Pacific coasts of Central America. Mr. Hemsley includes amongst the seed-drift stranded on the coast of Jamaica the seedvessels of Spondias (probably S. lutea) and of Astrocaryum (Bot. Chall. Exped., iv. 299, 304).

Those of Spondias lutea were found by me floating in the Guayaquil River and stranded on the beaches of Ecuador and of the Pacific and Atlantic coasts of the Panama isthmus. This is the Hog-plum, which in tropical America and the West Indies is both wild and cultivated. Its buoyant “stone” has a covering of cork-like air-bearing tissue. This is a remarkable case of non-adaptation in the matter of buoyancy. The seedvessels cut across contained sound seeds; and they are provided with the essential qualities of “long floaters.”

Summary.

(1) The strand-district of the west coast of South America is divided into four zones:—

(a) The Convolvulus soldanella zone of Southern Chile.

(b) The Desert or Plantless zone of Northern Chile.

(c) The Sesuvium zone of Peru.

(d) The Mangrove zone of Ecuador and Colombia.

(2) The mangroves do not extend south of Ecuador or, more strictly, south of Tumbez (3° 30ʹ S.).

(3) The absence of mangroves on the tropical coasts of Chile and Peru is attributed to the Humboldt current, which has so influenced the climate that it has converted the sea-border of North Chile into a desert and that of Peru into a region of semi-sterility.

(4) It is considered that this has been effected through the prevailing winds acquiring drying qualities on crossing the cold waters of the current in tropical latitudes.

(5) To establish this it is shown that when the Humboldt current leaves the coast at Cape Blanco mangroves thrive in the Gulf of Guayaquil, and that when it strikes the coast again near Santa Elena Point and courses along that seaboard to the equator we find the Peruvian conditions of semi-sterility reproduced.

(6) The probability that the arid climate of Peru is in our own time extending northward into Ecuador is pointed out; and from the presence of old coral blocks on the Peruvian beaches it is considered likely that when these corals throve the mangroves extended far down the coast of Peru.

(7) It is shown from the presence of the same species of mangroves on the Pacific and Atlantic coasts of America that there must have been, not long ago, a communication between these two oceans across Central America; but it is at the same time observed that this could not be inferred from shore-plants with buoyant seeds that, like Entada scandens, occur inland, since, although they occur on both sides of the continent, their distribution can be explained by the transport of their seeds by rivers to the Atlantic and Pacific coasts, such as we see in operation on the Panama isthmus in our own day.

(8) Stress is laid on the great development of mangroves in the Gulf of Guayaquil and in the Guayas estuary; and it is pointed out that there are in this locality two varieties of Rhizophora mangle, a large and a small variety, the first approaching in some of its characters the Asiatic species, R. mucronata, and being akin also to the seedless intermediate form of Fiji.

(9) Amongst other matters dealt with in this chapter are the floating seed-drift of the Guayas or Guayaquil River and the shore plants and stranded seed-drift of the Panama isthmus. From the locality last named we learn that rivers bring down from the interior to the Atlantic and Pacific coasts much the same seed-drift, and that from this centre littoral plants with buoyant seeds can be distributed over the whole tropical zone.

CHAPTER XXXIII
SEED-DISPERSAL AND GEOLOGICAL TIME

The shifting of the source of the Polynesian plants from the New to the Old World.—The floral history of Polynesia stated in terms of geological time.—The suspension of the agencies of dispersal in later periods.—Parallel differentiation in the course of ages of climate, bird, and plant.—New Zealand.—Insects and bats as agents in plant-dispersal.—The effective agency of sea-birds in other regions.—The observations of Ekstam.—The Spitzbergen controversy.—The efficacy of ducks as distributors of aquatic plants.—Summary.

In the matter of the dispersal of seeds by birds in the tropical Pacific, there are at least two questions which my readers must have frequently put to themselves. The one would be concerned with the shifting of the source of the Polynesian plants from America to the Old World, which occurred probably near the close of the Tertiary period. The other would be connected with the suspension of the work of dispersal over a large portion of Polynesia, which has become more and more pronounced as we approach our own day.

Suggested Cause of the Shifting of the Source of the Polynesian Plants from the New to the Old World.—In previous chapters we have discussed the various epochs in the plant-stocking of these islands. There was first the age of Coniferæ, in which the islands of the Western Pacific were only concerned, an age prior to the appearance of the volcanic groups of the Tahitian and Hawaiian regions, and placed in the Secondary period. Then followed, in the Tertiary period, after the birth of Hawaii and Tahiti, and when the island-groups of the Western Pacific were mainly submerged, the general dispersion from America of the Compositæ, Lobeliaceæ, and other orders, now represented only by genera peculiar to the Hawaiian and Tahitian islands. Last of all, towards the close of the Tertiary period, when the island-groups of the Western Pacific had re-emerged, a general dispersion of Old World plants, mainly Malayan, took place over all the present archipelagoes of the tropical Pacific.

Since the climate of Hawaii must have, to a great extent, shared in the vicissitudes of the continental climates of the northern hemisphere before, during, and after the Glacial epoch, it is assumed that in the Ice Age no tropical plants reached the group, and that only the plants now represented in its mountain-flora could have then reached there. The area of active dispersion of tropical plants was pushed far south. During the Ice period, Indo-Malayan plants doubtless crowded into the equatorial region of the Western Pacific; but, cramped and confined within this limited area, they were cut off by a climatic barrier from the cool latitudes of Hawaii. As the cold ages passed away, migratory birds, confined during that period to the southern hemisphere, would extend their ranges north, sometimes reaching Hawaii, and transporting to it the seeds of New Zealand and Antarctic genera, now represented by endemic species on its mountain-slopes. The Indo-Malayan plants, with the increasing warmth in the climate of the northern hemisphere, would overrun the Pacific, set free from their prison in the south-west portion of that ocean. Dispersal, we might imagine, would be at first very active over the whole ocean.

My point is, then, that whilst the Malayan era of the plant-stocking began after the Ice Age in the northern hemisphere, the dispersion of the New Zealand and Antarctic genera over the Pacific took place during that period; whilst, as before noticed, the dispersion of the Compositæ, Lobeliaceæ, and other orders, represented now in Hawaii by endemic genera, would be pre-Glacial and well back in the Tertiary epoch. I would, therefore, suggest the following scheme, in illustration of the floral history of the tropical Pacific.

(1) The Age of Conifers of the Western Pacific during the Mesozoic period, and before the appearance of the Hawaiian and Tahitian archipelagoes.

(2) The Age of Compositæ and Lobeliaceæ, and of other genera. This is an era of American plants, and it is referred to the Tertiary period. In it only the newly-formed Hawaiian and Tahitian groups shared, the islands of the Western Pacific being largely submerged.

(3) The Age of Malayan plants, regarded as mainly post-Glacial, and subsequent, therefore, to the re-emergence of the Western Pacific islands at the close of the Tertiary period.

Dispersion then was general over the Pacific. The distribution of the New Zealand and Antarctic genera, plants that take a subordinate part in the floras of the Pacific islands, is regarded as having occurred during the glaciation of the northern hemisphere.

On the Suspension of the Agencies of Dispersal in the Tropical Pacific.—If the remark of Drake del Castillo that genera possessing only non-endemic species in the Pacific islands owe their presence in this region to existing agencies of dispersal looks something like a truism, we must remember that, assuming Nature to be uniform in her methods, it involves not merely the original co-operation of the same agencies with genera that own only peculiar species, but also the subsequent suspension of the work of these agencies.

The nature of the connection between freedom of dispersal and specific differentiation is well brought out by Beccari in contrasting the species of Ficus and the palms of Borneo; whilst out of fifty-five species of Ficus collected by him in that island, 30 per cent. were apparently peculiar, 85 per cent. of the 130 Borneo palms had not been found elsewhere. In the English edition of his Nelle Foreste di Borneo he says that “the explanation lies in the fact of the facile dissemination of the various species of Ficus through the agency of birds, an explanation which applies to all trees which produce edible fruits specially relished by animals.” He shows, also, that the same principle applies within the limits of the genus Ficus, since of those Bornean species known to him as belonging to the section Urostigma, which possesses fruits most preferred by birds (pigeons, hornbills, &c.), nearly all (fourteen out of sixteen) are found elsewhere; whilst of ten species belonging to the section Covellia, where the fruits are more or less hidden and inconspicuous, and with difficulty discovered by birds that would effectively distribute the species, four, at the most, are found elsewhere. “Such facts,” he goes on to say, “show that in tropical countries the various kinds of Ficus are, to a large extent, biologically connected with birds, which, perhaps, on their part, also owe some of their peculiarities in the shape of the bill or in the plumage to the nature and coloration of the fruits which form their food.”

Whilst Dr. Beccari as a botanist lays especial stress on the biological connection in Malaya between the plant, as illustrated by the genus Ficus, and the bird, Mr. Perkins, as a zoologist, is similarly emphatic on the biological connection in Hawaii between the bird, as illustrated by the peculiar family of the Drepanididæ, and the plant. The plants here are the arborescent Lobeliaceæ and the Freycinetias. To the flowers of the arborescent Lobeliaceæ the nectar-feeding Drepanids are particularly partial; and the development of the extreme forms of these birds, as Mr. Perkins observes in the Fauna Hawaiiensis, “is not comprehensible without a knowledge of the island flora.” Not only does he point to the modifications in the form of the bill of the bird in connection with the tubular form of the flowers; but in at least one species of these arborescent Lobeliaceæ he shows that it is dependent on the Drepanid for its fertilisation, and he inclines to the view that changes such as that of lengthening of the bill may have taken place side by side with the increasing length of the tubular flowers. In connection with the Freycinetias of Hawaii, Mr. Perkins regards the bill of the Ou, a finch-like Drepanid of the genus Psittacirostra, as “entirely formed and adapted for the purpose of picking out the component parts” of the fruiting inflorescence.

That in an isolated island-group birds and plants often “differentiate” together is a fact well known in distribution. In Hawaii, for instance, as I learn from Mr. Perkins, quite 45 per cent. of the birds are peculiar; whilst according to Dr. Hillebrand 80 per cent. of the flowering plants are confined to the group. Then, again, in the Galapagos Islands, half of the plants and two-thirds of the birds are confined to that archipelago. At the other end of the series we have the Azores, with about a tenth of its plants peculiar, and about 4 per cent. of its birds peculiar to the islands, and Iceland with no endemic plants and, as far as I can gather, few peculiar birds.

Accepting Mr. Charles Dixon’s view (The Migration of Birds, 1897) that specific differentiation does not occur along lines of migration, we must assume that the differentiation of the avifauna of an isolated group like Hawaii began with the breaking off of its regular communication through birds with the outside world. I do not consider that in the past these Pacific archipelagoes received their birds in any haphazard fashion, as, for instance, in the guise of stragglers that had lost their way. From the circumstance pointed out to me by Mr. Perkins that 25 of the 67 genera of Hawaiian birds are peculiar, we must postulate a high antiquity for the bird fauna dating far back into the Tertiary period. Mr. Perkins, who kindly supplied me with his general views of the nature of the Hawaiian fauna, tells me that it is “positively oceanic-insular and could be continental only on the supposition that everything continental had been at some time destroyed and that the group had been subsequently re-stocked as would any oceanic island.”

The view naturally presents itself that in past ages birds in the Pacific were much more uniform in their characters, and the agencies of dispersal far more active in their operations and far more general in their range than in more recent times, “It may be accepted (says Mr. Dixon) as an axiom of geographical distribution that all existing species are surviving relics of more ancient forms or ancestral types, whose dispersal in a remoter past was more continuous, and whose affinities and characteristics were therefore more homogeneous.” I assume that in past ages the differentiation of birds has largely been favoured by differentiation of climate acting through the limitation of their ranges. To these changes, plants, so often biologically connected with birds, have largely responded.

There is, of course, no difficulty in imputing to birds the capacity of reaching Hawaii in the mid-Pacific, and there are many regular migrants now (sea-birds, waders, ducks, &c.). The only difficulty is in the estimation of the time occupied in the trans-oceanic journey. According to Gätke the journey, which is 1,500 to 2,000 miles, ought to be accomplished within the limit of fifteen hours, which he regards as “the longest spell during which a bird is able to remain on the wing without taking sustenance of any kind.” As he considers that a bird might cover the 1,600 miles between Newfoundland and Ireland in nine hours (Heligoland as an Ornithological Observatory, p. 140), the Hawaiian traverse would offer to him no difficulties. It has frequently occurred to me in this connection that in ancient times, when the volcanoes of the mid-Pacific were in full activity, their light at night-time would have often given a direction to the migrating bird, and that they might have sometimes determined the line of migration across the Pacific.

It has not been possible to discuss here the capacity of pigeons to cross an ocean, a subject bearing directly on the floras of all the Pacific groups (excepting Hawaii, which possesses no indigenous Columbæ) and as concerning these islands generally presenting no difficulty. Dr. H. de Varigny, who amongst his other studies has long displayed an active interest in plant-dispersal, has directed my attention to two very important papers on the flight of pigeons in the Revue Scientifique, one by M. A. Thauziès (April 30, 1898) and the other by M. M. Dusolier (Nov. 28, 1903). That land birds, as well as swimmers and waders, cross the Atlantic is well known, and in this connection the reader might profitably consult Prof. Heilprin’s Geographical Distribution of Animals (vol. 58, Internat. Sci. Ser. 1887).

Much might be said of these matters, but it would be out of place here; and I will content myself with stating the view above indicated that the suspension of the agencies of seed-dispersal over the Pacific is probably connected with a general principle affecting the whole plant-world. The tendency in the course of ages has been towards the differentiation of climate, bird, and plant, the range of the bird being largely controlled by the climate, and the range of the plant being mainly dependent on the range of the bird.

It is evident that in some cases the plants themselves may make the endemism of a flora more pronounced by creating their own difficulties and by standing in the way of their own dispersal to outside regions. It has been shown that some of the endemic Hawaiian genera (see [Note 68]) have deteriorated in their capacity for dispersal by birds; and similar remarks are made with reference to the genera Sicyos (page [365]) and Eugenia (page [350]). Genera with stone fruits like Elæocarpus possess in the different species stones of various sizes, some of them suitable in point of size for conveyance in a bird’s body over an ocean, others so large that one could only predicate for them a limited capacity for distribution by birds over a few hundred miles of sea. One, for instance, could safely assume that species of Elæocarpus, with stones an inch and over in size, that occur in Fiji and Hawaii, are not suited for distribution over an ocean now; whilst other species found in New Zealand and Rarotonga have stones less than half this size, which are quite fitted for distribution by birds over broad tracts of ocean (page [337]).

This brings us to discuss the relative difficulties presented from the dispersal-standpoint by the forest floras of Hawaii, Fiji, and New Zealand. It is with the forest floras that nearly all the difficulties of distribution lie; and I hope I shall not be considered presumptuous, or at all events too heterodox, in expressing the opinion that judging from the details given in Kirk’s Forest Flora of New Zealand those islands present no greater difficulties for the student of plant-distribution, if we exclude the Coniferæ, than either Fiji or Hawaii. Indeed, even with the Conifers included, New Zealand presents fewer problems than Fiji, and Hawaii has its own special difficulties connected with the inland species of the Leguminosæ. There is, on the other hand, no special New Zealand difficulty. It possesses the Conifers in common with Fiji; and it shares with Fiji and Hawaii genera like Elæocarpus and Sideroxylon, that take a foremost place amongst the trees of the Pacific forest flora presenting puzzling points to the student of distribution. The existence of Elæocarpus in New Zealand admits of a simpler explanation than the occurrence of the same genus in Hawaii. Pandanus in Fiji is a more difficult genus from the standpoint of dispersal than Corynocarpus in New Zealand, and in fact, than any of the non-coniferous genera of forest trees in that region.

Whilst it is likely that birds of the genus Porphyrio have, up to almost recent times, been active in distributing the seeds of New Zealand plants outside the region (see p. [296]), it would seem that the fruit-pigeons, as represented by a solitary peculiar species of Carpophaga, have long since ceased to be active in this direction. It is true that Sir W. Buller gives a long list of trees, including Corynocarpus, Elæocarpus, Litsea, Olea, Podocarpus, and many others, the fruits of which are appreciated by this pigeon; but since the bird is confined to this region its efforts in plant-dispersal possess only a local interest. Mr. G. M. Thomson, indeed, has expressed the opinion (Trans. and Proc. N.Z. Instit. vol. 33) that in recent times not a single plant has been added through the agency of birds to the New Zealand flora. Besides the regular migratory birds two cuckoos only reach the region, the one from Australia and the other from Polynesia; whilst Australian birds which had managed to survive the long flight across the ocean have been met with only at times on the west coast of the North Island. From the standpoint adopted in this work we should have expected that, with the exception of current-dispersed plants, New Zealand would be out of touch with the world outside. Varied only by occasional inrushes of plants, its history, dating back to the Mesozoic age, has been one of insular isolation.

When, however, we apply the principles of plant-dispersal in the Pacific, deducible from the study of the Hawaiian flora, we learn that the stocking of New Zealand with its plants could have been carried out (with the exception of the Coniferæ and a few other genera like Fagus that are in a geological sense ancient denizens in this region) by the agencies that stocked Hawaii with its flora. New Zealand genera like Elæocarpus, Sideroxylon, Sophora, etc., that are represented in the forests of Hawaii, could not be taken to illustrate any former continental connection. If we, so-to-speak, put the New Zealand forest flora in the Hawaiian sieve, all will pass through with the exception of Fagus, the genera of the Coniferæ, and plants of similar history in high southern latitudes. This residuum belongs more to the palæobotanist than to the student of means of dispersal.

I should be inclined to think that the tropical genera of the New Zealand flora, more especially of the forest flora, reached that region during the glaciation of the northern hemisphere, when the Indo-Malayan plants were, so-to-speak, cornered in the Western Pacific. Yet it must be noted that these are, as a rule, genera that either display an indifference to the varying thermal conditions of different latitudes or are known to at times extend their range beyond the tropics. Thus Elæocarpus and Freycinetia are equally at home in the temperate rain-forests of New Zealand and in the tropical rain-forests of Polynesia and Malaya; whilst widely-spread tropical genera like Pittosporum and Peperomia, that occur in New Zealand, exhibit their power of adaptation to varying climates by extending outside the tropics in other regions and by their vertical range in the Hawaiian mountains, where they are found alike at low elevations a few hundred feet above the sea and at altitudes of 6,000 or 7,000 feet.

All these plants, however, are in a relative sense recent intruders. When the student of dispersal looks at the long list of the conifers of the New Zealand forest flora and reflects that he knows but little of their means of dispersal, and that if his acquaintance were far greater it would not avail him much, he has no choice but to take his place behind the earlier investigators of the flora, and to see in these trees evidence in favour of a remote continental period, probably referable to the mesozoic age.

A Discussion of some Means of Dispersal.—Not many authors seem to have discussed the possibility of insects as agents of seed-dispersal in the Pacific. They appear to me quite suited for transporting the spores of ferns and lycopods as well as the minute seeds of plants like the orchids and the begonias. Darwin, who allowed few possible means of dispersal to escape his notice, procured the germination, as my readers will remember, of grass seeds found in the dung of Natal locusts. When on the barren summit of Mauna Loa, I noticed that the recently dead bodies of some butterflies, that had been carried up the slopes from the forests below by the “southerly updraught,” were already attacked by bugs, parasites that must have been transported from the lower regions by some of the numerous larger insects that are blown up the slopes.

In [Note 61] the occurrence of the wind-blown insects on the summit of Mauna Loa is described. That insects can be transported into the upper regions of the atmosphere by ascending air-currents was long ago remarked by Humboldt, and the subject has been discussed with his usual acumen by Whymper (Travels amongst the Great Andes of the Equator). Carried along in the higher air-currents these insects might finally be deposited at places far distant from their home. One reads occasionally extraordinary accounts of a rain of insects. A very circumstantial account was given to me when I was on Keeling Atoll of a shower of dragon-flies that fell on the islands, their remains being found in quantities in the lagoon. Dragon-flies, it is known, are often found at sea far from land, and one species has been observed nearly all over the world, including the Pacific islands. In this connection it is interesting to recall Mr. M’Lachlan’s remark in his article on the dragon-fly in the seventh volume of the Encyclopædia Britannica that some of the earliest fossil forms seem to have been washed ashore after having been drowned at sea.

Another creature that has been often ignored as a possible agent in seed-dispersal is the bat. Bats are found all over the world, including the oceanic groups, and one can scarcely doubt that they must have often transported seeds, at all events in their hair. They are found at times high up in mountainous regions, and Sir H. Johnston, in his recent work on the Uganda Protectorate, refers to the occurrence of bats at an altitude of 13,000 feet. The large frugivorous bats (Pteropidæ) are known to be very destructive feeders; but I doubt whether they swallow the seeds. Dr. Warburg, as is remarked in [Chapter XXV], says that they feed on the flowers of Freycinetias, and I have already observed that they visit the flowers of Geissois ternata in Fiji (p. [394]). In this fashion Dr. Warburg regards them as agents in pollenisation; and it seems to me that if, as appears likely, they are attracted by trees with large, brightly-coloured flowers, they would often aid in the dispersal of the minute seeds of trees like Metrosideros.

Until recently sea-birds, and some particular birds of passage, have been generally considered as only fitted for dispersing seeds in their plumage. That they can also transport seeds inside their bodies is shown below. Dr. R. Brown in his book entitled Our Earth and its Story, 1888, gives a general account of plant-dispersal with numerous references to the Literature on the subject. On the direct route between Scotland and Cape Farewell in Greenland snow-buntings (Plectrophanes nivalis) and other birds of passage frequently alight, as we are told, on ships when hundreds of miles from land. Dr. Brown says that when taking this voyage he examined dozens of these birds. Only in one case, however, did he find any seeds, namely, in the case of a snow-bunting which carried, attached to its plumage, an achene of, perhaps, a Ranunculus, and in its gizzard a seed like that of a Suæda. My discovery of a small, hard seed in the gizzard of a Cape-pigeon (Daption capensis) 550 miles east of Tristan da Cunha has been referred to by Mr. Hemsley in his introduction to the Botany of the “Challenger” Expedition (p. 45). On p. [188] I have mentioned the probable dispersal of the seeds of Cæsalpinia by frigate-birds and boobies; and in [Note 59] reference is made to some large seeds found in the crop of the Fulmar petrel.

Gulls, when they nest at the coast, where the sea-thrift (Armeria vulgaris) and the sea-campion (Silene maritima) thrive, or inland amongst the heather-covered slopes, must often carry the seeds of these plants from place to place in their plumage (see Notes [15] and [16]); but, as shown below, they can also disperse plants with fleshy fruits which at times form their food. Gulls, geese, and arctic grouse take an important part in the dispersal of seeds in the cold latitudes of the northern hemisphere; and few things are more suggestive in this way to the student of distribution than the data supplied by Ekstam, Hesselman, Sernander, and others for the region including Spitzbergen, Nova Zembla, and Arctic Norway. The history of the discussion relating to the flora and fauna of Spitzbergen reproduces in its main features the various stages in the controversy that has been waged in connection with the Pacific islands.

When Ekstam published, in 1895, the results of his observations on the plants of Nova Zembla, he observed that he possessed no data to show whether swimming and wading birds fed on berries; and he attached all importance to dispersal by winds. On subsequently visiting Spitzbergen he must have been at first inclined, therefore, to the opinion of Nathorst, who, having found only a solitary species of bird (a snow-sparrow) in that region, naturally concluded that birds had been of no importance as agents in the plant-stocking. However, Ekstam’s opportunities were greater, and he tells us that in the craws of six specimens of Lagopus hyperboreus shot in Spitzbergen in August he found represented almost 25 per cent. of the usual phanerogamic flora of that region, in the form of fruits, seeds, bulbils, flower-buds, leaf-buds, &c. This observer now also realised the importance of gulls and geese in dispersing certain types of plants in those latitudes. Species of Larus, he says, consume greedily all kinds of berries, and especially those of Empetrum nigrum, the stones of which were found uninjured in their droppings by Professor Lagerheim in Arctic Norway. Geese, as we are also informed, are hearty plant-eaters in Spitzbergen; and Ekstam found in their droppings the fruits of Oxyria digyna as well as an abundance of uninjured bulbils of Polygonum viviparum, some of which proved to be capable of growth (See Ekstam in Tromso Museums Aarshefter, vols. 18 and 20, 1895-7).

The result of Ekstam’s observations in Spitzbergen was to lead him to attach a very considerable importance in plant-dispersal to the agency of birds; and when in explanation of the Scandinavian elements in the Spitzbergen flora he had to choose between a former land connection and the agency of birds, he preferred the bird.

I have gone into some detail in this matter because the Spitzbergen controversy in some respects might have equally centred around New Zealand or some of the large continental islands of the tropical Pacific. There is at first the endeavour in the absence of precise knowledge to disregard the bird and to look for a land connection. With the increase in our acquaintance with the efficacy of bird-agency in seed distribution there is the abandonment of such a view. In both localities, however, there are the same counter-indications of the insect faunas, and the same considerations are raised by the absence or presence of larger animals in the regions concerned. The principal difference lies in the frozen sea, and yet, strangely enough, it does not seem to affect the problem much. It would indeed appear that the questions raised by the floras and faunas of the Pacific islands are not peculiarly Pacific in their character; and it is probable that the difficulties here presented are repeated in one form or other in the case of large islands over all the globe.

On the efficacy of Ducks and other Waterfowl in the Distribution of Aquatic Plants.—It is highly probable that aquatic plants, like the beach plants distributed by the currents and the ferns and lycopods distributed mainly by the winds, have changed much less in the course of ages than the plants of the inland forest. This in all three cases is chiefly due to the uninterrupted freedom of communication by means of the dispersing agency.

Wild ducks and their kind are active agents in the distribution of the seeds of aquatic plants; but it is curious that the early experiments of Caspary went far to discredit them in this respect. As quoted by Dr. Schenck in his Die Biologie der Wassergewächse, 1886, he fed tame ducks with the seeds of water-lilies and found that in a short time they thoroughly digested the seeds. Those familiar with the seeds of our British species of Nuphar and Nymphæa will not be surprised at such a result; but, unfortunately, the inference drawn from this experiment has been by some extended to aquatic plants in general. Since the seeds or seed-vessels of some aquatic or semi-aquatic plants of the genera Potamogeton, Sparganium, &c., appeared to me to be quite fitted for conveyance without injury in a duck’s body, I made several years ago a number of observations on this subject, the results of which were published in Science Gossip for September, 1894.

Out of 13 wild ducks obtained in the London market and stated to have been sent from Norfolk and Holland, eleven contained in their stomach and intestines 828 seeds, which I thus classed:—

• 295 seeds of Sparganium in 8 birds
• 41 seeds of Potamogeton in 3 birds
• 270 seeds of Cyperaceæ in 5 birds
• 222 not identified

In the case of four birds the germinating capacity of the seeds was tested, and in three cases very successfully. The seeds of Potamogeton, Sparganium, and of the Cyperaceæ germinated readily in water, but few of them failing, the process beginning in a few days or a few weeks. At that time I was conducting an extended series of observations on the postponement of germination of the seeds of aquatic plants, the results of which were published in the Proceedings of the Royal Physical Society of Edinburgh for 1897. It was there shown that the seeds of these plants often postpone their germination to the second and even to the third spring. It thus happened that, whilst seeds obtained from the stomach and intestines of the wild duck germinated in a few days or weeks, I had to wait often a year and more for such a result with seeds in their ordinary condition. This was well brought out in another experiment made on a domestic duck, which I have described on page [369]. That wild ducks are to be regarded in the light of “flying germinators” is thus very evident.

Summary.

(1) In explanation of the shifting of the source of the Polynesian plants from the New to the Old World, it is suggested that during the glaciation of the northern hemisphere the Indo-Malayan plants entering this region were “cornered” in the tropical Western Pacific, and were only set free after the cold period had passed away, when they overran Polynesia.

(2) Whilst the age of the Conifers is placed in the Mesozoic period, that of the Compositæ is accredited to the Tertiary period, and the era of Malayan immigration followed the glacial epoch.

(3) The suspension to a great extent of the agencies of plant dispersal in the Pacific in later times is connected with a general principle affecting the whole plant-world. With the secular drying up of the globe the differentiation of climate, bird, and plant have gone on together, the range of the bird being mainly controlled by the climate and the range of the plant being largely dependent on the bird.

(4) Accepting Hawaii as entirely insular in its history, it is pointed out that the principles deducible from the study of its flora can be applied to the forest-flora of New Zealand, with the exception of the Conifers and some genera that are ancient denizens of Antarctic latitudes, and indicate a remote continental age dating back to the Mesozoic period. It is suggested that the Indo-Malayan element in its flora arrived there during the glaciation of the northern hemisphere.

(5) Insects and bats have probably been effective agents in seed-dispersal in the Pacific, and it is shown that sea-birds carry seeds in their stomach and intestines as well as in their feathers.

(6) It is shown that birds of the grouse family, gulls, and geese are active seed-dispersers in cold northern latitudes, and that the discussion of their influence in stocking Spitzbergen with its plants reproduces many of the points of the controversy concerning the floras of the continental islands of the South Pacific.

(7) The results of experiments and observations are cited to establish the efficacy of ducks in distributing the seeds of aquatic plants, the seeds ejected in their droppings germinating in a few days or weeks, whilst those remaining in the pond or river often do not germinate for a year or more.

CHAPTER XXXIV
GENERAL ARGUMENT AND CONCLUSION

The problems concerned in the study of the floras of the Pacific islands from the standpoint of dispersal are here approached through the buoyant quality of the seed and fruit; and it is shown when dividing the plants into two groups, those with buoyant and those with non-buoyant seeds or fruits, that there has been at work through the ages a great sorting process, by which the plants belonging to the group first named have been mostly gathered at the coast. Its operation may be also observed within the limits of a genus, where the species possessing seeds or fruits that float is stationed at the coast, whilst the species with seeds or fruits that sink makes its home inland.

When the principle here involved is applied to the British flora, it presents itself as part of a much wider principle, by which plants endowed with buoyant seeds and fruits have been stationed at the water-side, whether on a river-bank, or beside a lake or pond, or on a sea-beach. The broader principle proves in its turn to belong to a far larger scheme, in which the fitness or unfitness of a plant to live in a physiologically dry station appears as the primary determining quality, the xerophyte (the plant of the dry station), provided with buoyant seed or fruit, finding its way to the coast, and the hygrophyte (the plant growing under more moist conditions), that is similarly endowed, establishing itself by the side of the river, or the lake, or the pond.

When dealing with the general character and composition of the strand-plants of the tropical Pacific, it is shown that in Fiji the beach-plants often assert their primary xerophilous habit or fitness for occupying any dry station by extending into the inland plains on the dry sides of the islands. The Fijian shore-plants are divided into three formations, those of the beach, those of the mangrove-swamp, and those of intermediate stations on the borders of the swamps. The great majority of the Fijian shore-plants are dispersed by the currents. The Tahitian Islands, which are representative of Eastern Polynesia, lack the mangroves and most of the plants that grow at the margin of a mangrove-swamp; and their strand-flora is mainly composed of plants of the beach, such as are dispersed by the currents far and wide in tropical regions. The Hawaiian strand-flora is very meagre in its character, lacking not only the plants of the mangrove and intermediate formations, but almost all the large-fruited beach-trees of the South Pacific. Since Hawaii possesses but few current-dispersed shore-plants that are not found in the New World, reasons are given for the inference that such shore-plants were originally brought by the currents from America, and not from the South Pacific.

We are led on various grounds to the conclusion that tropical shore-plants distributed by currents belong to two great regions, the American including the west coast of Africa, and the Asiatic, or Old World Region, which includes the African east coast. It is held that America is so placed with regard to the currents, that it is a distributor, and not a recipient of tropical shore-plants dispersed by that agency. From this it follows that all cosmopolitan tropical beach-plants that are dispersed by the currents have their homes in America.

The results of observation and experiment are given to show that there is no direct relation between the specific weight of seeds and fruits and the density of sea-water. Yet, although the floating or sinking of a seed or fruit is but an accidental attribute, it has had indirectly a far-reaching influence not only on plant-distribution, but on plant-development. In accordance with this want of relation between the specific weight of seeds and fruits and the density of sea-water, the great variety of structures concerned with buoyancy are regarded in the main, after a detailed examination of their character, as not arising from adaptation. Rather, it is urged, is buoyancy connected with structures that now serve a purpose for which they were not originally developed. Nature, it is held, has never concerned herself directly with providing means of dispersal of any sort.

In the discussion of the relation between the littoral and inland Pacific floras, it is shown, as a result of the examination of those genera possessing both shore and inland species, that they have been on the whole developed on independent lines. Two special difficulties in explaining the modes of dispersal of plants of the Pacific islands here come into prominence. There is the Hawaiian difficulty, where with genera containing both shore and inland species only the last are found in Hawaii; and although the shore-plants are known to be dispersed by the current, the inland plants display little or no capacity for this or any other mode of dispersal. Here belong the Leguminous genera Canavalia, Erythrina, Mezoneuron, and Sophora, and the Apocynaceous genus Ochrosia; and it is assumed that the inland Hawaiian species are derived from a current-dispersed shore-plant that has since disappeared from the group. The Fijian difficulty is displayed in those genera where both coast and inland species occur in the islands, but no known existing means of dispersal across an ocean can be postulated for the inland plants, though the shore species are distributed by the currents. Of such genera Pandanus is the best example, and it is pointed out that this genus presents the same difficulty in the Mascarene Islands, in which case the agency of the extinct Columbæ is invoked.

As illustrating the methods of observation and experiment employed by the author, the Leguminous shore-plants Afzelia bijuga, Cæsalpinia bonducella, and Entada scandens are discussed at length; and in the chapters on the enigmas of the Leguminosæ in the Pacific it is pointed out that the behaviour of the plants of this order is a source of much perplexity, and that they conform to no single rule of dispersal.

Coming to the inland plants of this region, the Fijian, Tahitian, and Hawaiian groups are taken as the chief centres of distribution in the Pacific. After discussing the relative sizes, the altitudes, and the climates of these three archipelagoes, it is shown that Hawaii, on account of the far greater altitude of the islands, is characterised by a special mountain flora, and that it is comparable with Fiji, and to a great extent also with Tahiti, only as regarding the plants of the levels below 4,000 or 5,000 feet.

The first era of the plant-stocking is designated the Age of Ferns, and it is observed that, whilst in Hawaii nearly half of the ferns and lycopods are peculiar to that group, very few new species have been developed in the Fijian and Tahitian regions.

The next era in the floral history of these islands is represented in the first era of the flowering plants. This is indicated by the endemic genera, which are particularly numerous in Hawaii, relatively scanty in Fiji, and very few in Tahiti. On account of their preponderance, the era is designated the Age of Compositæ and Lobeliaceæ. The genera of these two orders, though mainly characteristic of Hawaii, are also to be found in the Tahitian region, but they are absent from the Fijian area. Chiefly American in their affinities, their dispersion over the Pacific took place during the Tertiary submergence of the archipelagoes of the Western Pacific, in which are included the groups of the Fijian area (Fiji, Samoa, Tonga). These early forms of Compositæ and Lobeliaceæ are often arborescent in habit; and it is observed that Tree-Lobelias also occur high up the slopes of lofty mountains in tropical regions, as in Equatorial Africa, under conditions similar to those prevailing on the slopes of the Hawaiian mountains, where the Tree-Lobelias, termed by Dr. Hillebrand “the pride of our flora,” abound.

The other Hawaiian endemic genera, marking the first chapter in the history of the flowering plants, arrange themselves in two groups, one chiefly American in general affinities, and containing highly differentiated Caryophyllaceæ, Labiatæ, &c.; the other largely Malayan, and indicating the close of the first era of the flowering plants, when the main source of the plants was shifted from America to the Old World. The Fijian endemic genera, which are few in number, miscellaneous in appearance, and disconnected in character, are regarded as having probably acquired their endemic reputation through their failure at their sources in the regions to the west.

The second era of the flowering plants is indicated by the non-endemic genera. Here we are concerned on the one hand with a mountainous flora mainly Hawaiian, in which genera from the New Zealand and Antarctic floras take a conspicuous part, and on the other with a low-level flora chiefly derived from Indo-Malaya, and including the plants of the lower slopes of Hawaii below 4,000 and 5,000 feet, and the floras in mass of Fiji and Tahiti.

On account of their lower altitude, the extensive mountain flora of Hawaii is but scantily developed in Tahiti, and is represented by a mere remnant in Fiji and Samoa. Two-thirds of the Hawaiian non-endemic mountain genera contain only species restricted to the group, and, although amongst these disconnected genera, Acæna, Gunnera, Coprosma, Lagenophora, &c., of the New Zealand and Antarctic floras take a prominent part, a large proportion of the genera like Ranunculus, Rubus, Artemisia, Vaccinium, and Plantago represent generally the flora of the north temperate zone on the summits of tropical mountains. The Tahitian mountain flora, scanty as it is when judged by the non-endemic genera, displays much kinship with the Hawaiian mountain flora; but this kinship is mainly confined to genera from high southern latitudes, such as Coprosma, Cyathodes, Astelia, &c. In the possession on its mountain slopes of the three genera of the Coniferæ, Dammara, Podocarpus, and Dacrydium, the Fijian region is distinguished from that of Tahiti and Hawaii; and it is assumed that they mark the site of a continental area in the Mesozoic period, when the Tahitian and Hawaiian groups did not exist.

The era of the non-endemic genera, in so far as it is concerned with the low-level flora of Hawaii and the floras in mass of the areas of Fiji, Samoa and East Polynesia, is termed Malayan, because many of the genera are thence derived. Here we are dealing with all the oceanic groups of the tropical Pacific, and not with a portion of them, as in the case of the Age of Coniferæ, in the Secondary period, that was limited to the Western Pacific, or in the case of the Age of Compositæ and Lobeliaceæ that was restricted during the Tertiary epoch to the Hawaiian and Tahitian regions. The first part of this era, as is indicated by the endemic species, is an age of complete isolation in Hawaii, and of partial isolation in the groups of the southern region. Amongst the genera typical of this period are Pittosporum, Gardenia, Psychotria, Cyrtandra, and Freycinetia. A later period in this era of the general dispersal of Malayan plants over the Pacific is one where the extremely variable or polymorphous species plays a conspicuous part, as represented in such genera as Alphitonia, Dodonæa, Metrosideros, Pisonia, and Wikstrœmia, the general principle being that each genus is at first represented by a widely ranging very variable species, which ultimately ceases to wander and settles down, and becomes the parent of different sets of species in the several groups.

The facts of distribution in this age of general dispersion are just such as we might look for in the case of a general dispersal over the oceanic groups of the Pacific, with the altitudes of the islands playing a determining part. But it should be remarked that the greater number of the genera that have entered the Pacific from the Old World have not advanced eastward of the Fijian region, half of the Fijian genera not occurring in the Hawaiian and Tahitian regions. The explanation of this is to be found, not in any lack of capacities for dispersal, but in a want of opportunities. The story of plant-distribution in the Pacific is bound up with the successive stages of decreasing activity in the dispersing agencies. The area of active dispersion, as illustrated by the non-endemic genera, at first comprised the whole of the tropical Pacific. It was afterwards restricted to the South Pacific, and finally to the Western Pacific only. The birds that carried seeds all over this ocean became more and more restricted in their ranges, probably on account of increasing diversity of climatic conditions. The plants of necessity responded to the ever narrowing conditions of bird-life in this ocean, and the differentiation of the plant and the bird have taken place together.

During the stages of decreasing activity in the dispersing agencies, the widely-ranging highly variable species continued to be an important factor in the development of new species in the different groups. The rôle of the polymorphous species has always been a conspicuous one in the Pacific.

Yet, as in the case of the Cyrtandras, it is shown that the display of great formative power within a genus is not a peculiarity of an insular flora; that the isolation of an oceanic archipelago does not exclusively induce “endemism,” but only intensifies it; that the development of new species may be nearly as active on a mountain in a continent as on an island in mid-ocean; and that this is equally true of a land genus, like Embelia, exposed to an infinite variety of conditions, and of an aquatic genus, like Naias, where the conditions of existence are relatively uniform all the world over.

In framing a scheme by which the eras of the floral history of the Pacific are brought into correlation with those of geological time, the age of the Coniferæ is placed in the Secondary period, that of the Compositæ and Lobeliaceæ in the Tertiary period, whilst the era of Malayan immigration is regarded as mainly post-glacial. The age of the Coniferæ is concerned only with the Western Pacific, since the Hawaiian and Tahitian islands had not then been formed. The age of the Compositæ and Lobeliaceæ is concerned only with Hawaii and Tahiti, since the islands of the Western Pacific were then more or less submerged. That of the Malayan plants affects the whole Pacific as at present displayed to us.

In the chapter on the viviparous mangroves of Fiji it is shown that both the Asiatic and the American species of Rhizophora (R. mucronata and R. mangle) exist in that group, and that there is in addition a seedless form, the Selala, which, although intermediate in character between the two other species, comes nearest to the Asiatic plant. Reasons are given for the belief that the Selala is derived from the Asiatic species (R. mucronata), not as the result of a cross but as connected with its dimorphism; and in support of this it is pointed out that on the Ecuador coast of South America, where only the American species exists, a dimorphism is also displayed, one of the forms approaching in several of its characters the Fijian Selala, though fruiting abundantly and bearing the impress of a closer connection with the typical American species than with the Asiatic plant. The view that Rhizophora mangle reached the Western Pacific from America is rejected, and it is considered that this species was originally as widely diffused in the Old World as in America, and that it now survives only in a few places in the tropics of the Old World. The results of detailed observations on the modes of dispersal and on the germinating process both with Rhizophora and Bruguiera are given; and the absence, as a general rule, of any period of rest between the fecundation of the ovule and the germination of the seed is established.

A special chapter is devoted to the significance of vivipary, and it is considered that a record of the history of vivipary on the globe is afforded in the scale of germinative capacity that begins with the seedling hanging from a mangrove and ends with the seed that is detached in an immature condition from an inland plant. It is suggested that with the drying up of the planet in the course of ages the viviparous habit, which was once nearly universal, has been for the most part lost except in the mangrove swamp, which to some extent represents an age when the earth was enveloped in cloud and mist and the atmosphere was saturated with aqueous vapour. The lost habit is at times revived in the abnormal vivipary of some inland plants, and traces of it are seen in the abnormal structure of the seeds of some genera of the Myrtaceæ, like Barringtonia, and in the seeds of genera of other orders. With the desiccation of the planet and the emergence of the continents there has been continual differentiation of climate resulting in seasonal variation and in the development of the rest-period of the seed.

With the secular drying of the globe and the consequent differentiation of climate is to be connected the suspension to a great extent of the agency of birds as plant-dispersers in later ages, not only in the Pacific Islands but over all the tropics. The changes of climate, bird, and plant have gone on together, the range of the bird being controlled by the climate, and the distribution of the plant being largely dependent on the bird.

The history of climate, the history of the continents and of the oceans, the history of life itself, but only in the sense below defined, all belong to that of a desiccating world, or rather of a planet once sunless and enveloped in mist and cloud, that through the ages has been drying up. Life’s types were few and the sea prevailed, and one climate reigned over the globe. With the diminution of the aqueous envelopes the continents began to emerge, climates began to individualise, and organisms commenced to differentiate, and thus the process has run on through the past, ever from the general to the special both in the organic and in the inorganic world.

The same story of a world drying up is told by the marine remains left stranded far up some mountain slope, or by the bird akin to no other of its kind that Time has stranded on some island in mid-Pacific. The bird generalised in type that once ranged the globe is now represented over its original range by a hundred different groups of descendants, confined each to its own locality. Climate, once so uniform, now so diversified, has by restricting the range of the bird favoured the process of differentiation, and the plant dependent on the bird for its distribution has in its turn responded to these changes.

The rôle of the polymorphous species belongs alike to the plant and to the bird. A species that covers the range of a genus varies at first in every region and ultimately gives birth to new species in some parts of its range. Then the wide-ranging species disappears and the original area is divided up into a number of smaller areas each with its own group of species. Each smaller area breaks up again, and forms, yet more specialised, are produced; and thus the process of subdivision of range and of differentiation of form goes along until each island in an archipelago owns its bird and each hill and valley has its separate plants. This is not the path that Evolution takes, since beyond lies extinction whether of plant or of bird. Such is the upshot of the process of differentiation exhibited in the development of species and genera in the Pacific Islands, or, indeed, in any oceanic groups. It can never do more than produce a Dodo or a Kiwi, or amongst the plants a Tree-Lobelia.

Evolution here and elsewhere is a thing apart from species and genera, which are but eddies on the surface of its stream. It is a scheme of life introduced into a much conditioned world, and adaptation in endless forms is the price it has had to pay. The whole story of life on this earth is a story of a sacrifice, of an end to be won, but of a price to be paid. Immortality is in the scheme, but death is the price of adaptation. The same theme runs through our conceptions of the spiritual life. There is the same duality, evolution adapting its scheme to the exigencies of the physical world, the good principle ever in conflict with the evil, and at times compelled to adapt itself to attain its ends. There is the tale of adaptation in the one case and of sacrifice in the other, and success is reached in both.

APPENDIX
List of Notes

[Note 1]. On the number of known species of Fijian flowering plants.

[Note 2]. The littoral plants of Fiji.

[Note 3]. Results of long flotation experiments on the seeds or seedvessels of tropical littoral plants.

[Note 4]. Table illustrating the degree of buoyancy of the seeds and fruits of inland Fijian plants.

[Note 5]. The inland Fijian plants possessing buoyant seeds or fruits.

[Note 6]. Table showing the degree of buoyancy of the seeds and fruits of some inland Hawaiian plants.

[Note 7]. Some inland Hawaiian plants possessing buoyant seeds or fruits.

[Note 8]. The pyrenes of Morinda.

[Note 9]. The buoyancy of the fruits of Calophyllum.

[Note 10]. The buoyancy experiments on British plants.

[Note 11]. The effect of sea-water immersion on the germinating capacity of seeds and seed-vessels.

[Note 12]. The buoyancy of the fruits of Galium aparine.

[Note 13]. The buoyancy of the seeds of Convolvulus sepium.

[Note 14]. Other long flotation experiments.

[Note 15]. The occurrence inland of Silene maritima.

[Note 16]. The buoyancy of the seeds or fruits of the British beach-plants that also occur inland.

[Note 17]. The buoyancy of the seeds or fruits of the British littoral plants that frequent salt-marshes and muddy shores.

[Note 18]. The buoyancy of the seeds or fruits of the British littoral plants that are confined to the beach.

[Note 19]. On germination in sea-water.

[Note 20]. On the maximum heights reached by some shore plants in their extension inland in Vanua Levu, Fiji.

[Note 21]. On the dwarfing of shore plants when extending inland in the “talasinga” plains in Vanua Levu.

[Note 22]. The “talasinga” plains of Vanua Levu.

[Note 23]. Schimper’s grouping of the Indo-Malayan strand flora.

[Note 24]. Grouping of some of the characteristic plants of the strand flora of Fiji.

[Note 25]. The strand flora of the Tahitian region.

[Note 26]. The Fijian shore plants not found in Tahiti.

[Note 27]. The intruders into the beach flora from the inland plants of Tahiti.

[Note 28]. The littoral plants of the Hawaiian islands.

[Note 29]. Botanical notes on the coast plants of the Hawaiian islands.

[Note 30]. The beach drift of the Hawaiian islands.

[Note 31]. The inland extension of the shore plants of the Hawaiian islands.

[Note 32]. The Fijian species of Premna.

[Note 33]. De Candolle’s list of plants dispersed exclusively by currents.

[Note 34]. The littoral plants of the eastern-most Polynesian islands.

[Note 35]. Distribution of the littoral plants with buoyant seeds or fruits that occur in the Fijian, Tongan, Samoan, Tahitian, and Hawaiian Groups.

[Note 36]. Hawaiian plants with buoyant seeds or fruits known to be dispersed by the currents either exclusively or with the assistance of frugivorous birds.

[Note 37]. On vivipary in the fruits of Barringtonia racemosa and Carapa obovata.

[Note 38]. On the temperature and density of the surface water of the estuaries of the Rewa River in Fiji and of the Guayaquil River in Ecuador.

[Note 39]. On the Pacific species of Strongylodon.

[Note 40]. Precautions in testing seed-buoyancy.

[Note 41]. The buoyancy of the seeds of Convolvulus soldanella in fresh-water and sea-water compared.

[Note 42]. On secular changes in sea-density.

[Note 43]. On the mucosity of small seeds and seed-like fruits when wet.

[Note 44]. Upon the effects of inland extension on the buoyancy of the seeds or fruits of littoral plants.

[Note 45]. Tabulated results of the classification, according to Schimper’s application of the Natural Selection Theory, of the buoyant seeds and fruits of tropical littoral plants.

[Note 46]. On the modes of dispersal of the genus Brackenridgea.

[Note 47]. On the transport of gourds by currents.

[Note 48]. On the useless dispersal by currents of the fruits of the Oak and of other species of Quercus, as well as of the Hazel (Corylus).

[Note 49]. On the distribution of Ipomœa pes capræ, Convolvulus soldanella, and Convolvulus sepium.

[Note 50]. On the structure of the seeds and fruits of Barringtonia.

[Note 51]. On a common inland species of Scævola in Vanua Levu, Fiji.

[Note 52]. On the capacity for dispersal by currents of Colubrina oppositifolia.

[Note 53]. On the genus Erythrina.

[Note 54]. On the genus Canavalia.

[Note 55]. The inland extension of Scævola kœnigii.

[Note 56]. On the capacity for dispersal by currents of Sophora tomentosa, S. chrysophylla, and S. tetraptera.

[Note 57]. On the species of Ochrosia.

[Note 58]. On Pandanus.

[Note 59]. Seeds in petrels.

[Note 60]. Schimper on the halophilous character of littoral Leguminosæ and of shore plants generally.

[Note 61]. Meteorological observations on the summit of Mauna Loa.

[Note 62]. On the relative proportion of vascular cryptogams in Fiji.

[Note 63]. On the table of vascular cryptogams of Tahiti, Hawaii, and Fiji.

[Note 64]. On the distribution of the Tahitian ferns and lycopods.

[Note 65]. Distribution of some of the mountain ferns of Hawaii that are not found either in Fiji or in Tahiti.

[Note 66]. Endemic genera of ferns in Hawaii.

[Note 67]. On the dispersal of Compositæ by birds.

[Note 68]. On some of the Hawaiian endemic genera excluding those of the Compositæ and Lobeliaceæ.

[Note 69]. On the germination of Cuscuta.

[Note 70]. On beach-temperature.

[Note 71]. On the buoyancy of the seeds or seed-vessels of some Chilian shore plants.

[Note 72]. On the southern limit of the mangrove formation in Ecuador.

[Note 73]. Additional note on the temperature of the dry coast of Ecuador, between the island of Puna and the equator.

[Note 74]. Observations on the temperature of the Humboldt current from Antofagasta northward between January and March, 1904.

[Note 75]. On the stranded massive corals of the genus Porites (?) found on the coast of North Chile and Peru at Arica, Callao, and Ancon.

[Note 76]. Stranded pumice on English and Scandinavian beaches.

[Note 77]. On the mode of dispersal of Kleinhovia hospita.

[Note 78]. On the “Sea”, an unidentified wild fruit tree in Fiji.

[Note 79]. On willow-leaved river-side plants.

[Note 80]. Mr. Perkins on the Hawaiian Lobeliaceæ.

[Note 81]. On the vertical range of some of the most typical and most conspicuous of the plants in the forests on the Hamakua slopes of Mauna Kea, Hawaii.

[Note 82]. Aboriginal weeds.

[Note 90]. On the buoyancy of the seeds of Euphorbia amygdaloides and E. segetalis.

[Note 91]. Mr. E. Kay Robinson on Aster tripolium.

NOTE 1 (page [13])
On the Number of Known Species of Fijian Flowering Plants

Rather over 600 species of flowering plants are included in Seemann’s Flora Vitiensis, excluding the weeds and the plants introduced by man. Horne’s collections would probably add another 300 species; and many more remain to be discovered.

NOTE 2 (page [13])
The Littoral Plants of Fiji

In the following table are incorporated the results of an extensive series of observations and experiments on the buoyancy of the seeds and fruits of the shore plants made by the author during his sojourn of two years in Fiji, and based not only on prolonged buoyancy-tests, but also on systematic examination of the stranded and floating seed-drift, both of sea and river. The details would occupy many chapters: and it is only possible here to give the bare results. Since Professor Schimper went over much the same ground in the Malayan region, one enjoys in many cases the great advantage of his authority; but a fair proportion of the results are new; and, besides, there are a number of plants included, the buoyancy of whose seeds or fruits has long been well established. In all cases the seed or fruit is taken as it presents itself for dispersal by the currents. Many of the plants are discussed with some detail in various parts of this book, as indicated in the reference column of the table.

Since the Gramineæ and the Cyperaceæ contain very few species suited for direct transport by the currents over wide areas of sea, this list may be regarded as containing nearly all the littoral flowering plants possessing seeds or seed-vessels with any buoyancy of importance.

Nearly all the Tahitian strictly littoral plants are represented in Fiji, and the few that have not been found there yet, such as Sesbania grandiflora, Heliotropium anomalum, &c., may exist, as in the first-named species, in the neighbouring Tongan group, and may probably even exist in Fiji. Two other Tahitian littoral plants, that are widely spread in the Pacific, namely, Suriana maritima and Sesuvium Portulacastrum, are found in Tonga, and are included in my list of Fijian shore plants, though not yet recorded from that group, where, however, they will, without a doubt, be found by some future observer.

Table showing the Buoyancy of the Seeds or Fruits of the Littoral Plants of Fiji, excluding the Grasses and, with one exception, the Sedges

The letters placed before the plant name indicate that the species is also found in Hawaii (H), in Tahiti (T), and in the Marquesas (M). The Marquesan locality is only given where the plant is not in Tahiti.

The abbreviations in the reference column are as follows:

S=Schimper; G=Guppy; P=Earlier authorities and particularly the list given by Hemsley in the Introduction to the Botany of the Challenger Expedition.

Species.	Family.	Buoyancy of seeds or fruits.		Authorities.	Pages of further reference. See also Index.
		Float for months.	Sink at once or in a week or two.
HT Calophyllum inophyllum	Guttiferæ.	+	...	S.G.P.	18
HT Hibiscus tiliaceus	Malvaceæ.	+	...	S.G.P.	21
Hibiscus diversifolius (Jacq.)	Malvaceæ.	+	...	G.	21
HT Thespesia populnea	Malvaceæ.	+	...	S.G.P.	[Note 3]
H Gossypium tomentosum (Nutt.)	Malvaceæ.	...	+	G.
Heritiera littoralis	Sterculiaceæ.	+	...	S.G.P.	45, 48
T Kleinhovia hospita	Sterculiaceæ.	+	...	G.	21
T Triumfetta rhomboidea	Tiliaceæ.	...	+?
T Triumfetta procumbens	Tiliaceæ.	...	+	G.	45
T Suriana maritima	Simarubeæ.	+	...	S.G.
Carapa moluccensis	Meliaceæ.	+	...	S.G.P.	45
Carapa obovata	Meliaceæ.	+	...	S.G.P.	45
T Ximenia americana	Olacineæ.	+	...	S.G.	113
Smythea pacifica (Seem.)	Rhamneæ.	+	...	G.P.	106
HT Colubrina asiatica	Rhamneæ.	+	...	G.	137
HT Dodonæa viscosa	Sapindaceæ.	+	...	S.G.
HT Tephrosia piscatoria	Papilionaceæ.	...	+	G.	45
M Desmodium umbellatum	Papilionaceæ.	...	+	G.
HT Dioclea violacea	Papilionaceæ.	+	...	G.P.	82
T Canavalia obtusifolia	Papilionaceæ.	+	...	S.G.P.	[Note 54]
T Canavalia sericea	Papilionaceæ.	+	...	G.	[Note 54]
T Canavalia ensiformis, var. turgida.	Papilionaceæ.	+	...	S.G.P.?	[Note 54]
HT Mucuna gigantea	Papilionaceæ.	+	...	S.G.P.	81
T Erythrina indica	Papilionaceæ.	+	...	S.G.P.
HT Strongylodon lucidum	Papilionaceæ.	+	...	G.	82
HT Vigna lutea	Papilionaceæ.	+	...	S.G.	139
Dalbergia monosperma	Papilionaceæ.	+	...	S.G.	106
Derris uliginosa	Papilionaceæ.	+	...	S.G.P.	111
Pongamia glabra	Papilionaceæ.	+	...	S.G.P.
T Sophora tomentosa	Papilionaceæ.	+	...	S.G.	[Note 56]
T Inocarpus edulis	Papilionaceæ.	+?	...	G.P.
HT Cæsalpinia Bonducella	Cæsalpinieæ.	+	...	S.G.P.	193
T Cæsalpinia Bonduc	Cæsalpinieæ.	+	...	G.P.	193
Afzelia bijuga	Cæsalpinieæ.	+	...	G.	173
Cynometra sp.	Cæsalpinieæ.	+?	...	S.G.
Entada scandens	Mimoseæ.	+	...	G.P.	181
Acacia laurifolia	Mimoseæ.	...	+	G.	164
T Leucæna Forsteri	Mimoseæ.	...	+	G.
T Serianthes myriadenia	Mimoseæ.	...	+	G.	424
Parinarium laurinum	Rosaceæ.	+	...	G.P.
Eugenia Richii	Myrtaceæ.	...	+	G.
T Barringtonia speciosa	Myrtaceæ.	+	...	S.G.P.
Barringtonia racemosa	Myrtaceæ.	+	...	G.
Rhizophora mucronata	Rhizophoreæ.	+	...	S.G.P.
Rhizophora mangle	Rhizophoreæ.	+	...	S.G.P.
Bruguiera Rheedii	Rhizophoreæ.	+	...	G.P.	&nbsp
HT Terminalia Katappa	Combretaceæ.	+	...	S.G.P.
M Terminalia littoralis	Combretaceæ.	+	...	S.G.P.
Lumnitzera coccinea	Combretaceæ.	+	...	S.G.P.
T Gyrocarpus Jacquini	Combretaceæ.	+	...	G.	423
T Pemphis acidula	Lythraceæ.	+	...	S.G.
T Luffa insularum (Gray)	Cucurbitaceæ.	+	...	G.	426
HT Sesuvium Portulacastrum	Ficoideæ.	...	+	G.
HT Morinda citrifolia	Rubiaceæ.	+	...	S.G.P.
T Guettarda speciosa	Rubiaceæ.	+	...	S.G.P.
T Wedelia biflora	Compositæ.	+	...	G.
HT Scævola Koenigii	Goodeniaceæ.	+	...	S.G.P.

Table showing the Buoyancy of the Seeds or Fruits of the

Littoral Plants of Fiji, excluding the Grasses, and with

one exception, the Sedges (continued)

Species.	Family	Buoyancy of seeds or fruits.		Authorities.	Pages of further reference. See also Index.
		Float for months.	Sink at once or in a week or two.
T Cerbera Odollam	Apocynaceæ.	+	...	S.G.P.
T Ochrosia parviflora	Apocynaceæ.	+	...	G.P.
HT Cordia subcordata	Boraginaceæ.	+	...	S.G.P.
T Tournefortia argentea	Boraginaceæ.	+	...	S.G.P.
HT Ipomœa pes capræ	Convolvulaceæ.	+	...	S.G.P.
H Ipomœa glaberrima (Boj.)	Convolvulaceæ.	+	...	G.
Aniseia uniflora	Convolvulaceæ.	+	...	G.
T Premna tahitensis	Verbenaceæ.	+	...	G.	[Note 32]
Clerodendron inerme	Verbenaceæ.	+	...	S.G.
HM Vitex trifolia	Verbenaceæ.	+	...	G.
HT Cassytha filiformis	Lauraceæ.	+	...	G.
T Hernandia peltata	Lauraceæ.	+	...	S.G.
HT Wikstrœmia fœtida	Thymelæaceæ.	...	+	G.
Drymispermum Burnettianum	Thymelæaceæ.	...	+	G.
T Euphorbia Atoto	Euphorbiaceæ.	...	+	S.P.G.
Excæcaria Agallocha	Euphorbiaceæ.	+	...	S.G.
T Casuarina equisetifolia	Casuarineæ.	...	+	G.
HT Tacca pinnatifida	Taccaceæ.	+	...	G.	19
HT Cocos nucifera	Palmeæ.	+	...	P.
HT Pandanus odoratissimus	Pandaneæ.	+	...	S.G.P.
Crinum asiaticum	Amaryllideæ.	...	+	G.P.
Scirpodendron costatum	Cyperaceæ.	+	...	G.	407
Cycas circinalis	Cycadeæ.	+	...	G.P.

NOTE 3 (page [13])
Results of Long Flotation Experiments on the Seeds or Seed-vessels of Tropical Littoral Plants

At various times during the past twenty years I have made lengthened experiments in England on the buoyancy in sea-water of the seeds or seed-vessels of beach plants collected by me in the Solomon Islands, the Fijis, Hawaii, Keeling Atoll, &c. In all the species enumerated below, the floating powers were retained after twelve months’ immersion, the seed-contents being to all appearance unharmed. In six species I succeeded in getting the seeds to germinate after the experiment; and there can be no doubt that the number of successful results would have been largely increased, if I had not been obliged to resort to very primitive methods in conducting the experiments. Some of the results are referred to in a note to my paper on the flora of Keeling Atoll, dated about 1889; and if I remember aright, Mr. Hemsley mentioned those relating to Thespesia populnea and Ipomœa grandiflora in the Annals of Botany, not long after. The others have not been previously published. In one instance (Cæsalpinia bonducella) the flotation experiment was prolonged to two and a half years, the seeds floating buoyantly and being apparently quite sound at the end of the experiment.

As demonstrating that tropical seeds can be transported unharmed by currents through cold latitudes, it should be noted that all these experiments were conducted in England. In the cases of the Keeling Atoll seeds the experiment was carried on through a very severe winter, the vessel of sea-water being exposed to a degree of cold that kept fresh-water frozen for three weeks on the same table. This did not prevent the subsequent germination of the seeds of Thespesia populnea and Ipomœa grandiflora. The same thing was established in a more natural way by Lindman, who planted seeds of Entada scandens and Mucuna urens, that had been stranded on the Norwegian coast, and found that they retained their germinating capacity (see Sernander, p. 7).

The following are the seeds or seed-vessels that remained afloat after a year’s flotation in sea-water, those that subsequently germinated being preceded by G. In the other cases the germinating capacity was not tested; but they were always sound in appearance when cut across at the close of the experiment.

• G Thespesia populnea (Malvaceæ)
• Dioclea (violacea?) (Papilionaceæ)
• G Mucuna gigantea, D C (Papilionaceæ)
• G Mucuna urens, D C (Papilionaceæ)
• Mucuna, sp. (Papilionaceæ)
• Mucuna, sp. (Papilionaceæ)
• G Strongylodon lucidum, Seem. (Papilionaceæ)
• Sophora tomentosa, (Papilionaceæ)
• G Cæsalpinia bonducella (Cæsalpinieæ)
• Entada scandens (Mimoseæ)
• Morinda citrifolia (Rubiaceæ)
• Scævola Koenigii (Goodeniaceæ)
• Cordia subcordata (Boragineæ)
• Tournefortia argentea (Boragineæ)
• G Ipomœa grandiflora, Lam. (Convolvulaceæ)
• Tacca pinnatifida (Taccaceæ)

NOTE 4 (page [13])
Table illustrating the Degree of Buoyancy of the Seeds and Fruits of Inland Fijian Plants

(Unless otherwise indicated, the seeds or fruits sink at once or in a day or two)

• Abrus precatorius.
• Acacia Richii.
• Ageratum conyzoides.
• Alphitonia excelsa.
• Alpinia sp.
• Alyxia (scandens?).
• Artocarpus incisa.
• Artocarpus integrifolia.
• Barringtonia edulis (1 month)
• Barringtonia sp.
• Bauhinia sp.
• Bischoffia javanica.
• Cæsalpinia sp.
• Calophyllum spectabile (2-4 weeks).
• Calophyllum Burmanni (4-10 days).
• Cananga odorata.
• Canarium sp.
• Canarium sp.
• Canna indica.
• Citrus aurantium (3-4 weeks).
• Citrus decumana (1 month).
• Citrus limonum (5 weeks).
• Citrus vulgaris, R. (6-7 weeks).
• Coix lachryma (2-7 days).
• Commersonia platyphylla.
• Cordyline sepiaria.
• Couthovia corynocarpa (a few days).
• Cucumis acidus (a few days).
• Cucurbita sp. (several months).
• Cupania sp.
• Dammara vitiensis (7-10 days).
• Dioscorea sativa (a few days).
• Dioscorea sp.
• Dracontomelon sylvestre.
• Dracontomelon sp.
• Elæocarpus sp.
• Elæocarpus sp. (a few days).
• Eranthemum sp.
• Eugenia malaccensis (2-4 weeks).
• Eugenia effusa? (4-7 days).
• Eugenia confertiflora? (10-12 days).
• Eugenia rariflora (a few days).
• Eugenia corynocarpa (a few days).
• Eugenia rivularis (a week).
• Fagræa Berteriana (a few days).
• Ficus Harveyi (7-10 days).
• Ficus scabra (7-10 days).
• Ficus sp. (7-10 days).
• Gardenia vitiensis (4-5 weeks).
• Geissois ternata.
• Geophila reniformis.
• Gnetum gnemon.
• Grewia sp.
• Guettarda sp. (a few weeks).
• Hibiscus Abelmoschus (months).
• Hibiscus seculentus.
• Hydrocotyle asiatica (months).
• Ipomœa batatas.
• Ipomœa insularis (nil or months).
• Ipomœa peltata (weeks or months).
• Ipomœa turpethum (nil or weeks or months).
• Ipomœa sp. (7-10 days).
• Lindenia vitiensis (weeks or months).
• Maba sp. (7-10 days).
• Macaranga sp. (1-2 weeks).
• Melastoma denticulatum.
• Micromelum minutum.
• Momordica Charantia (a few days).
• Morinda Forsteri.
• Mussænda frondosa.
• Myristica sp. (3-7 days)
• Myristica sp. (3-7 days)
• Myrmecodia sp.
• Nelitris vitiensis (a few days).
• Nephelium pinnatum (a few days).
• Ophiorrhiza leptantha.
• Phyllanthus sp.
• Phyllanthus sp.
• Piper Macgillivrayi.
• Pittosporum sp.
• Pleiosmilax vitiensis.
• Portulaca (lutea?).
• Portulaca quadrifida.
• Premna serratifolia.
• Pritchardia pacifica.
• Psychotria sp.
• Psychotria sp.
• Psychotria sp.
• Psychotria sp.
• Psychotria sp.
• Ptychosperma sp.
• Rhaphidophora vitiensis.
• Sapota sp. (a few days)
• Sapota sp. (a few days)
• Scævola floribunda.
• Spondias dulcis (a month).
• Sterculia sp. (seeds nil, fruits months).
• Stylocoryne sambucina (2 or 3 days).
• Tabernæmontana (orientalis?) (a few days).
• Tacca maculata (nil or a few days).
• Trichospermum Richii (a few days).
• Urena lobata.
• Veitchia Joannis.
• Veitchia sp.

NOTE 5 (page [14]).
The Inland Fijian Plants possessing Buoyant Seeds or Fruits

They come under the following heads:

(a) Plants of the stream-border or the pond-side or of the inland swamp, e.g., Lindenia vitiensis and Hydrocotyle asiatica. The extension of the principle by which plants with buoyant seeds or fruits are located, not only at the sea-side but at the water-side generally, is here involved, as explained in [Chapter III.]

(b) Plants following the rule deduced by Schimper for Terminalia, that when a genus comprises several species possessing buoyant fruits, only those having fruits with the greatest floating power are found at the coast, the least buoyant plants occurring inland; examples, Calophyllum and Guettarda.

(c) Plants that like Ipomœa behave irregularly in respect to seed-buoyancy, a difference in behaviour often associated with varying stations both at the coast and inland.

(d) Plants with dehiscent buoyant capsular fruits, like Sterculia, where dehiscence takes place on the tree and the seeds have no buoyancy. Although the unopened fruit may float a long time, it does not in that condition come under the influence of the currents.

(e) Plants like Citrus Decumana, Gardenia, sp., &c., that, although apparently exceptions to the principle, do not offer much opposition to it, since the first is most at home at the river-side and the second often displays a decided inclination for a station at the coast.

(f) Genuine exceptions to the principle, such as Hibiscus Abelmoschus (see page [21]).

NOTE 6 (page [15])
Table showing the Degree of Buoyancy of the Seeds and Fruits of some Inland Hawaiian Plants

(Unless otherwise stated, the seeds or fruits sink at once or in a day or two)

• Acacia Koa.
• Aleurites moluccana (1-2 weeks).
• Alyxia olivæformis.
• Argemone mexicana.
• Argyreia tiliæfolia (nil or months).
• Bidens pilosa.
• Campylotheca sp.
• Canavalia galeata.
• Capparis sandwicensis.
• Cassia Gaudichaudii.
• Cassia occidentalis.
• Cheirodendron Gaudichaudii.
• Colubrina oppositifolia (weeks).
• Commelina nudiflora.
• Coprosma ernodeoides.
• Coprosma sp.
• Coprosma sp.
• Cyathodes Tameiameiæ (a few days).
• Cyrtandra sp. (a few days).
• Cyrtandra sp. (a few days).
• Cyrtandra sp. (a few days).
• Dianella odorata (a few days).
• Dracæna aurea.
• Eclipta alba (months).
• Erythrina monosperma.
• Gossypium tomentosum (a week).
• Gossypium barbadense (a few days).
• Gossypium sp. cultiv. (a few days).
• Hibiscus Youngianus (weeks).
• Hydrocotyle verticillata (weeks).
• Ipomœa bona nox (nil or months).
• Ipomœa insularis.
• Ipomœa pentaphylla.
• Ipomœa reptans.
• Ipomœa tuberculata.
• Jacquemontia sandwicensis.
• Jussiæa villosa (a few days).
• Lobeliaceæ (Clermontia).
• Maba sandwicensis.
• Metrosideros polymorpha.
• Mezoneuron kauaiense (pod, a week).
• Mucuna urens (months).
• Myoporum sandwicense.
• Olea sandwicensis, see page [364].
• Phyllostegia grandiflora.
• Phyllostegia mollis.
• Plectronia odorata.
• Pritchardia Gaudichaudii (5 or 6 weeks).
• Ricinus communis (7-10 days).
• Rubus Macraei.
• Scævola Chamissoniana.
• Scævola Gaudichaudii.
• Sida fallax.
• Sisyrinchium acre.
• Solanum aculeatissimum.
• Sophora chrysophylla (pod, 1-2 weeks).
• Viola Chamissoniana.
• Waltheria americana.

NOTE 7 (page [15])
Some Inland Hawaiian Plants possessing Buoyant Seeds or Fruits

Three of these, Eclipta alba, Hibiscus Youngianus, and Hydrocotyle verticillata, frequent wet places, and come under the principle that water-side plants generally have buoyant seeds or fruits. The buoyancy of the seeds of Argyreia tiliæfolia and of Ipomœa bona nox varies with station and may be explained as under Ipomœa in [Note 5]. The floating power of the fruits of Colubrina oppositifolia may be akin to that of inland species of Terminalia as indicated in [Note 5], since another species of the genus C. asiatica, which is a coast plant, has very buoyant seeds. Mucuna urens was no doubt originally, as it now is in tropical America, a littoral plant. The buoyant fruits of Pritchardia Gaudichaudii offer a genuine exception to the principle (see page [330]).

NOTE 8 (pages [18], [112])
The Pyrenes of Morinda

The pyrenes of the two Malayan inland species of Morinda (M. umbellata and M. longiflora) examined by Professor Schimper do not possess the bladder-like cavity to which those of M. citrifolia owe their floating power, and it is to be inferred from his remarks (p. 183) that they have little or no buoyancy. The pyrenes of a Fijian inland species, near M. Grayi, had no floating power as tested by me, and they lacked the bladder-like cavity.

NOTE 9 (page [18])
The Buoyancy of the Fruits of Calophyllum

Professor Schimper found that whilst the fruits of Calophyllum inophyllum, the shore tree, remained afloat after 126 days, those of C. amœnum, an inland species, sank in from three to fourteen days, both possessing similar buoyant structures, but to a less degree in the case of the inland species. This genus presents a parallel case to Terminalia referred to on page [17]; but the general discussion of the subject will be found in [Chapter XIII.] According to the above authority C. Calaba, a West Indian coast tree, has buoyant fruits. The same is also true of the fruits of a large inland tree in the Solomon Islands experimented on by me (Solomon Islands, p. 305). It would thus appear that the fruits of the genus are as a rule buoyant, and that, as in Terminalia, the least buoyant fruits belong to the inland species. Professor Schimper also shows (p. 182) that the diminished floating power of the fruits of the inland species is associated with diminution in thickness of the buoyant seed-shell which is most developed in the buoyant fruits of the strand species.

NOTE 10 (page [24])
The Buoyancy Experiments on British Plants

The experiments in all cases were made to test the floating power of the seed or fruit in the condition in which it is detached from the plant. It usually makes very little difference whether sea-water or fresh water is employed, since in my numerous experiments there were but few exceptions to the general rule that seeds or seed-vessels that sink in fresh water sink also in sea-water. This subject is discussed in [Chapter X.] However, it may be here observed that the chief effect of the increased density of sea-water is merely to increase the proportion of buoyant seeds or fruits in any particular species.

It is necessary in such experiments to imitate Nature as much as possible. The seed or fruit, as the case may be, must be experimented upon in the condition in which it falls from the plant, or in the condition in which it would be ultimately found in river and pond drift. The seed or fruit should be thoroughly wetted, and air-bubbles removed.

Prolonged drying has but a slight effect on the great majority of seeds and seed-vessels experimented on, and this is just as true of tropical plants. Those that sink at once in the mature and fresh condition rarely float more than a day or two even after drying for a year. The usual effect is to increase the floating capacity of seeds and fruits already buoyant, and not to develop the capacity.

The results given in the table refer only to sound seeds. In fresh-water experiments, in nearly all cases, the seeds ultimately germinate in the water, and this is the usual cause of the close of the experiment. In an ordinary collection of floating seed drift from a pond or river, germination will go on for years at each successive spring, the postponement of germination being a very striking feature with a fair proportion of seeds in river and pond-drift. This subject is dealt with in detail in my paper published in the Proceedings for 1897 of the Royal Physical Society of Edinburgh.

The Table of Results of Observations and Experiments on the

Buoyancy of the Seeds or Seed-vessels of more than 300

British Flowering Plants

Explanation of Table.—The capacity of floating for months is thus indicated, ++; of floating for 1 to 4 weeks, +; and where sinking occurs at once or within a week there is no entry. When buoyancy continued in my experiments after 6 and 12 months, it is indicated by Roman numerals (VI and XII). A=an aquatic plant; M=a beach plant; R=a river-side or pond-side plant; var.=variable in floating power.

R	Thalictrum flavum
A	Ranunculus aquatilis
R	Ranunculus hederaceus
R	Ranunculus flammula
	Ranunculus ficaria
R	Ranunculus sceleratus VI	++	Var.
	Ranunculus acris
	Ranunculus repens XII	++	Var.
R	Caltha palustris	+
	Berberis vulgaris
A	Nymphæa alba
A	Nuphar luteum
	Papaver rhœas
	Papaver dubium
	Chelidonium majus
	Rœmeria hybrida
M	Glaucium luteum
	Barbarea vulgaris
R	Nasturtium officinale
R	Nasturtium sylvestre
R	Nasturtium amphibium
	Arabis hirsuta
	Arabis thaliana
R	Cardamine pratensis
	Cardamine hirsuta
	Alliaria officinalis
	Brassica campestris
	Brassica alba
M	Cochlearia officinalis
M	Alyssum maritimum
	Draba verna
	Thlaspi arvense
	Capsella bursa pastoris
M	Cakile maritima	+
M	Crambe maritima	+
M	Raphanus maritimus	+
	Reseda luteola
	Helianthemum vulgare
R	Viola palustris
	Viola canina
	Viola tricolor
	Polygala vulgaris
	Silene cucubalus
M	Silene maritima
	Lychnis diurna
	Sagina procumbens
M	Arenaria peploides (Honckeneya) XII	++
	Mœnchia erecta
	Cerastium vulgatum
R	Stellaria aquatica
	Stellaria media
	Stellaria graminea
	Stellaria holostea
	Spergularia rubra
M	Spergularia marina
	Spergula arvensis
R	Montia fontana
R	Elatine hydropiper
	Hypericum perforatum
	Hypericum quadrangulum
R	Hypericum elodes
	Linum usitatissimum
	Linum angustifolium
	Malva rotundifolia
	Malva sylvestris
	Oxalis acetosella
	Oxalis corniculata
	Impatiens parviflora
R	Impatiens fulva VI	++
	Acer campestre
	Ilex aquifolium
	Euonymus europæus	+
	Ulex europæus
	Cytisus scoparius
	Ononis arvensis
nbsp;	Medicago lupulina
	Medicago denticulata
	Trifolium incarnatum
	Lotus corniculatus
	Anthyllis vulneraria
	Vicia sativa
	Lathyrus pratensis
M	Lathyrus maritimus	++
R	Spiræa ulmaria	+
	Fragaria vesca
	Potentilla tormentilla
	Potentilla sp.
R	Potentilla comarum XII	++
	Alchemilla arvensis
	Rosa arvensis
	Cratægus oxyacantha	+
R	Epilobium hirsutum
	Epilobium parviflorum
R	Lythrum salicaria
R	Peplis portula
	Bryonia dioica
	Cotyledon umbilicus
	Saxifraga granulata
	Saxifraga tridactylites
R	Chrysosplenium alternifolium
R	Chrysosplenium oppositifolium
	Drosera rotundifolia	+
A	Myriophyllum spicatum
A	Myriophyllum alternifolium
R	Hydrocotyle vulgaris XII	++
M	Eryngium maritimum	++
R	Cicuta virosa
R	Apium graveolens
R	Apium nodiflorum
R	Apium inundatum
R	Sium latifolium	++
R	Sium angustifolium	++
R	Œnanthe crocata VI	++
R	Œnanthe phellandrium	+
	Æthusa cynapium
M	Crithmum maritimum	++
R	Angelica sylvestris XII	++
R	Peucedanum palustre	++
	Pastinaca sativa
	Chærophyllum sylvestre
	Smyrnium olusatrum
	Hedera helix
	Viscum album
	Sambucus nigra
R	Galium palustre VI	++
	Galium mollugo
	Galium aparine
	Centranthus ruber
	Valerianella olitoria
	Eupatorium cannabinum
M	Aster tripolium
R	Bidens cernua VI	++
R	Bidens tripartita VI	++
	Chrysanthemum segetum
	Chrysanthemum leucanthemum
	Matricaria inodora
M	Matricaria inodora, var. maritima	++
	Matricaria chamomilla
	Achillea millefolium
	Artemisia vulgaris
	Artemisia absinthium
	Tussilago farfara
	Tussilago petasites
	Senecio vulgaris
R	Senecio aquaticus
R	Senecio palustris
	Carduus nutans
	Carduus lanceolatus
R	Carduus palustris
	Carduus arvensis
	Tragopogon pratensis
	Tragopogon porrifolius
	Tragopogon echioides
	Leontodon autumnalis
	Sonchus oleraceus
	Taraxacum dens leonis
	Crepis virens
	Crepis fœtida
	Lapsana communis
A	Lobelia Dortmanna
R	Hottonia palustris
R	Lysimachia vulgaris	+
R	Lysimachia thyrsiflora	+
M	Glaux maritima
	Anagallis arvensis
M	Samolus valerandi
	Pinguicula lusitanica
	Ligustrum vulgare
R	Menyanthes trifoliata	++
R	Limnanthemum nymphæoides	+
	Convolvulus arvensis
	Convolvulus sepium XII	++
M	Convolvulus soldanella XII	++
	Cuscuta europæa
	Lithospermum officinale
R	Myosotis palustris
	Myosotis arvensis
	Myosotis versicolor
	Lycopsis arvensis
	Symphytum officinale
	Borago officinalis
	Datura stramonium
	Solanum dulcamara
	Solanum nigrum
	Linaria vulgaris
	Linaria cymbalaria
	Scrophularia nodosa
R	Scrophularia aquatica
R	Veronica anagallis
R	Veronica beccabunga
	Veronica agrestis
	Veronica arvensis
	Bartsia odontites
	Rhinanthus crista galli VI	++	Var.
R	Pedicularis palustris	++
	Salvia verbenaca
R	Lycopus europæus XII	++
R	Mentha aquatica VI	++
	Thymus sp.
	Calamintha officinalis
	Nepeta glechoma
	Nepeta cataria
	Prunella vulgaris
R	Scutellaria galericulata XII	++
	Stachys betonica
	Stachys sylvatica
R	Stachys palustris	++
	Stachys arvensis
	Galeopsis tetrahit
	Ballota nigra
	Lamium purpureum
	Lamium album
	Lamium galeobdolon
	Teucrium scorodonia
	Ajuga reptans
	Verbena officinalis
M	Armeria vulgaris
	Plantago major
	Plantago media
	Plantago lanceolata
M	Plantago maritima
M	Salicornia herbacea
M	Salsola kali	+
M	Suæda fruticosa
M	Suæda maritima
	Chenopodium album
M	Beta maritima
	Atriplex patula VI	++
R	Rumex aquaticus	+
	Rumex crispus	++
	Rumex obtusifolius
R	Rumex hydrolapathum XII	++
R	Rumex conglomeratus XII	++
	Polygonum aviculare
M	Polygonum maritimum
	Polygonum convolvulus
R	Polygonum amphibium
R	Polygonum persicaria
	Polygonum lapathifolium
R	Polygonum hydropiper
	Euphorbia helioscopia
	Euphorbia segetalis
M	Euphorbia paralias	++
	Euphorbia peplus (?)
	Euphorbia amygdaloides
A	Ceratophyllum demersum
A	Callitriche aquatica
	Urtica dioica
	Ulmus campestris
R	Alnus glutinosa XII	++
	Betula alba
	Corylus avellana	+
	Quercus robur	+
	Pinus sylvestris	+
	Taxus baccata
R	Typha latifolia
R	Typha angustifolia
R	Sparganium ramosum XII	++
R	Sparganium simplex VI	++	Var.
R	Sparganium minimum	++
	Arum maculatum
R	Calla palustris XII	++
A	Lemna minor	++
A	Lemna gibba
A	Naias marina
A	Zannichellia palustris
A	Ruppia maritima
A	Potamogeton natans XII	++
A	Potamogeton oblongus VI	++
A	Potamogeton lucens VI	++
A	Potamogeton perfoliatus	++
A	Potamogeton crispus
A	Potamogeton densus
A	Potamogeton obtusifolius
A	Potamogeton pusillus
R	Butomus umbellatus
R	Sagittaria sagittifolia VI	++	Var.
R	Alisma plantago VI	++	Var.
R	Alisma ranunculoides
A	Alisma natans
R	Damasonium stellatum
R	Scheuchzeria palustris	++
R	Triglochin palustre
R	Triglochin maritimum
A	Hydrocharis morsus ranæ
R	Iris pseudacorus XII	++
	Iris fœtidissima
	Tamus communis
	Fritillaria meleagris	++
	Scilla nutans
	Narthecium ossifragum
R	Juncus communis
	Juncus glaucus
R	Juncus articulatus
	Juncus squarrosus
R	Juncus bufonius
M	Juncus maritimus
	Luzula campestris
R	Cladium mariscus	++
R	Blysmus rufus	++
R	Scirpus palustris
A	Scirpus fluitans
	Scirpus setaceus
	Scirpus holoschœnus
R	Scirpus lacustris
M	Scirpus maritimus	+
	Scirpus sylvaticus
	Eriophorum alpinum
	Eriophorum vaginatum
	Eriophorum polystachion
R	Carex leporina	++
	Carex stellulata XII	++
R	Carex canescens	++
	Carex remota XII	++
R	Carex paniculata XII	++
R	Carex vulpina XII	++
R	Carex acuta VI	++
	Carex hirta
R	Carex flava VI	++	Var.
	Carex distans
	Carex panicea
R	Carex pseudocyperus VI	++
R	Carex ampullacea	+
R	Carex paludosa XII		++
	Carex sp.
	Carex sp.
R	Leersia oryzoides
R	Alopecurus geniculatus
	Agrostis sp.
	Agrostis sp.
R	Poa aquatica
R	Poa fluitans
	Melica nutans
R	Arundo phragmites	+?

Total of the original list: 320 species belonging to 192 genera and 65 families. Of these, about 260 were tested by the author, the data for the remaining species being mainly derived from the writings of Thuret, Kolpin-Ravn, and Sernander, with a few from those of Darwin and Martins.

Note.—Whilst this work has been going through the press, the author has added thirteen species, seven genera, and two families to the list above given; but the general inferences are not affected by the additions. The corrected total would, therefore, be 333 species, 199 genera, and 67 families.

On the effect of drying on the buoyancy of seeds and seed-vessels

It has been already observed that this is as a rule but slight, and that in the great majority of cases the effect of prolonged drying for many months, or even for years, is at the most to give a seed or fruit originally non-buoyant a floating power of a few days’ duration. This is a subject to which I have paid especial attention in my experiments, since, of course, much depends on it in the way of dispersal by currents. It is obvious that a seed or fruit possessing impermeable coverings at the time of its separation from the parent can scarcely be compared with one where the coverings only attain their water-proof capacity by drying. Most gardeners know that seeds which dry easily take up moisture easily, and the principle applies in a varying degree to the great majority of seeds and fruits.

Darwin was inclined to attach importance to adventitious buoyancy acquired by drying; and in the Origin of Species he refers to instances offered by the fruits of the Hazel (Corylus), the Asparagus, and Heliosciadium. In [Note 48] I have referred to the cases of the Oak and the Hazel; and, indeed, we have only to examine the beach-drift in various parts of the world, and to look at their respective stations, to learn that this is not an effective mode of dispersal. Buoyancy of seed or fruit is only one of many other qualities that is concerned with distribution by currents. Nature does not act in this way in seed-distribution, and there can be little doubt that the author of the Origin of Species would have been the first to abandon this view, if his researches had been continued. It should be especially noted that plants of the sea-beach, where the floating power happens to be nil, or limited only to a week or two, would have derived great advantage from the drying of their seeds or fruits if it was really effective in aiding dispersal by currents. However, with plants like Cakile maritima, Eryngium maritimum, Glaucium luteum, &c., the effect of drying is very small.

NOTE 11 (page [25])
The Effect of Sea-water Immersion on the Germinating Capacity of Seeds and Seed-vessels

Berkeley, Darwin, Martins, and others, long ago established the capacity of seeds to germinate after prolonged immersion in sea-water. The reader will find a resumé of their results in the appendix to Mr Hemsley’s volume on the Botany of the Challenger Expedition. The subject is well illustrated in the original papers of those authors, and in my later papers on the flora of Keeling Atoll, and on the seed-drift of the Thames.

I may here remark that the earlier observers often pay more attention to the retention of the germinating capacity after sea-water immersion than to the degree of buoyancy. For this reason I have not been able to make great use of the buoyancy results of Martins, since he frequently does not distinguish between temporary and long-sustained buoyancy, an objection also pointed out by Thuret and Hemsley.

NOTE 12 (page [27])
The Buoyancy of the Fruits of Galium aparine

Norman and Sernander (see p. 172) attribute considerable buoyancy to these fruits on account of the hollow cavity in each. I used to find them in England in floating river-drift in autumn; and Norman observed them on the Scandinavian beaches. They do not, however, float long, as the cavity is open; and in two sets of my experiments they sank within a few days.

NOTE 13 (page [29])
The Buoyancy of the Seeds of Convolvulus sepium

This plant seeded freely in 1893 in the Lower Thames Valley, as at Molesey. I kept some of the seeds afloat for thirty-three months, of which the first nine months were spent in sea-water and the rest in fresh-water. One seed, at the end of the period, germinated healthily in the fresh-water.

NOTE 14 (page [26])
Other Long Flotation Experiments

Whilst keeping my collections of Thames seed-drift in water from year to year, I obtained a number of records of long “flotations.” Thus in several cases, as with Bidens cernua and different species of Carex, germination of the floating fruit took place in the water after a period of two years. The same is also true of the seeds of Iris pseudacorus and of the drupes of Sparganium ramosum. The last-named remained afloat in the vessels, with the seed still sound, after four years; and the fruits of Carex paludosa germinated afloat after three years in water. Many drift fruits and seeds did not germinate freely in the vessels until the second spring, that is, after a lapse of eighteen months; and in those cases where the experiments were still further prolonged, a few germinated in the vessels in the third and sometimes even in the fourth year.

NOTE 15 (pages [33], [280])
The Occurrence Inland of Silene maritima

Prof. Schimper appeared to be in doubt as to the inclusion of this littoral plant amongst those found in elevated mountain districts. However, an interesting note on the occurrence of this plant on the summit of one of the inland Norwegian mountains is given by Sernander (p. 405), and is referred to by me on page [280] of this work.

NOTE 16 (page [34])
The Buoyancy of the Seeds or Fruits of the British Beach-plants that also occur Inland

My experiments in the case of Armeria vulgaris, Artemisia, Cochlearia officinalis, Plantago, the maritime forms of Spergularia rubra with and without winged seeds, and Silene maritima disclose little or no floating capacity even after prolonged drying. Thuret obtained similar results for the Spergularia. It is unlikely that other plants of the group possess any floating power worth speaking of. As indicated in [Note 71], the fruits of Raphanus maritimus float only for 7 to 10 days.

Nature disperses the fruits of Armeria vulgaris inclosed in the persistent calyx; but in this condition they float only for 2 to 4 days in sea-water, and the buoyancy of the capsule and seed is still more limited. They are sufficiently light to be blown some distance by strong winds, and the stiff hairs would cause them to adhere to a bird’s plumage in the case of gulls nesting where the plants grow.

Reference to Matricaria inodora is made under [Note 18].

NOTE 17 (page [35])
The Buoyancy of the Seeds or Fruits of the Group of British Littoral Plants that frequent Salt Marshes and Muddy Shores

Aster tripolium. The achenes, with or without the pappus, sink in fresh and salt water in a day or two even after a year’s drying.

• Glaux maritima
• Plantago maritima
• Samolus valerandi
• Suæda fruticosa
• Suæda maritima

The small seeds, or the seed-like nucules as in Suæda, have but little floating power even after prolonged drying.

Salicornia herbacea. Would be dispersed probably by floating portions of the plant, which, however, soon break down and the liberated seeds sink. The floating seedling thrives in sea-water and could be carried great distances (see [Note 19]).

Salsola kali. I experimented on this plant, both on the coast of Devonshire and in Chile, with the same results in both localities whether in the fresh state or after drying for weeks. The fruit sinks, but when the plant dries the fruit is often detached inclosed in the perianth, and floats in that condition in sea-water for a few days. Portions of the plant of various sizes bearing mature fruits all sank within ten days. It would seem at first sight, from the observations of Prof. Martins, that the fruits float for several weeks; but his experiments were mainly directed to testing the powers of germination after sea-water immersion; and it is often not at all clear whether flotation is implied or even to be correctly inferred. There is a slight suspicion of germination on the plant. Sea-birds doubtless aid in the dispersion of this plant; the dry crisp portions of the plant carrying fruits catch readily in one’s clothes on account of the prickly-pointed leaves.

Scirpus maritimus. The fresh fruits float a few weeks in sea-water in most cases, but 10 per cent. remain afloat after two months. After drying for some months 30 per cent. remain floating after two months’ immersion.

• Triglochin maritimum
• Triglochin palustre

The fruits float a few days or a week. Drying somewhat increases the buoyancy. Sir W. Buller in New Zealand found in the gullet of Anas superciliosa, the Grey Duck, numbers of the fruits of Triglochin triandrum.

NOTE 18 (page [35])
The Buoyancy of the Seeds or Fruits of the British Littoral Plants that are confined to the Beach

Arenaria (Honckeneya) peploides. The seeds float for many months in sea-water unharmed, 75 per cent. floating after a year. They never germinate in sea-water; but on being transferred to fresh water after many months in sea-water they germinate healthily in a few days. These seeds only float a few days in fresh water, all sinking within 10 days, and even after a year’s drying they sink in a week or two. Precisely the same results were produced in my experiments in 1892 on Cornish seeds, and in 1904 on Devonshire seeds. In the great contrast between their floating capacity in sea-water and in fresh water the seeds of this plant defy the general rule that seeds that float a long time in sea-water float also a long time in fresh-water. According also to Sernander the seeds float a long time in the sea. He says that the capsules float, but since they ultimately dehisce this could scarcely be efficacious in dispersal. Floating portions of the plant also aid in its dispersal, according to the same authority (p. 174). The plant forms great extended masses on the pebbly shores of Spitzbergen (Ekstam, p. 28).

Beta maritima. Thuret found that the dried fruits of this plant floated only two or three days in sea-water; whilst in my sea-water experiments the freshly gathered fruits floated only one or two days. Sernander speaks of them as fitted for dispersal from shore to shore; but this could only be to a limited extent. Martins and Thuret established by experiment the capacity of the germination of seeds of other species of Beta after long immersion in sea-water; and the first seems to imply that those of Beta vulgaris float for many weeks; but I am inclined to think an error lies here.

Cakile maritima. The fruits, even after long drying, float, as a rule, only a week and sink within ten days, the same results being afforded in my sea-water experiments in 1893 on fruits from Cornwall, and in 1904 on fruits from Devonshire. The fruits are common in the stranded drift on the north coast of Devonshire and may often be seen germinating there. They are also frequent in the beach drift of the Scandinavian coasts (Sernander, p. 156).

Crambe maritima. The fruits were kept floating by Sernander more than 13 days (p. 165). Martins implies that they floated for 45 days. Darwin says that they germinated after 37 days’ immersion in sea-water, but does not specify that they floated all the time.

Crithmum maritimum. The ripe fruits readily separate into the two carpels, which are very buoyant and float in sea-water for months. In my experiments, 95 per cent. remained afloat after 10 months. It is remarkable that whilst in sea-water the spongy covering of the carpels retains its vitality, in fresh-water it becomes sickly and decays and the carpels lose their floating power, so that they float weeks instead of months as in the sea-water. The carpels are extremely light, being washed up in the spray and blown up by the wind amongst the lightest of the stranded drift of the Devonshire beaches. In a moderate gale they are often blown off the beach and up the cliff-faces.

Convolvulus soldanella. From 40 to 50 per cent. of the seeds float after six months in sea-water, and about 30 per cent. float after eighteen months, retaining up to the end their germinating capacity. Sernander implies that the plant is not found on the Scandinavian coast to the north of Nissum Fjord in Denmark. It is known, however, to occur in the south of Scotland. (I am indebted to Mr. Millett for his extremely kind assistance in experimenting on this plant about ten years since.)

Eryngium maritimum. The fruits float in sea-water, as a rule, only 3 or 4 days and all sink within a week. After drying for three months, the floating period is only increased by a day or two. Though not at all suited for transport for any distance by the currents, the carpels, on account of their long prickly calyx teeth, would readily become entangled in a bird’s plumage, and doubtless they are dispersed usually in that fashion.

Euphorbia paralias. The seeds float a long time unharmed in the sea. In my experiments at least 90 per cent. remained afloat after six weeks in sea-water. On account of their small size they are liable to be overlooked in beach drift; but they are to be found stranded on the sands of our southern coasts, and they came under my notice in abundance in the seed-drift of the Sicilian beaches.

Glaucium luteum.—The seeds have no proper buoyancy even after prolonged drying. On account of their oiliness they will float at first on still water; but they can be made to sink at once or in a day by dropping water upon them. The mode of dispersal is problematical.

Lathyrus maritimus.—The seeds are evidently able to float a long time. They were, according to Sernander (p. 178), found in quantities by J. Schmidt cast up on some sand-islets near Falster in Denmark; and the plant is regarded by Norman as distributed over the coasts of Arctic Norway through the agency of the currents. They have, as observed by Schmidt, considerable floating powers. Some small leguminous seeds, seemingly of this species, which I found in the beach drift of Woollacombe Sands, Devonshire, floated uninjured for many weeks in sea-water.

Matricaria maritima, maritime variety of M. inodora. The fruits floated in my experiments unharmed after eight months in sea-water. In an experiment made some years since on the fruits of the inland form I noted that they had little or no buoyancy; but it is necessary to repeat the observation. Sernander (p. 181) supports Norman’s view that these plants are spread by the currents in Arctic Norway. The fruits occur in the Baltic sea-drift and also in fresh-water drift. M. inodora is found on sandy beaches in Nova Zembla. I am inclined to regard the maritime form from the dispersal standpoint as a distinct species.

Polygonum maritimum.—I have made observations on this plant in Devonshire, the Lipari Islands, and the coast of Chile. As in the case of several other species of Polygonum tested by me the fruits have little or no buoyancy, but inclosed in the perianth they float three or four days. The entire plant floats; but portions placed in sea-water sank within five or six days. Shore-birds can alone explain the wide distribution of this species.

The structural characters of some of these fruits or seeds are in their relation to buoyancy discussed on page [115]. It may be here observed that the valuable results obtained by Prof. Martins in testing the germinating capacity of the fruits and seeds of several of the shore-plants above mentioned, after long immersion in sea-water, are at times not to be depended on for the flotation indications, the persistence of the seed’s vitality being the special purpose of his research. His negative results as regards germination are not, however, always conclusive, since the period employed from April to June was quite insufficient. In many of my experiments seeds after long flotation in sea-water did not germinate for a year or more afterwards. If his investigation had been extended, the opinion that the Ranunculaceæ, the Malvaceæ, and the Convolvulaceæ are apparently least able to resist the action of sea-water would never have been formed. A very large amount of evidence now shows that most seeds or fruits that are at all well protected will germinate after long immersion in sea-water. But all experiments must be well safeguarded and extended over a year or two. The necessity of this was long since shown by Thuret. By employing double sets of seeds he ascertained that in a third of the species germination failed not only in the case of the seeds immersed in sea-water, but also in those that had not been placed in sea-water at all. Future investigators may, however, regard the buoyant qualities of seeds or fruits with their associated structural characters as offering now the true line of research. Observers beginning with Berkeley and Darwin down to the present time have quite established the fact that seeds as a rule germinate freely after long sea-water immersion.

NOTE 19 (page [35])
On Germination in Sea-water

During my experiments on the buoyancy of about 270 British plants, about a fourth of them (including most of those with buoyant seeds or fruits) were subjected to prolonged immersion in sea-water from periods varying from six to thirty-three months. If we except plants like Aster tripolium, Salicornia herbacea, Triglochin maritimum, &c., that live normally in salt marshes, or on the muddy banks of estuaries, only one of the whole number, namely, Ranunculus sceleratus, displayed the capacity of germination in sea-water. Amongst the plants that failed may be mentioned the following that are confined to the sea-beach—Arenaria peploides, Cakile maritima, Convolvulus soldanella, Eryngium maritimum, Euphorbia paralias, Glaucium luteum, and we may here include Crithmum maritimum of the rocky coasts. Of the beach-plants that also grow inland, Silene maritima and Spergularia rubra (excepting the form found on muddy coast flats) likewise failed. Amongst the plants of miscellaneous inland stations that failed were Atriplex patula, Bidens cernua, B. tripartita, Calla palustris, several species of Carex both from dry and wet situations, Convolvulus arvensis, C. sepium, Hydrocotyle vulgaris, Iris pseudacorus, several species of Juncus, Lycopus europæus, Mentha aquatica, Ranunculus repens, Rhinanthus crista galli, several species of Rumex, Scutellaria galericulata, Sparganium ramosum, &c.

In nearly all the plants that failed to germinate in sea-water the capacity of readily germinating in fresh water was displayed. The restraining power of immersion in sea-water was illustrated over and over again in my experiments. During the course of an experiment seeds removed from the sea-water vessel and placed directly in a vessel of fresh water kept beside the other germinated in a few days, whilst those left in the sea-water never germinated, though often kept there for months after. It was also noticeable that a previous sea-water immersion favoured early germination in fresh water. It may be added that most of the experiments were on floating seeds and seedvessels, though germination also occurred in the sunken state.

It was ascertained in the exceptional case of Ranunculus sceleratus, that although germination took place in sea-water, it was only after a prolonged soaking of months had prepared the way. Of a number of its seed-like fruits placed in fresh water and in sea-water in April and kept under the same conditions, those in fresh water germinated freely in a week or two, whilst those in sea-water did not begin to germinate until the following October. Whilst the floating seedlings produced by germination in fresh water grew vigorously and developed roots, those resulting from germination in sea-water and left in the vessel only attained a length of four millimetres in two months, developed no roots, and showed only the first leaf. The sea-water seedlings were pale green, and in their stout fleshy appearance contrasted greatly with the slender fresh-water seedlings.

With regard to the germination in sea-water of the plants of the salt marsh and of the mud-flats of estuaries, the following observations may be made. With Aster tripolium the seeds germinate readily in sea-water even when its density is raised by evaporation to 1·040; and I think that by a carefully graduated series of experiments they could be induced to germinate in brine. The seeds of Salicornia herbacea germinate in sea-water more readily than in fresh water; and the sea-water seedling is much the more vigorous and healthy of the two. I kept the floating seedlings in sea-water for about ten weeks from the date of germination, when they had developed the second joint and were throwing out rootlets. After that, unless placed in salt-mud, they became sickly and died. The floating seedling can evidently disperse the species. I found with Spergularia marina, the maritime form of S. rubra, that seeds of the plants growing on a sandy beach did not germinate in sea-water, only those from plants growing on muddy coast-flats doing so. But the sea-water seedlings, unlike those of Salicornia herbacea, but like those of Ranunculus sceleratus, when left in sea-water did not thrive. The seeds of Triglochin maritimum, as well as those of T. palustre, behave very similarly in sea-water, germinating readily, the liberated seedlings thriving afloat and producing the plumule. The ultimate test of the capacity for germinating in sea-water seems to lie in the behaviour of the seedling when left in the sea-water. Unless it belongs to a characteristic plant of the salt marsh or of the estuary, like Salicornia, it makes but little attempt at growth whilst afloat in sea-water, showing no rootlets, though at times developing the plumule.

The germination of seeds in sea-water also attracted the notice of Darwin; but his results in some respects are scarcely those I should have looked for (Gardener’s Chronicle, May, 1855, and Journ. Linn. Soc., vol. i., p. 130, 1857). Out of the seeds of 87 plants placed in sea-water to test their capacity of germination when afterwards planted, in three cases, those of Tussilago farfara, Convolvulus tricolor, and the garden Orache (Atriplex), the seeds germinated under the water, the freed seedlings, as with the two first named plants, living in the sea-water for some time after. Darwin was evidently himself surprised at these results, and I am quite unable to understand them. In England and in the tropics I have carried on prolonged sea-water experiments on the seeds of at least fifteen species of Convolvulus and Ipomœa (including the beach plants C. soldanella and I. pes capræ) and have never obtained such a result. The seeds will nearly always germinate well in fresh water; but in sea-water the process begins, as indicated by the swollen seed, and then aborts, the embryo dying (see page [83]). The seeds of Atriplex patula, though a long time in sea-water in my experiments, made no attempt to germinate there. Neither Prof. Martins, who experimented upon the effects of sea-water immersion on the seeds of nearly 100 plants, including many coast species, nor M. Thuret, who experimented in sea-water on the seeds of 251 plants, the experiments being in some cases prolonged for more than a year, make any reference, as far as I could gather from their writings, to any cases of germination in sea-water. Darwin’s results, however, are always significant in matters of dispersal; and perhaps one of my readers will be able to experiment again on his three plants.

When in Hawaii, I made some observations on the germination of Batis maritima in sea-water, a plan with which I was also familiar in its home in the salt-water pools of the coast of Peru. The mature fruits, on being freed from the parent plant in sea-water, float away, and in from one to two weeks they break down from decay, setting free the seeds. The seeds float in sea-water indefinitely, their buoyancy only terminating with their germination, the first seeds germinating afloat about six weeks after the breaking down of the fruit, whilst the rest continue to float in the sea-water during the next three months, some of them germinating at intervals, and all of them doing so eventually. Strange to say, although the seedlings remained healthy whilst afloat in the sea-water, they made no effort either to separate the cotyledons or to produce a plumule.

NOTE 20 (page [42]).
On the Maximum Heights reached by some Shore Plants in their Extension Inland in Vanua Levu, Fiji

Since they occupy the “talasinga” districts described in the following note, these shore plants would be expected to extend as high as those districts extend, namely, to about 1,500 feet above the sea. This indeed represents their limit excepting in one instance; but many fall considerably short of this elevation.

• Canavalia obtusifolia, variety, 700 feet, rare.
• Cassytha filiformis, 950 feet.
• Cerbera Odollam, 1,200 feet: 2,600 feet in one exceptional case on the slopes of Mbatini.
• Colubrina asiatica, 400 feet.
• Cycas circinalis, 1,100 feet.
• Derris uliginosa, 1,000 feet, rare.
• Ipomœa pes capræ, 1,300 feet.
• Morinda citrifolia, 700 feet.
• Scævola Kœnigii, not common inland, and rarely over 100 feet above the sea; but it may occur miles from the beach, as near Vatu Levoni, where a few stunted plants were growing five miles from the coast.
• Vitex trifolia, 1,300 feet, usually more or less unifoliolate and procumbent.

Unless otherwise stated all the plants above named are common inland, as also are Premna tahitensis, Tacca pinnatifida, Tephrosia piscatoria, Hibiscus tiliaceus, &c.; but I have made no note of Thespesia populnea occurring far off the beach.

NOTE 21 (pages [42], [43])
On the Dwarfing of Shore Plants when extending Inland into the “Talasinga” Plains in Vanua Levu.

Premna tahitensis, 9 or 10 feet high at the coast, may here be only 3 feet high. Other trees like Morinda citrifolia become also stunted. Cerbera Odollam, a moderate-sized tree at the coast, may in the “talasinga” plains be only 4 to 6 feet high, but it here displays distinct varietal characters. Whilst the shore trees of Cerbera Odollam have broad leaves (length 3 times the breadth) with obtuse points, and short, stout flower-peduncles (1¹⁄₂-2 inches), the inland or “talasinga” species has long lanceolate leaves (length 7 or 8 times the breadth), and long, slender flower peduncles (3 inches). However, intermediate forms are common, the broad-leaved coast tree approaching the inland plant and vice versâ.

NOTE 22 (page [43])
The “Talasinga” Plains of Vanua Levu, Fiji

Amongst the most conspicuous features of the north and north-west or lee sides of the large islands of Vanua Levu and Viti Levu are the extensive rolling plains that extend from the sea-border for some miles inland to the foot of the mountains. It is to those of the first-named island that the following remarks strictly apply; but no doubt they will serve equally well for those of the other island. In the first volume on the geology of Vanua Levu, reference is frequently made to this subject, and the reader may profitably look at the remarks there made.

Here the mountain-forests more or less abruptly cease, and we have an undulating region of grass, reeds, and ferns dotted over with Casuarinas, Pandanus trees, Cycads, Acacias, and shrubby growths. Though the list of plants characteristic of these plains is not small, they are not, as a rule, numerous in any one locality, and the general appearance is one of aridity. A dry, crumbling soil, often deeply stained by iron-oxide, is plentifully exposed; and blocks of basic volcanic rocks in all stages of disintegration are strewn over the surface in many localities. Rivers, fed by the heavy rainfall of the forested slopes of the mountains, traverse these regions, but, as a rule, receive no tributaries; and the districts have, in fact, well earned the name given to them by the natives of the “talasinga,” or sun-burnt, lands.

The vegetation, though sparse and scanty in comparison with that of the forests, is sufficiently varied when it comes to be more closely examined. In one locality we may have extensive tracts covered with Gleichenia, Pteris, and other ferns of the bracken habit. In another, tall reeds (Eulalia) and grasses cover large areas. Here, more than one species of Tacca (T. pinnatifida and T. maculata) thrive. There, the Turmeric (Curcuma longa) abounds. Trailing over the soil in one place we notice Ipomœa pes capræ, in another the Yaka (Pachyrrhizus trilobus), and in another the procumbent unifoliolate form of Vitex trifolia. Amongst the shrubs and small trees we observe in different localities the Sama (Commersonia echinata), the Mbulei (Alstonia plumosa—one of the rubber plants), Mussænda frondosa, Melastoma denticulatum, and Nelitris vitiensis, the Nunga-nunga. Dodonæa viscosa, found in similar regions in Australia and New Zealand, abounds in places; and here and there may be seen species of Hibbertia, another Australian genus. Fagræa Berteriana, the Mbua tree, grows abundantly in certain districts, as in the Mbua plains, and Gardenias are at times abundant. One or two characteristic beach-plants have been already mentioned, and amongst others particularly frequent in these plains are Cassytha filiformis, Cerbera Odollam, Morinda citrifolia, and Premna tahitensis.

When these talasinga districts approach the forests, patches of wood occur at intervals, and we observe here the Candle-nut Tree (Aleurites moluccana), the Vunga (Metrosideros polymorpha), and the Thau-kuro (Casuarina nodiflora). Such are some of the botanical features of these districts; but the reader will acquire a sufficiently correct general notion of the floral physiognomy of these regions if he bears in mind their most conspicuous characters, those of an undulating region more or less covered with ferns, tall reeds, and grass, and dotted over, either separately or in clumps, with Casuarinas (C. equisetifolia), Screw-pines (Pandanus odoratissimus), Cycads (C. circinalis), and Acacias (A. Richii, &c.).

However, the peculiar vegetation of these plains often ascends the lower slopes of the mountains, reaching to various elevations. In Vanua Levu it often ceases at 900 or 1,000 feet, but it may only reach to 400 or 500 feet, and, on the other hand, not uncommonly it ascends to as much as 1,500 feet, the greatest elevation recorded by me being 1,600-1,700 feet in the Sealevu district. It extends miles inland, and where conditions are suitable it may reach the heart of the island.

Different explanations have been offered of the origin of the peculiar vegetation of the leeward slopes of these islands. It is, however, a phenomenon that is presented over much of the globe by islands lying in the track of regular winds, the weather, or wet, side being densely wooded, whilst the lee, or dry, side is covered with grass, ferns, and similar vegetation. The predisposing cause must be climatic; and although Mr. Horne’s explanation attributing it to the effect of fires and to a faulty system of native cultivation (pp. 80, 132) may be doubtless true in certain localities, the influences at work here must be the same as are at work in other islands and on continental coasts in other parts of the world.

But for all that it is not easy to give a definite explanation even from a meteorological standpoint. Those who are interested in this subject will recall the desert districts of Australia and the dreary sandy wastes of the coast of Northern Chile and Peru; and they will be cautious in venturing on a definite explanation even with such relatively unimportant examples of the same principle as are exhibited by the islands of Fiji. Dr. Seemann, writing of these “talasinga” plains (p. xii), remarks that “their very aspect is proof that rain falls in only limited quantity,” the mountainous backbone of the islands intercepting, as he holds, much of the rainfall. But the subsequent observations of Mr. Holmes, at Delanasau, in the “talasinga” district on the north-west side of Vanua Levu, have shown that there is by no means a small rainfall in this locality, the average rainfall, for instance, for the seven years ending December, 1877, being 113 inches, which must be quite two-thirds or three-fourths of the fall on the weather side of the island (see p. 215); whilst the average number of days on which rain fell was 156. The true cause would seem to lie in the excessive dryness of the air on the lee side of the islands between the rains, and the whole matter may, perhaps, be one rather for the hygrometer than for the rain-gauge. I have no comparative data bearing on this point; but Mr. Holmes, whose observations as here quoted are from Horne’s Year in Fiji, found that the mean relative humidity for 1875 at 1 P.M. was 63, which is certainly very low for the tropics. I may remark that, as far as personal experience goes, the climate on the lee side of Vanua Levu is much more enervating, much less healthy, and the air is far more “drying” than on the side exposed to the trade-wind.

Geological characters, as I found, explained nothing in this connection, the “talasinga” vegetation sometimes occurring on basaltic areas, at other times on the “soapstone” or calcareous mud-stone, and again on coarser tufaceous rocks. In my volume on the geology of Vanua Levu (p. 57), it is pointed out that the extensive disintegration of the basaltic rocks, that are exposed on these plains in places, affords evidence of the great antiquity of these “talasinga” districts in their present unforested condition. The extent to which these rocks have weathered downward is remarkable. In some places they are decomposed to a depth of ten feet and more. The same inference is to be drawn from the occurrence of fragments of limonite, or bog-iron ore, over these plains, marking as they do original swampy tracts that, with a few exceptions, have long since disappeared. Such deposits indicate that these plains have been for ages in the same condition. ... It may be added that, according to Mr. Lister and Mr. Crosby, the features of the “talasinga” plains occur in the Tongan Group on the leeward sides of the islands of Eua and Vavau.

NOTE 23 (page [43])
Schimper’s Grouping of the Indo-Malayan Strand-flora

It is divided into four formations—the Mangrove, the Nipa, the Barringtonia, and the Pes-capræ. The two last make up my Beach-formation, the Barringtonia formation comprising the trees, shrubs, &c., immediately lining the beach, and the Pes capræ including the creepers and bushes of the beach itself. In the Pacific islands it is not always easy to preserve this distinction. The Nipa formation corresponds in some respects with my Intermediate or Transition formation, lying as it does between the mangrove-belts and the woods of the interior; but the swamp-palm (Nipa fruticans) that forms it in the mass is not found in Fiji or, indeed, in the Pacific islands, excepting the Solomon and Caroline Groups.

NOTE 24 (page [44])
Grouping of some of the Characteristic Plants of the Strand-flora of Fiji

(a) Beach-formation.—Calophyllum inophyllum, Thespesia populnea, Triumfetta procumbens, Carapa moluccensis, Canavalia obtusifolia, Vigna lutea, Pongamia glabra, Sophora tomentosa, Cæsalpinia Bonducella, Acacia laurifolia, Barringtonia speciosa, Terminalia Katappa, Gyrocarpus Jacquini, Pemphis acidula, Morinda citrifolia, Guettarda speciosa, Wedelia biflora, Scævola Kœnigii, Cordia subcordata, Tournefortia argentea, Ipomœa pes capræ, Cassytha filiformis, Hernandia peltata, Pandanus odoratissimus, &c.

(b) Mangrove-formation.—Carapa obovata, Rhizophora mucronata, Rhizophora mangle, Bruguiera Rheedii, Lumnitzera coccinea, Scirpodendron costatum, &c. (See below.)

(c) Intermediate or Transition-formation.—Hibiscus tiliaceus, Heritiera littoralis, Smythea pacifica, Derris uliginosa, Entada scandens, Barringtonia racemosa, Cerbera Odollam, Clerodendron inerme, Vitex trifolia, Excæcaria Agallocha, &c.

N.B.—It is not possible to draw a definite line between the plants of the mangrove swamp and those of the tracts around. Several of the plants placed in the intermediate formation, such as Heritiera littoralis, Entada scandens, Excæcaria Agallocha, &c., are just as much at home amongst the mangroves. In the same way it is often difficult to distinguish between the Beach and the Intermediate formations, and plants like Cerbera Odollam, Hibiscus tiliaceus, and Vitex trifolia belong equally to both.

NOTE 25 (page [47])
The Strand-flora of the Tahitian Region

Drake del Castillo’s Flore de la Polynésie française deals mainly with the Society or Tahitian Islands, but also with the Marquesas, Paumotus, Gambier Islands, and Wallis Island. The last-named, however, lies in Western Polynesia, and is not dealt with in this connection. There is no reason to believe, judging from the general character of the islands and from Cheeseman’s memoir on the Rarotongan flora, that the strand-plants of the islands of the Cook and Austral Groups, which also belong to this region, differ materially from those of the Tahitian islands proper. Rarotonga, however, possesses Entada scandens, not recorded as a growing plant from any other part of East Polynesia, excepting perhaps Mangaia in the same group.

NOTE 26 (page [48])
The Fijian Shore-plants not found in Tahiti

Although most of these plants, such as Barringtonia racemosa, Clerodendron inerme, Entada scandens, Excæcaria Agallocha, Heritiera littoralis, Smythea pacifica, &c., have fruits that float for months, and could have reached Tahiti as readily as some of the beach-plants that have successfully established themselves, there are a few like Dalbergia monosperma, Derris uliginosa, and Scirpodendron costatum, the fruits of which only float for weeks, and it is possible that they may have been unable to reach there.

NOTE 27 (page [49])
The Intruders into the Beach-flora from the Inland Plants of Tahiti

Drake del Castillo mentions several, such as species of Boerhaavia, that could only be occasional intruders; but it is noteworthy that Gardenia tahitensis appears to be a genuine recruit from inland. The xerophilous habit of the Pacific Gardenias and their station, usually near the coast, however, would render this possible.

NOTE 28 (page [52])
The Littoral Plants of the Hawaiian Islands

Species.[[5]]		Origin.			Distribution.			Characters of fruit or seed.
		Indigenous.	Introduced.		Old World.	New Worlds.	Both Worlds.	Size.[[6]]		Buoyancy.		Attract frugivorous birds.
			By aborigines in ancient times.	By Europeans soon after discovery.				Large.	Small.	Float for months.	Sink at once or in a few days.
	Acacia Farnesiana	...	...	+	...	...	+	...	+	weeks	...	?
	Cæsalpinia Bonducella	+	...	...	...	...	+	...	+	...	+	...
	Calophyllum inophyllum	...	+	...	+	...	...	+	...	+	...	...
	Cassytha filiformis	+	...	...	...	...	+	...	+	+	...	+
	Colubrina asiatica	+	...	...	+	...	...	...	+	+	...	...
	Cordia subcordata	...	+	...	+	...	...	...	+	+	...	...
E	Cuscuta sandwichiana	+	...	...	...	...	...	...	+	...	+	...
	Cocos nucifera	...	+	...	...	...	+	+	...	+	...	...
P	Gossypium tomentosum	+	...	...	...	...	...	...	+	...	+	...
P	Heliotropium anomalum	+	...	...	...	...	...	...	+	...	+	...
	Heliotropium curassavicum	+	...	...	...	...	+	...	+	...	+	...
	Herpestis Monnieria	+	...	...	...	...	+	...	+	...	+	...
	Hibiscus tiliaceus	...	+	...	...	...	+	...	+	+	...	...
	Ipomœa glaberrima	+	...	...	+	...	...	...	+	+	...	...
	Ipomœa pes capræ	+	...	...	...	...	+	...	+	+	...	...
E	Jacquemontia sandwicensis	+	...	...	...	...	...	...	+	...	+	...
E	Lipochæta integrifolia	+	...	...	...	...	...	...	+	...	+	...
	Morinda citrifolia	...	+	...	+	...	...	...	+	+	...	?
	Mucuna gigantea	+	...	...	+	...	...	+	...	+	...	...
	Pandanus odoratissimus	...	+	...	+	...	...	+	...	+	...	...
	Portulaca oleracea	...	+	...	...	...	...	...	+	...	+	...
	Scævola Kœnigii	+	...	...	+	...	+	...	+	+	...	+
	Sesuvium Portulacastrum	+	...	...	...	...	...	...	+	...	+	...
	Tacca pinnatifida	...	+	...	+	...	+	...	+	+	...	...
	Tephrosia piscatoria	+	...	...	+	...	...	...	+	...	+	...
	Terminalia Katappa	...	...	+	+	...	...	+	...	+	...	...
	Thespesia populnea	...	+	...	+	...	...	...	+	+	...	...
	Tribulus cistoides	+	...	...	...	+	...	...	+	...	+	...
	Vigna lutea	+	...	...	...	...	+	...	+	+	...	...
	Vitex trifolia	+	...	...	+	...	...	...	+	+	...	+

Species.[[5]]		Origin.			Distribution.
		Indigenous.	Introduced.		Old World.	New Worlds.	Both Worlds.
			By aborigines in ancient times.	By Europeans soon after discovery.
	Acacia Farnesiana	...	...	+	...	...	+
	Cæsalpinia Bonducella	+	...	...	...	...	+
	Calophyllum inophyllum	...	+	...	+	...	...
	Cassytha filiformis	+	...	...	...	...	+
	Colubrina asiatica	+	...	...	+	...	...
	Cordia subcordata	...	+	...	+	...	...
E	Cuscuta sandwichiana	+	...	...	...	...	...
	Cocos nucifera	...	+	...	...	...	+
P	Gossypium tomentosum	+	...	...	...	...	...
P	Heliotropium anomalum	+	...	...	...	...	...
	Heliotropium curassavicum	+	...	...	...	...	+
	Herpestis Monnieria	+	...	...	...	...	+
	Hibiscus tiliaceus	...	+	...	...	...	+
	Ipomœa glaberrima	+	...	...	+	...	...
	Ipomœa pes capræ	+	...	...	...	...	+
E	Jacquemontia sandwicensis	+	...	...	...	...	...
E	Lipochæta integrifolia	+	...	...	...	...	...
	Morinda citrifolia	...	+	...	+	...	...
	Mucuna gigantea	+	...	...	+	...	...
	Pandanus odoratissimus	...	+	...	+	...	...
	Portulaca oleracea	...	+	...	...	...	...
	Scævola Kœnigii	+	...	...	+	...	+
	Sesuvium Portulacastrum	+	...	...	...	...	...
	Tacca pinnatifida	...	+	...	+	...	+
	Tephrosia piscatoria	+	...	...	+	...	...
	Terminalia Katappa	...	...	+	+	...	...
	Thespesia populnea	...	+	...	+	...	...
	Tribulus cistoides	+	...	...	...	+	...
	Vigna lutea	+	...	...	...	...	+
	Vitex trifolia	+	...	...	+	...	...

Species.[[5]]		Characters of fruit or seed.
		Size.[[6]]		Buoyancy.		Attract frugivorous birds.
		Large.	Small.	Float for months.	Sink at once or in a few days.
	Acacia Farnesiana	...	+	weeks	...	?
	Cæsalpinia Bonducella	...	+	...	+	...
	Calophyllum inophyllum	+	...	+	...	...
	Cassytha filiformis	...	+	+	...	+
	Colubrina asiatica	...	+	+	...	...
	Cordia subcordata	...	+	+	...	...
E	Cuscuta sandwichiana	...	+	...	+	...
	Cocos nucifera	+	...	+	...	...
P	Gossypium tomentosum	...	+	...	+	...
P	Heliotropium anomalum	...	+	...	+	...
	Heliotropium curassavicum	...	+	...	+	...
	Herpestis Monnieria	...	+	...	+	...
	Hibiscus tiliaceus	...	+	+	...	...
	Ipomœa glaberrima	...	+	+	...	...
	Ipomœa pes capræ	...	+	+	...	...
E	Jacquemontia sandwicensis	...	+	...	+	...
E	Lipochæta integrifolia	...	+	...	+	...
	Morinda citrifolia	...	+	+	...	?
	Mucuna gigantea	+	...	+	...	...
	Pandanus odoratissimus	+	...	+	...	...
	Portulaca oleracea	...	+	...	+	...
	Scævola Kœnigii	...	+	+	...	+
	Sesuvium Portulacastrum	...	+	...	+	...
	Tacca pinnatifida	...	+	+	...	...
	Tephrosia piscatoria	...	+	...	+	...
	Terminalia Katappa	+	...	+	...	...
	Thespesia populnea	...	+	+	...	...
	Tribulus cistoides	...	+	...	+	...
	Vigna lutea	...	+	+	...	...
	Vitex trifolia	...	+	+	...	+

[5]. There are three endemic species here included which are preceded by E. Two species preceded by P are confined to Polynesia. Most of the plants are at present typically littoral, though often also occurring inland.

[6]. All fruits or seeds, an inch or over in size, that could not have been transported to Hawaii by birds are regarded as large.

NOTE 29 (page [54])
Botanical Notes on the Coast-plants of the Hawaiian Islands

[The following remarks have been extracted from my journals and represent some of the field-notes of journeys made in the more interesting localities.]

(1) Walk along the Puna Coast, Hawaii, from Punaluu to Hilo (Dec. 26, 1896, to Jan. 6, 1897).—For the first two to three miles to Kamehame Point, the following plants were noticed on the flows of smooth ropy lava that formed the cliff-bound coast—Capparis sandwichiana, Jacquemontia sandwicensis, Ipomœa insularis, Lipochæta lavarum, Portulaca villosa, Tephrosia piscatoria, Tribulus cistoides, Waltheria americana, &c. Beyond this point Scævola Kœnigii was abundant in places on the old lava-flows near the sea, and further on patches of Myoporum sandwicense growing, not as a tree 20 to 30 feet high, as in the mountains, but as a prostrate shrub with fleshy leaves. Vegetation similar to that above described occurred on the surface of the old lava-flows that constituted the cliff-bound sea-border as far as Kapapala Bay. On the sandy beach at Kapapala Bay grew Ipomœa pes capræ, serving as host to Cuscuta sandwichiana. In the vicinity of the house at Keauhou there were a few Coco palms and Pandanus trees, whilst Capparis sandwichiana and Morinda citrifolia were growing on the adjacent lava-fields.

Morinda citrifolia and Tephrosia piscatoria grew on the lava flows between Keauhou and Apua. On the beach at Apua, Ipomœa pes capræ and Scævola Kœnigii were abundant, the last extending a few hundred yards inland on the lava. Further east the inland bush, made up of Cyathodes tameiameiæ, Metrosideros polymorpha, &c., descended to the coast to within a few hundred yards of the sea. In crossing the lava coast plains to Kapa-ahu I observed Morinda citrifolia growing frequently out of the cracks in the bare lava-rock, and an occasional solitary tree of Erythrina monosperma growing also from the fissures.

Before reaching Kapa-ahu we passed the site of an old coast village, named Laepuki, where there were growing from forty to fifty Coco-nut palms, as well as another village, represented by a solitary house, and named Kamomoa, where there were 27 Coco-nut palms and a few Pandanus trees. Kapa-ahu, with its numerous Coco-nut palms, was more like a South Sea coast village than any before seen; and the coast vegetation suddenly acquired a South Pacific character.

At Pulama, for instance, about a mile west of Kapa-ahu, where the ancient lava-flows, fairly vegetated, terminate at the sea in cliffs 20 or 25 feet high, there is a curious and quite unexpected development of a littoral flora such as we should see in the South Pacific. Here, growing on the broken lava surface at the brink of the cliffs and overlooking the sea, thrive Cæsalpinia Bonducella, Cocos nucifera, Ipomœa pes capræ, Ipomœa glaberrima, Morinda citrifolia, Pandanus odoratissimus, Scævola Kœnigii, Sesuvium Portulacastrum, Thespesia populnea, and Vigna lutea. This shore-belt of characteristic littoral plants is backed by vegetation more inland in its character, amongst which Aleurites moluccana, Dodonæa viscosa, Erythrina monosperma, Ipomœa insularis, I. bona nox, Osteomeles anthyllidifolia, &c., are to be observed. Such a shore-belt of typical littoral plants is rarely to be found in the large island of Hawaii; and its usual position at the margin of cliffs, and raised 20 or 25 feet above the sea, is rather suggestive of an uplift in recent times of this part of the coast.

Between Kapa-ahu and Kalapana is a low country occupied mostly by Guavas, and often turfy. At Kalapana, which is a large village situated on a grassy plain by the sea, Coco palms and Pandanus trees abound, and Mucuna gigantea and Cæsalpinia Bonducella are frequent near the coast, whilst Ipomœa pes capræ is common on the beach. Calophyllum inophyllum is planted near the houses. Here Osteomeles anthyllidifolia in its dwarfed form descends to the edge of the cliffs. About half a mile beyond Kalapana is the hamlet of Kaimu, and here among the Coco palms close to the beach I noticed four Loulu palms (Pritchardia Gaudichaudii). Beyond Kaimu the trees and shrubs of the inland wood, Metrosideros polymorpha, Cyathodes tameiameiæ, &c., descend on the spurs of old lava-flows close to the coast; whilst Pandanus and Morinda citrifolia with Mucuna gigantea are common near the sea as far as Kehena, where there are plenty of Coco palms. I approached Opihikao through as fine a Pandanus forest as I have ever seen, the large Bird’s Nest Fern (Asplenium nidus) growing half-way up their trunks, adding picturesqueness to the scene, whilst Mucuna gigantea was a common climber. Beyond Opihikao the inland woods descend to the coast. Thence on to Makuu the coasts are mostly occupied by Pandanus forests, and the lower coast road from Makuu to Hilo traverses a region where these Pandanus trees abound, extending far inland. Scævola Koenigii and Ipomœa pes capræ are common on the coast near Coco-nut Island, Hilo Bay.

It may be added that the agency of the wild goat explains the dispersal of Myoporum sandwicense, Morinda citrifolia, Tephrosia piscatoria, Waltheria americana, &c., over the almost bare surfaces of the lava flows on the Puna coast. Goat droppings were frequent under the patches of Myoporum and Waltheria. In some of them I found the entire seeds of Portulaca oleracea and the small cocci of Euphorbia pilulifera, weeds common in the district.

(2) Coasts of the Kalae Promontory and its Vicinity, Hawaii.— This is the most southerly portion of the group, and it is on the eastern coasts of this district that many of the North American drift logs are embayed and stranded. At Kamilo, to the east of the promontory, there is a long beach of calcareous sand where Heliotropium anomalum, Scævola Kœnigii, and Tribulus cistoides grow in abundance, whilst Sesuvium Portulacastrum thrives on the beach and in brackish pools. Portulaca lutea (Sol.), Ipomœa glaberrima (Boj.), and Jacquemontia sandwicensis also occur. Where the beach-sand has encroached on the adjacent lava surface, the Scævola covers extensive tracts off the beach, and is stunted. I noticed a solitary thicket of Thespesia populnea on the beach.

The actual headland of Kalae is wind-swept and covered with grass, amongst which Portulaca villosa and Sida fallax thrive. By the sea occur Scævola Kœnigii and Ipomœa pes capræ, and there is some Sesbania tomentosa near the point. Waiheiaukini beach is shut in between the lofty arid slopes of the promontory on one side and a modern lava-flow on the other side. Here Scævola Kœnigii grows in quantity, together with Ipomœa pes capræ, Tribulus cistoides, Sida fallax, and Jacquemontia sandwicensis, whilst Cuscuta sandwichiana is abundant, finding its hosts in the first four plants just named.

(3) South Kona Coast, Hawaii.—The coast here, as exemplified by that between Kapua and Hoopuloa, is mostly bare lava. Here and there, a little coral sand collects amongst the lava blocks of the rubbly shore, and it is in such places that Scævola Kœnigii and Ipomœa pes capræ find a home and apparently thrive, whilst Hibiscus tiliaceus and Morinda citrifolia grow behind. I observed Cordia subcordata and one or two specimens of Pritchardia Gaudichaudii by the coast on the south side of Milolii. Around a brackish pool at Kapua I observed Heliotropium curassavicum, and Acacia Farnesiana was to be seen growing on the beach at Okoe. On the lava coast between Hoopuloa and Papa, two miles to the north, Tephrosia piscatoria was very abundant.

(4) North Kona Coast, Hawaii.—I examined the coast between Kailua and Kiholo. White beaches are common south of Keahole Point, the coast further north being usually lava-bound with sandy beaches here and there. Heliotropium anomalum, Ipomœa pes capræ, and Sesuvium Portulacastrum are the commonest beach plants on this coast. Scævola Kœnigii is also abundant in places, whilst Tribulus cistoides and Morinda citrifolia are also fairly common on the beaches. The Morinda also grows on the adjacent lava flats; but on both sand and rock it is evidently usually self-sown, since seedlings are to be seen near the older plants. Heliotropium curassavicum is to be seen here and there on the sand all along the coast, but nearly always associated with H. anomalum. Jacquemontia sandwicensis occurs occasionally on the beach; and Cuscuta sandwichiana is abundant in places, growing generally on Ipomœa pes capræ, but sometimes on Scævola Kœnigii. Brackish water ponds are common on the coast inside the beaches, Ruppia maritima flourishing in the water, with Sesuvium Portulacastrum growing at the edges. Sometimes Hala trees (Pandanus odoratissimus) fringe the borders of the pools. I noticed Pritchardia Gaudichaudii on the coast at Kiholo, and I learned that Cordia subcordata was once common here as on other parts of the Kona coast; but it has died out as in most other localities.

(5) Kohala Coast, Hawaii.—Several littoral plants are scantily represented on the beach of black sand at the mouth of the Waimanu valley, especially Ipomœa pes capræ, Morinda citrifolia, Pandanus odoratissimus, and Scævola Kœnigii. The Pandanus covers the adjacent precipitous slopes up to a height of several hundred feet above the sea. Ipomœa pes capræ is abundant on the sand dunes backing the beach at Waipio. I observed Naias marina in the Waipio River just inside the mouth. No one seems to have recorded the plant from the group since Chamisso found it in Oahu.

(6) Hamakua Coast, Hawaii.—Not many opportunities presented themselves on this cliff-bound coast of finding littoral plants. At the mouth of a gulch between Ookala and Laupahoehoe I found growing at the coast Vitex trifolia (var. unifoliolata) in quantity, together with Morinda citrifolia, Scævola Kœnigii, and Pandanus odoratissimus, the last-named clothing the hill-slopes overlooking the sea.

(7) The Coasts of Oahu.—The littoral vegetation of the south-east portion of the island from Diamond Head round to Waimanalo is, as a rule, scanty. Ipomœa pes capræ and Tribulus cistoides prevail to Koko Head, and on the rubbly coast between that headland and Makapuu Point occur Tephrosia piscatoria, different species of Lipochæta, &c. Between Makapuu Point and Waimanalo, Scævola Kœnigii and Vitex trifolia (var. unifoliolata) are fairly abundant, the former growing on the rocky slope at the base of the cliffs, and raised perhaps some 20 feet above the sea. Along the whole east coast of the island the littoral vegetation is rarely well represented. However, Ipomœa pes capræ is common everywhere, whilst Scævola Kœnigii occurs frequently, and here and there a few plants of Morinda citrifolia are seen on the beach, while thickets of Hibiscus tiliaceus mark in some localities the mouths of streams.

On the north coast of Oahu, as on the Waialua and Waimea beaches, the one-leaved variety of Vitex trifolia is common, together with Ipomœa pes capræ and Euphorbia cordata; whilst Acacia Farnesiana is frequent on the Waialua beach, its pods being much appreciated by the cattle. Occasionally, as by the bridge at Waimea, Colubrina asiatica and Thespesia populnea are to be noticed.

Shore vegetation is a little better represented on the beaches at and near Kaena Point, the north-west corner of the island. Here on the sand we find often in abundance Heliotropium anomalum, the same variety of Vitex trifolia, Scævola Kœnigii, and Ipomœa pes capræ; whilst on the rocks bordering the beach occur Gossypium tomentosum, Jacquemontia sandwicensis, Tribulus cistoides, Vigna lutea, and more than one species of Lipochæta, the last being derivatives from the inland flora.

On the west coast of the island true shore-plants play an inconspicuous part. Ipomœa pes capræ is common on the beaches, and such plants as Acacia Farnesiana, Jacquemontia sandwicensis, Gossypium tomentosum, and Tribulus cistoides immediately border the beach. Ipomœa tuberculata is a frequent intruder as well as the recently introduced Algaroba tree (Prosopis dulcis). Acacia Farnesiana also extends inland, covering entire large areas and forming in the Waianae valley extensive thickets impenetrable for the cattle. It occupies great districts near the coast in different parts of Oahu, and with Hibiscus tiliaceus is to be found far inland. The cattle are active dispersers of its seeds. (See [Note 30].)

True beach plants are infrequent at the mouth of Pearl Harbour, although the coast is well suited for them. Here I found Heliotropium anomalum, H. curassavicum, Jacquemontia sandwicensis, Lipochæta integrifolia (a true beach plant), Herpestis Monnieria, &c. Batis maritima occurs in one or two localities around Oahu, but it is, according to Hillebrand, of recent introduction.

NOTE 30 (page [58])
The Beach-drift of the Hawaiian Islands

It was pointed out by Dole long ago in one of the Hawaiian Club Papers (1868) that the existing currents bring to this archipelago only huge pine logs from Oregon, but no tropical fruits; and Hillebrand (p. xiv.) refers to the driftwood of pine logs from the north-west coast of America, stranded on the shores of these islands. This drift seems to collect in quantity in particular localities, as on the south-east coast of Hawaii between Honuapo and the Kalae promontory (especially on the Kamilo beach near Kaluwalu) and on the east coast of Oahu; and probably there are other favourable localities for catching the drift on the northern shores of Maui and Molokai.

It was on the south-east coast of Hawaii (on the beach at Kamilo and on the eastern side of the Kalae promontory) that this drift came particularly under my notice. Here the logs are stranded in abundance, in sufficient quantity, in fact, to build a town, and they were employed for building purposes by the manager of the neighbouring sugar-cane plantation. Several of the logs are of huge size, as much as 4 feet in diameter; and they are known locally as “white cedar” and “red cedar,” and characterised as Oregon timber. Some of them are extensively burrowed by the “teredo” and other boring mollusks. Others recently stranded are covered with barnacles (Lepadidæ), whilst others that have lain long on the beach are bare. I have seen these logs occasionally washed up at Punaluu and at different places on the lava-bound Puna coast. They apparently first strike the Puna coast, and are drifted along until they become embayed near the Kalae promontory, and ultimately stranded. Mingled with them on the beaches Pandanus trunks occur in number; they evidently hail from those parts of the Puna coast where Pandanus forests prevail, and thus they indicate the direction of the drift on the coast of this island. In places there was a considerable amount of small vegetable débris, sometimes partially concealed by the sand, and containing seeds and fruits in fair quantity.

The following seeds and fruits were collected:—

• Pandanus drupes, common; most of them fresh-looking, but a few much worn.
• Thespesia populnea, a few seeds.
• Ipomœa pes capræ, seeds, fair numbers.
• Ipomœa bona nox, seeds, a few.
• Ipomœa glaberrima, seeds, a few.
• Argyreia tiliæfolia, fruits and seeds, a few.
• Strongylodon lucidum, seeds, a few.
• Cæsalpinia Bonducella, a single seed.
• Vigna lutea, seeds, a few.
• Calophyllum inophyllum, a few fruits.
• Ricinus communis, a few, the seeds either free or in the cocci, and often empty or decaying.
• Aleurites moluccana, seeds, common, none sound, either empty or containing a rotten kernel: also a single fruit.
• One or two seeds not identified.

There was seemingly a total absence of the fruits or seeds of any littoral plant not found in these islands, such as I was familiar with in the South Pacific. In the mass this seed-drift could have been derived from the neighbouring coasts of the island. This is especially indicated in the cases of the fruits and seeds of Aleurites moluccana, Ricinus communis, and Argyreia tiliæfolia. The sound seeds of Aleurites do not float, the buoyant seeds being always empty, or nearly so; and the presence of the seeds in beach-drift, as explained on page [419], is due partly to the buoyancy of the empty seed and partly to the decay of the stranded fruit, the fruits being able to float for a week or two. So, also, the seeds of Ricinus, whether free or inclosed in the coccus, do not, when sound, float longer than a week or ten days. The capsules of the Argyreia can float two or three weeks, whilst the seeds vary in their behaviour, as observed on page [20]. I noticed in places where the vegetable débris was heaped up and exposed to the sun’s heat, that some of the Ipomœa seeds were germinating. It is to be remarked that horse-dung and goat-dung are always common in the beach-drift of these islands. Seeds are sometimes to be seen in the stranded material; and it was evident that the droppings of these animals can float for some weeks before breaking down.... I may add that large sponges, apparently of no value, are thrown up in quantities on the east side of the Kalae promontory.

Excepting the pine logs, the only things coming under my notice in this beach-drift that could be characterised without hesitation as non-Hawaiian, were two well-worn pieces of acid pumice, less than an inch in size. One of them was incrusted partially by the tubes of annelids, and both of them had evidently been drifting about in the Pacific for a long period, perhaps for years. They were such as occur in abundance on the beaches of the South Pacific, and, in fact, on all the shores of the Pacific Ocean, both temperate and tropical. Although I carefully searched the stranded drift of many beaches in this group, no other specimens of drift pumice were found.

On different parts of Oahu the beach-drift was always made up of materials derived from the vegetation of the coast adjacent. Of most frequent occurrence were the seeds of Ipomœa pes capræ and Vigna lutea, and the fruits of Scævola Kœnigii, Vitex trifolia, and Pandanus odoratissimus. In addition, the empty seeds of Aleurites moluccana were numerous, and there were occasional seeds of Thespesia populnea, Colubrina asiatica, and Mucuna gigantea. On one beach there were a number of fruits of Terminalia Katappa, showing but little signs of ocean travel, and evidently derived from trees in the vicinity. This tree was introduced by Europeans; but it is not unlikely that in a generation or two it will become, without man’s aid, one of the characteristic beach trees of Oahu. It may be remarked that the pods of Acacia Farnesiana, a shrub now growing abundantly in Oahu near the sea, are washed up in great quantities on the beaches of the west coast of this island, and the seeds are to be seen germinating in numbers on the beach, the seedlings striking into the sand. The pods float unharmed in sea-water for four or five weeks, but the seeds, when freed, sink.

Although the above evidence gives no indication of tropical drift of non-Hawaiian origin on the beaches, it is probable, for reasons adduced in [Chapter VIII.], that, in the winter, drift may be brought from tropical America.

NOTE 31 (page [59])
The Inland Extension of the Shore-plants of Hawaii

Cæsalpinia Bonducella.—According to Hillebrand, this plant, so characteristic of the littoral floras of tropical regions, grows “in gulches of the lower plains on all the islands,” no reference being made to its occurrence on the beaches. It is very rarely to be seen on the beaches of the large island of Hawaii; but it is to be found on the lava-bound coasts, and from there it extends inland usually on old lava-flows for five or six miles, and reaches sometimes considerable elevations. In one locality I found it at 2,000 feet above the sea (see page [188]).

Cassytha filiformis.—Though a typical shore-plant in Fiji and other tropical localities, it is rarely so in these islands. Hillebrand says nothing of its station. It grows well in the lower open wooded regions, and is frequently found amongst the blocks of old lava-flows near the coast.

Cuscuta sandwichiana.—Unlike its fellow parasite Cassytha filiformis, this species of Cuscuta, which is confined to this group, never came under my notice away from the beach; and Hillebrand speaks of finding it only at the coast (see page 366).

Ipomœa pes capræ, as I observed it in the islands of Hawaii and Oahu, is confined to the beach or to neighbouring sand-dunes. Hillebrand makes no reference to its occurrence inland. This species in these islands offers thus a great contrast to its behaviour in Fiji.

Scævola Kœnigii.—Whilst most at home on the sandy beaches, this plant is also frequently met with in the island of Hawaii on scantily vegetated lava-flows near the coast; but I never noticed it more than a few hundred yards from the sea.

Tephrosia piscatoria.—Though it may occur on the beach, it is generally found as described by Hillebrand on the rocky or rubbly ground at the back of the beach, as well as further inland. It is common on the old lava-fields of the island of Hawaii near the coast; and, according to the natives, its seeds are disseminated by the wild goats that frequent these localities.

Tribulus cistoides.—Hillebrand observes that this plant is found along the sea-shore and on the lower plains. I found it most frequently on the beaches and on the old lava-flows near the sea.

Vitex trifolia, var. unifoliolata.—It is confined, as Hillebrand remarks, to the beaches. Neither in Oahu nor in Hawaii did I ever find it straying inland, which is the more remarkable since this variety, or one closely similar to it, is one of the most characteristic inland plants of the Fijian strand-flora.

Vigna lutea.—This plant was found by me growing on the beaches and in their vicinity. Hillebrand merely speaks of it as “growing at short distances from the shore.”

Some of the trees, usually littoral in their station in the tropical Pacific, which are regarded as having been introduced in early times into the Hawaiian group by the Aborigines (see [Chapter VII.]), behave, nevertheless, quite like indigenous plants in the inland regions and in the lower levels. This is true, for instance, of Hibiscus tiliaceus and Pandanus odoratissimus, the last-named forming forests at the sea-board extending in places far up the mountain slopes. The same, however, may be said of other plants known to have been introduced since the discovery of the islands, as in the cases of Cactus Tuna and of Ricinus communis; and it also applies to Aleurites moluccana, the Candle-nut Tree, which, although it could only have been introduced by the Aborigines, now forms forests on the lower slopes of the mountains.

NOTE 32 (pages [19], [112], [165])
The Fijian Species of Premna

I was much interested in the small trees and shrubs of this genus in Fiji, more especially on account of the relation between the shore and inland species. This is an Old World genus containing some eighty species mainly characteristic of tropical Asia and Malaya, and represented in the South Pacific archipelagoes by two species, one Premna taitensis or tahitensis, spread over the region and very near P. integrifolia, an Asiatic species; the other Premna serratifolia, an Asiatic plant found in Fiji, the Marquesas, and other groups. Without endeavouring to give a precise value to the Fijian plants, I will merely describe the prevailing forms, which are, however, connected by intermediate varieties. These trees, I may add, are known by the same name in the various Pacific groups, “Avaro” or “Avalo” in Tahiti, “Alo-alo” in Samoa, “Yaro” and “Yaro-yaro” in Fiji.

The Fijian plants may be thus described.... (a) Premna serratifolia, an inland tree, growing in open woods and on the outskirts of the forest, 25 to 30 feet high, more or less hairy, leaves coarsely serrated with long tapering points, putamen prominently tuberculated and thick-walled.

(b) Premna taitensis or P. integrifolia, a low straggling coast tree or shrub of the beaches, the coral islets, the swampy borders of the estuaries, and the inland talasinga plains, its usual height being eight to ten feet, except in the inland plains, where it is dwarfed, and three to five feet high. It is more or less glabrous, the leaves being typically entire with obtuse or retuse and mucronate apices. The putamen is thin-walled and relatively smooth. (c) Intermediate forms found generally in the inland plains or talasinga regions.

On the Modes of Dispersal.—Speaking generally, the small drupes of both species float at first, but the soft parts are soon removed by decay, and the stone is freed. In the case of the coast species, P. taitensis, the stones float indefinitely and are often found afloat in rivers. In the case of the inland tree, P. serratifolia, most of the stones sink at once, whilst the others sink in a few days. It is probable that currents are one of the effective agencies in distributing the coast species, but this could not apply to the inland tree. The fruits of both the inland and the coast species would attract birds, and the stones would resist injury in their crops. This is the agency advocated by Prof. Schimper for the shore species, P. integrifolia, of Indo-Malaya; and fruits referred with a query to this genus were found in the collection of seeds and fruits obtained by me from the crops of pigeons in the Solomon Islands (Bot. Chall. Exped., Introd. p. 46, part IV. p. 312).

On the Cause of the Buoyancy of the Stone or Putamen of the Coast Species.—This is primarily connected with the empty seed-cavities, the four-celled stone usually developing only one seed, the other cavities being empty. This inference was established by the dissection of a large number of stones, but it will be seen from the table below that one-seeded stones are also frequent in the case of the inland tree (P. serratifolia), where they as a rule sink. With either species the substance of the stone has no floating power, but with the shore species, on account of the thin-walled stone, the empty seed-cavities cause it to be specifically lighter than water whilst with the inland species the walls of the stone are so thick that the empty spaces of the unfilled seed-cavities do not effect the same result. It may be remarked that when the coast species grows in the inland plains the buoyancy of the stone is preserved.

	One-seeded stones.	Two-seeded stones.	Three-seeded stones.
Inland tree (P. serratifolia)	73 per cent.	23 per cent.	4 per cent.
Coast tree (P. taitensis)	92 per cent.	8 per cent.	—

NOTE 33 (page [63])
De Candolle’s List of Plants dispersed exclusively by Currents

Drepanocarpus lunatus; Ecastaphyllum Brownei; Mucuna urens, D.C.; Tephrosia piscatoria; Hibiscus tiliaceus; Rhizophora mangle; Guilandina Bonduc, Linn.; Ipomœa pes capræ; Canavalia obtusifolia.

I have experimented on the buoyancy of the fruits and seeds of all these plants excepting the two first named. In five species the seeds float in sea-water unharmed for several months. With Rhizophora it is the floating seedling that disperses the plant. Neither the pods nor the seeds of Tephrosia piscatoria are suited for dispersal by the currents.

NOTE 34 (page [64])
The Littoral Plants of the Easternmost Polynesian Islands

Except in the case of Hernandia peltata my authority here is the Botany of the “Challenger” Expedition. Mr. J. H. Maiden gives some further details of the flora of Pitcairn Island in a more recent paper (Austral. Assoc. Rep., Melbourne, 1901, vol. 8), and Hernandia peltata is included in his list.

NOTE 35 (page [68])
Distribution of the Littoral Plants with Buoyant Seeds or Fruits that are found in the Fijian, Tongan, Samoan, Tahitian, and Hawaiian Groups

This list probably contains nearly all the species of the Polynesian region, but it is not implied that these plants have been recorded from all the groups (vide infra).

(a) Species found only in the Old World.—Calophyllum inophyllum, Hibiscus diversifolius, Thespesia populnea, Heritiera littoralis, Kleinhovia hospita, Carapa moluccensis, C. obovata, Smythea pacifica, Colubrina asiatica, Mucuna gigantea, Erythrina indica, Strongylodon lucidum, Dalbergia monosperma, Pongamia glabra, Inocarpus edulis, Derris uliginosa, Afzelia bijuga, Barringtonia racemosa, B. speciosa, Rhizophora mucronata, Bruguiera Rheedii, Terminalia Katappa, T. littoralis, Lumnitzera coccinea, Pemphis acidula, Morinda citrifolia, Guettarda speciosa, Wedelia biflora, Scævola Kœnigii, Cerbera Odollam, Ochrosia parviflora, Cordia subcordata, Tournefortia argentea, Ipomœa glaberrima, I. grandiflora, I. peltata, Aniseia uniflora, Clerodendron inerme, Vitex trifolia, Hernandia peltata, Excæcaria Agallocha, Tacca pinnatifida, Cycas circinalis, Pandanus odoratissimus, Scirpodendron costatum.

(b) Species occurring in both the Old and New Worlds.—Hibiscus tiliaceus, Suriana maritima, Ximenia americana, Dodonæa viscosa, Canavalia obtusifolia, C. ensiformis, Vigna lutea, Sophora tomentosa, Cæsalpinia Bonduc, C. Bonducella, Entada scandens, Gyrocarpus Jacquini, Luffa insularum, Ipomœa pes capræ, Cassytha filiformis, Cocos nucifera.

(c) Species occurring in America to the exclusion of the Old World.—Dioclea violacea, Mucuna urens, Rhizophora mangle.

(d) Species found only in Polynesia.—Canavalia sericea, Mucuna platyphylla(?), Cynometra grandiflora, Serianthes myriadenia, Parinarium laurinum(?), Premna tahitensis.

Remarks.—Of these seventy plants there is not one that has not come within the scope of my observations and experiments. The West Coast of Africa is included in the American region for reasons given in [Chapter VIII]. For the other authorities on the buoyancy of these seeds and fruits reference should be made to the list given under [Note 2] and to other parts of this work. About one or two of the plants, like Ipomœa peltata, one scarcely knows whether they are most characteristic of the coast-flora or of the inland-flora.

NOTE 36 (page [72])
Hawaiian Plants with Buoyant Seeds and Fruits known to be dispersed by the Currents either exclusively or, as in a few Species, with the Assistance of Frugivorous Birds

Colubrina asiatica.—Usually regarded as confined to the Old World; but since nearly all the species are American, that continent may be considered as the probable home also of this species. Hillebrand gives it a locality in the West Indies.

Dioclea violacea.—Tropical America.

Mucuna gigantea.—Old World.

Mucuna urens.—America, and extending to the African West Coast, which is to be included in the American region of shore-plants.

Strongylodon lucidum.—Old World.

Vigna lutea.—Old and New Worlds.

Cæsalpinia Bonducella.—Old and New Worlds.

Scævola Kœnigii.—Usually regarded as confined to the Old World, but according to the synonymy accepted by some authors it is also to be ascribed to America. The genus is chiefly Australian, and it is possible that the littoral species may have reached America through the agency of birds, since all the species of the genus possess fruits that would attract frugivorous birds.

Ipomœa glaberrima (Boj.).—Old World.

Ipomœa pes capræ.—Old and New Worlds.

Vitex trifolia.—Old World. The genus is also dispersed by pigeons.

Cassytha filiformis.—Old and New Worlds. Like Scævola the genus is chiefly Australian, and here, also, the fruits of the littoral species are not only dispersed by the currents, but are known to be also disseminated by fruit-pigeons.

It is possible that birds may have taken a predominant part in the dispersal of the species of Scævola, Vitex and Cassytha.

There thus remain nine species for consideration. Of these two are exclusively American, three are found in both the Old and New Worlds and four are usually regarded as exclusively Old World plants, but one of them (Colubrina asiatica) has a fair claim to be regarded as of American origin. Thus it is quite possible that six out of these nine plants were brought to Hawaii from America through the agency of the currents.

NOTE 37 (page [78])
On Vivipary in the Fruits of Barringtonia racemosa and Carapa obovata

As observed by me in the Rewa delta, Fiji, there was no external evidence of such a process in the case of the fruits on the trees; but I did not pay very special attention to the matter, and it will be gathered from [Chapter XXX.] that the initial stage of germination may show no indication in the appearance of the fruit. More observation is needed for both species. As indicated in [Note 50], the structure of the seed of Barringtonia racemosa is suggestive of a lost viviparous habit. With regard to Carapa, Schimper (p. 43) remarks that he has never observed vivipary; but Miquel, in his Flora Indiæ Bataviæ, particularly speaks of the seeds germinating in the capsule. I think this is very likely, and that perhaps even the rupture of the capsule may be partly due to this cause.

NOTE 38 (page [78])
On the Temperature and Density of the Surface-water of the Estuaries of the Rewa River in Fiji, and of the Guayaquil River in Ecuador

(a) The Rewa Estuary.—My observations were made mostly in the warm, wet seasons, from October to January, 1897-99, and generally in the vicinity of the Roman Catholic Mission. The density varied usually between 1·000 and 1·010, the water being quite fresh after heavy rains inland. Though the density was usually greatest at high water, this was by no means always the case. The temperature of the water in dry weather varied from 79° to 84° F. With the river in flood after heavy rains it fell to 75° and 76°. As a rule, the fresher the water the lower the temperature, but this was not invariable. There was evidence of super-heating in the estuary, the water there having sometimes a temperature of 82° or 83°, when the water higher up the river as far as Viria was two or three degrees cooler, the sea-temperature being 79° to 80°. The average temperature of the water of the estuary during the season would be 80 to 81°.

(b) The Estuary of the Rio Guayas, also known, as the Guayaquil River.—My observations were made in the last week of February and in the first half of March, 1904. Whilst the sea-temperature a few miles off the Ecuador coast varied from 76° to 80° F., the water of the estuary from the mouth up to Guayaquil ranged from 79° to 86°, whilst rather higher up the river the temperature was about 79° or 80°. The super-heating of the estuary is thus directly indicated. It was well marked in the lower part of the estuary during one of my ascents of the river.

Surface-temperatures of estuary of the Guayaquil River, March 13, 1904, 11 a.m. to 4 p.m.; tide running up.

Sea-temperature 5-10 miles off the mouth	79·7
Estuary-temperature at the mouth, off Puna	82·7
Estuary-temperature 3 miles above Puna	84·4
Estuary-temperature 15 miles above Puna	86·5
Estuary-temperature 25 miles above Puna	82·5
Estuary-temperature off Guayaquil	81·8

The water of the estuary was, as a rule, cooler with the ebbing tide.

The density of the estuary-water at the mouth opposite Puna during the two days the ship was in quarantine ranged from 1·004 to 1·016, being generally about 1·010, and salter with the up-going tide. Off Guayaquil the water during the ebbing tide was quite fresh and, from an Ecuadorian standpoint only, potable, whilst at high water it may be a little brackish. The sea-water has much freer access to the channels in the mangrove-district at the back of the city of Guayaquil, where at high water I found the density to be 1·014.

Off Puna, on Feb. 25, I noticed that the surface-current which was running down the stream was from one to two fathoms deep, whilst below it was a strong current running up the river which carried my thermometer up against the surface-current.

NOTE 39 (page [82])
On the Pacific Species of Strongylodon

Hillebrand in his Hawaiian Flora, following Seemann, regards S. lucidum, Seem., and S. ruber, Vogel, as one species found in Fiji, Hawaii, and Tahiti, and by the former placed also in Ceylon. Hillebrand and Seemann are followed by Drake del Castillo as regards the Tahitian species. Taubert, in his monograph on the Leguminosæ (Engler’s Pflanz. Fam., Teil 3, Abth. 3, 1894), takes the same view of the Polynesian species and of its wide distribution. However, in the Genera Plantarum and in the Index Kewensis, the Asiatic and Polynesian species have been always kept apart. The two species of the genus mentioned in the first work are increased to five in the Index Kewensis, viz., one in Fiji (S. lucidum), one in Hawaii (S. ruber), two in Madagascar, and one in the Philippines.

NOTE 40 (page [88])
Precautions in Testing Seed-buoyancy

Many seeds and fruits require a few hours’ soaking before they sink; and when small they will rest a long time on the surface of still water, but a touch with the finger or a drop of water will send them to the bottom. A few will float a few days (3 or 4) before sinking; but such are included in the non-buoyant group. Only in rare cases does prolonged drying increase the period of flotation by more than a few days, examples being given at the end of the Table of Buoyancy results under [Note 10]. Adherent air-bubbles, a common cause of adventitious buoyancy, must always be removed.

NOTE 41 (page [91])
The Buoyancy of the Seeds of Convolvulus Soldanella in Fresh Water and Sea-water compared

The experiments were commenced at the close of September, 1894, and covered six months. At the end of this period in Mr. Millett’s experiment, 56 per cent. of the seeds were afloat in fresh water, and 62 per cent. in sea-water; whilst in my own experiment 72 per cent. floated in fresh water, and 65 per cent. in sea-water. I was indebted to Mr. Millett’s courtesy for the seeds.

NOTE 42 (page [96])
On Secular Changes in Sea-density

Exact data bearing on this subject are not at my disposal; but it would seem that geologists have formed conflicting conclusions from similar premises. There is the view that the composition of the ocean water was very different in early geological periods (Encycl. Brit., x., 221); but I should imagine that the character of the crustacean fauna of those seas would negative any great divergence from the present condition. Suess implies that the ancient seas carried the same minerals in solution that they do now, and it is to be inferred in a similar proportion (E. de Margerie’s French edition of Das Antlitz der Erde, ii., 343 and 345).

NOTE 43 (page [102])
On the Mucosity of Small Seeds and Seed-like Fruits when wet

I paid considerable attention to this subject from the standpoint of dispersal some years ago, and published most of the results in Science Gossip for Sept., 1894. This peculiar quality of seeds had been noticed by Dr. Kerner in his Pflanzenleben (vol. i., 1887-91), and was regarded as illustrating a mode of dispersal of seeds by adherence. As a rule, such seeds when placed in water become coated with mucus in a few minutes, or within an hour, and when allowed to dry on feathers they adhere as firmly as if gummed. I found that this quality is not affected by prolonged drying, as in the cases of Nepeta glechoma and Salvia verbenaca, where it was exhibited to the same degree after the seed-like fruits had been kept from one to three years. I especially tested about 110 British plants that were likely to display this quality, and found that about a dozen exhibited it in a marked degree, and if to these we add those plants with seeds that display it to a limited extent so that they merely become adhesive when wetted, the total would be nearly twenty. It will be noticed from the list subjoined that the plants showing marked mucosity belong to twenty genera and to ten families, the Labiatæ and Cruciferæ predominating. Although in some genera, like Plantago, there is reason to suppose that the seeds of all the species would behave in this fashion, it would be wrong to infer that this is usually the case, six genera being indicated below to which such a rule would not apply, and doubtless the number could be extended. These plants in England mostly occur at the roadside, on waste ground, and in dry meadows. It may be added that although in most cases the seeds appear in water to emit mucus, “exuded mucilage” being the expression used in the English edition of Kerner’s work, in some instances, as with Helianthemum vulgare, there appears to be a dissolving process affecting the outer seed-covering.

I. Plants with Seeds or Seed-like Fruits that emit Mucus to a Marked Degree when placed in Water.

• Arabis thaliana, G. Cruciferæ.
• Camelina sativa, K. Cruciferæ.
• Teesdalia, K. Cruciferæ.
• Capsella bursa-pastoris, G. Cruciferæ.
• Lepidium sativum, D. Cruciferæ.
• Helianthemum vulgare, G. Cistaceæ.
• * Viola tricolor (Field Pansy), G Violaceæ.
• Linum usitatissimum, D. Linaceæ.
• Linum, K. Linaceæ.
• * Matricaria chamomilla, K. G. Compositæ.
• * Senecio vulgaris, G. B. Compositæ.
• Collomia, K. Polemoniaceæ.
• Gilia, K. Polemoniaceæ.
• * Veronica beccabunga, S. Scrophulariaceæ.
• Ocimum basilicum, K. Labiatæ.
• Salvia verbenaca, G., &c. Labiatæ.
• Salvia, K. B. Labiatæ.
• * Nepeta glechoma, G. Labiatæ.
• * Dracocephalum, K. Labiatæ.
• Prunella vulgaris, G. Labiatæ.
• Plantago, K. Plantagineæ.
• Plantago major, lanceolata, maritima, G. Plantagineæ.
• Luzula campestris. G. Juncaceæ.

Explanation of Abbreviations.—The capital letter following the name indicates my authority, which is not necessarily the oldest in each case: B = Beal; D = Darwin; G = Guppy; K = Kerner; S = Scott Elliot. The respective works quoted will be found at the end of this volume. The papers of Darwin quoted will be found in Journ. Linn. Soc., “Botany,” vol. i., 1857, and in the Gardeners Chronicle for 1855.

The asterisk is placed before those genera of which other species examined by me exhibited no mucosity; these species are Arabis hirsuta, Viola canina, V. palustris, Matricaria inodora, Senecio aquaticus, Veronica agrestis, V. arvensis, Nepeta cataria, Dracocephalum canariensis.

II. Plants with Seeds or Seed-like Fruits which in my Experiments only exhibited Mucosity in a Slight Degree, becoming merely “Sticky” or Adhesive when placed in Water.

Arabis albida, Chrysanthemum leucanthemum, Lamium purpureum (occasionally), Thymus sp., Juncus bufonius, J. communis, J. glaucus, J. squarrosus.

III. Plants with Seeds or Small Fruits that exhibit Adhesiveness in the Dry State and are apt to stick to one’s fingers.

Adenostemma viscosum, Lycopus europæus, Piper Macgillivrayi, &c. One may include here also Lagenophora (see page [276]) as well as the familiar instances of Pisonia (page [347]) and Boerhaavia (page [356]).

NOTE 44 (page [121])
On the Effects of Inland Extension on the Buoyancy of the Seeds or Fruits of Littoral Plants

When in Fiji I experimented on the buoyancy of the following beach-plants that had extended far into the interior of Vanua Levu, as will be found described in [Note 22]. Those tested were Cassytha filiformis, Cerbera Odollam, Ipomœa pes capræ, Morinda citrifolia, Premna tahitensis, Scævola Kœnigii, and Tacca pinnatifida. In all but Cerbera Odollam, where I contented myself with establishing that the fruits floated buoyantly in sea-water, the experiments were prolonged for many weeks and often for several months; and in some cases, as with Ipomœa pes capræ, three or four experiments were made on seeds from different inland localities. The result was to establish in all cases that the floating powers were as great with the inland as with the coast plants of the same species; nor could any structural difference of importance be noticed. It should be observed that there is every reason to believe that the “talasinga” plains of Fiji have been occupied by the intruding beach-plants for many ages.

NOTE 45 (page [122])
Tabulated Results of the Classification, according to Schimper’s Application of the Natural Selection Theory, of the Buoyant Seeds and Fruits of the Tropical Littoral Plants on the Basis of the Structural Characters concerned in Buoyancy

Region.	Classification of species.						Proportion of non-adaptive and adaptive species.		Total number of species dealt with.
	Non-adaptive.				Adaptive.
	First group.		Second group.		Third group.		Percentage of non-adaptive species.	Percentage of adaptive species.
	Number.	Per cent.	Number.	Per cent.	Number.	Per cent.
Pacific Islands	27	40	10	15	30	45	55	45	67
Pacific Islands, tropical America, and Indo-Malaya	28	35	12	15	40	50	50	50	80

Note.—If to the last we add the eight British shore plants, the buoyant fruits of which are described in [Chapter XII.], three non-adaptive and five adaptive, we get a proportion of adaptive species for temperate and tropical regions of fifty-one per cent. This is probably fairly typical of the world generally; but it must be remembered by the reader that the author regards them all as non-adaptive. In that case, the table can be used for the numerical results of the three groups which are based only on structural characters without reference to any theory.

NOTE 46 (page [124])
On the Modes of Dispersal of the Genus Brackenridgea.

Seed-vessels of this genus found afloat in the New Guinea drift are described by Mr. Hemsley as having two curved cavities crossing each other one containing a seed, the other empty. “This empty cavity,” it is stated “gives the fruit its buoyancy” (Bot. Chall. Exped., iii., 289; plate 54) Dr. Beccari, in the English edition of his Wanderings in Borneo, p. 187, speaks of the closed air-containing cavities in the seed-vessels, or rather “stones,” of this genus as probably giving them buoyancy and thus enabling them to be dispersed by currents. He points out that the fleshy covering of these fruits would also aid their dispersal by birds. The Italian botanist implies that the two Bornean species grow in swamps. The Fijian species, as observed by me in flower in Vanua Levu, grew in the dry talasinga districts bordering the Mathuata coast, the locality where Seemann found the plant. One of the most recent accounts of the genus is given by Van Tieghem in his memoir on the Ochnaceæ in Ann. des. Sci. Nat. Bot., tome 16, 1902. According to him there are nine species, all from Malaya and New Guinea, with the exception of one in Fiji. Previous authors have also referred to Queensland and Zanzibar species. However, all the species have a limited distribution, a fact which plainly assigns to birds the principal share in the dispersal of the genus.

NOTE 47 (page [125])
On the Transport of Gourds by Currents

Small calabashes or bottle-gourds are not uncommonly to be found floating in the Fijian estuaries and stranded on the beaches; and I have also found them in the sea off the coasts. They are usually more or less globular, 3 or 4 inches across, and are evidently able to float for very long periods and to carry the seeds unharmed. Most of those I examined from the drift were dry inside and contained the seeds dried together into a loose ball about an inch in size. The seeds are not those figured in Gaertner’s De Fructibus et Seminibus, as belonging to Lagenaria vulgaris, and more resemble those of Cucurbita, but are non-buoyant. One of these gourds, picked up by me in the sea in Fiji, was placed in sea-water, and two months later was still floating buoyantly. After being then kept dry for seven months, it was broken open; and ten of the seeds were put in soil, two of them germinating in a few days.

In Ecuador gourds similar in size and shape were frequently observed by me floating in the drift of the Guayaquil River and stranded on the sea-beaches. The seeds are similarly caked together in a loose mass in the cavity of the fruit. Their characters indicate that they belong to another species of gourd; and they differ also from the Fijian seeds in their buoyancy, some of them in my experiments floating two months and afterwards germinating.

It has been known since the days of Ström and Gunnerus, two Norwegian naturalists of the 17th century, that gourds and calabashes are from time to time stranded with other Gulf-stream drift on the coasts of Norway. We learn from Sernander that those found are usually worked calabashes; but he alludes to one that was unworked and contained several seeds (see Sernander, p. 119).

It is scarcely likely that a seed-carrying gourd stranded on a beach would be able to establish the plant without the aid of man; but it seems highly probable that gourds have often been introduced into new countries by the currents and that man has afterwards cultivated them. These plants may be contrasted with that remarkable Cucurbit, Luffa insularum, a genuine littoral plant, the seeds of which, and not the fruits, are dispersed in the Pacific by the currents (see page [426]).

NOTE 48 (page [126])
On the Useless Dispersal by Currents of the Fruits of the Oak (Quercus robur) and other Species of Quercus, and also of the Hazel (Corylus avellana)

The fruits of different species of Quercus are of not infrequent occurrence in the seed-drift both of the temperate and tropical regions, being brought down by the rivers to the sea and then stranded on the neighbouring beaches. They were amongst the drift gathered by Mr. Moseley in the open sea, 70 miles off the New Guinea coast (Bot. Chall. Exped., iv., 294). I found them on the beaches of Keeling Atoll where no oak exists, and on the beaches of the south coast of Java; whilst Prof. Schimper noticed them among the stranded drift of the Java Sea, and Prof. Penzig found them stranded on the shores of Krakatoa. They also came under my notice on the Sicilian beaches and on the Italian coast at Cumæ. Those of Quercus robur are to be found on the English beaches and in the autumn drift of the Thames, but they soon sink and disappear from river-drift. They are referred to by Dr. Sernander as frozen with other floating seeds in the ice of the Scandinavian rivers; but he evidently does not regard them as possessing much independent floating power.

Some years ago the author made a number of experiments on the buoyancy of the acorns of Quercus robur, and he formed the conclusion that when freshly collected not more than 4 to 8 per cent. of mature fruits will float in fresh-water, and not more than about 10 to 12 per cent. in sea-water, but that in either case they all sink in a day or two. Immature acorns float much longer, and it is these that mostly figure in the drift. However, unlike most fruits of little initial buoyancy the mature fruits gain considerable floating power by drying. Of some that had been kept for seven months 20 per cent. floated after four weeks in sea-water and 15 per cent. after 10 weeks.... It may be added that, according to Thuret, the fruits of Quercus ilex have little or no floating power.

The buoyancy of the fruits of Quercus is due entirely to the cavity left by the shrinking of the kernel. I never remember to have found one with a sound seed amongst the drift in England and Sicily; and I should doubt much whether those in the tropical drift retain their germinating powers. But, apart from this, the genus Quercus finds in its own constitution or habit the greatest obstacle in most species to the adoption of a littoral station. However, there are exceptional tendencies displayed by the evergreen oaks; and this is very significant, since in their xerophilous leaves they possess the preliminary qualification for a station near the sea. Quercus ilex, it is well known, shows a partiality for the sea-air, and Q. virens, the “live oak,” flourishes near the sea in the southern states of America, a maritime variety being distinguished by botanists. One of the willow-oaks of America, Q. phellos, which grows in swampy land, also has a beach variety.

The Hazel-tree (Corylus avellana) must be placed in the same category with Quercus. I found the empty nuts commonly amongst the stranded drift of the Sicilian and English beaches. The fruits were also frequently noticed by Dr. Sernander in the Scandinavian sea-drift; but he says nothing of their empty condition. Mr. Darwin remarks, in the Origin of Species, that he found that fresh hazel-nuts sank, but that after drying a long time they floated for ninety days and subsequently germinated. The floating-power is no doubt due to the cavity arising from the shrinking of the kernel, and it is to this cause that Dr. Sernander attributed the slight initial buoyancy observed by him. However, the hazel, like the common oak, lacks the habit that would fit it for a station by the sea, and, whatever capacity its fruits may possess for dispersal by currents, it is quite useless for the spread of the species.

NOTE 49 (page [131])
On the Distribution of Ipomœa pes capræ, Convolvulus soldanella, and Convolvulus sepium

Whilst Ipomœa pes capræ is cosmopolitan in the tropical zones, Convolvulus soldanella is cosmopolitan in both the north and south temperate zones; but, as might be expected, the two species at times meet and their areas overlap. Thus, according to Mr. Cheeseman (Trans. New Zealand Inst., xx., 1887), they meet in the Kermadec Islands, in the South Pacific, in about latitude 30°. From my observations on the coast of Chile it would seem that C. soldanella in its northward extension fails somewhere between Valparaiso and Coquimbo, that is to say, between 33° and 30° S. lat. Gay merely refers to the plant as existing in North Chile, which in his time would include the coast between 33° and 24° S. lat. It intrudes within the “thirties” on the coast of California and is found in Madeira in about 33° N. lat. Ipomœa pes capræ in its turn extends into subtropical regions, being recorded from the Kermadecs, as above noted, and from the Bermudas in 32° N. lat. Owing probably to special physical conditions of the coast, which are referred to in [Chapter XXXII.], this plant is evidently limited to the tropics on the west coast of South America. It did not come under my notice on the beaches of North Chile, and it is apparently not mentioned by Gay in his work on the Chilian flora.

Convolvulus sepium, the frequent inland associate of the littoral C. soldanella over the temperate regions of the globe, belongs to the same section of the genus (Calystegia). Its extraordinary occurrence by itself in the island of St. Paul, in the Southern Ocean, about fifty yards from the shore (Bot. Chall. Exped., ii., 153, 264), almost suggests that we have here a dimorphic species with a littoral and an inland form; and its existence in the Azores is in this connection very remarkable. It may be here noted that of three plants raised from seeds found in the beach-drift near Palermo two had the foliage of C. sepium and one of C. soldanella. Perhaps one of my readers, in imitation of De Vries with Œnothera, might be able to settle this point by raising some hundreds of seedlings from the seeds of the beach species. It is possible that the relation between these two species of Convolvulus may be in some respects akin to that between Cæsalpinia Bonducella and C. Bonduc, two littoral plants that accompany each other over much of the tropical zone.

The student of dispersal will, however, find some curious gaps in the distribution of Convolvulus soldanella even in the temperate regions; and it will be curious to observe how they affect the distribution of C. sepium. He will have to answer the query of De Candolle:... “Admitting, if one wishes, that the currents have transported this marine species, how comes it that it chances to be in the Pacific and in Europe, without occurring on the east coasts of America and on the east and west coasts of Africa?” (Geographie Botanique, ii., 1056). He will have to explain why some botanists give C. soldanella a habitat in the tropics, as in the Indian region. Schimper, who investigated this point, says that he arrived at no certain result (p. 127). See Notes [13] and [41] and pages [29], [91], for further remarks on these two species of Convolvulus.

NOTE 50 (pages [79], [132])
On the Structure of the Seeds and Fruits of Barringtonia

As regards the fruits and their coverings, the littoral and inland species of Fiji evidently fall into different sections, the first named (B. speciosa and B. racemosa) being distinguished by their outer fibrous husk, to which the buoyancy is due, the last-named (B. edulis and an undescribed species) possessing a hard stone surrounding the seed, and here the fruits sink or float only for limited periods.

The fruits of B. edulis have an outer almost fleshy covering, a little fibrous at the outside, and the hard ligneous “stone,” containing an edible seed, requires a hammer to break it. They float heavily for three or four weeks, whereas those of the littoral species float for many months. In the case of another inland species found by me growing as a small tree 12 feet high on the slopes of Mount Seatura in Vanua Levu at an elevation of 1,000 feet above the sea, the seed was similarly protected by a hard “stone” that could only be broken with an axe, and the fruit was non-buoyant, with thin and perishable outer coats.

This mountain species of Fiji, which I may name Barringtonia seaturæ, has the general habit of B. racemosa, with which the natives persisted in linking it; whilst the fruit and foliage come nearer to those of B. edulis. The leaves are entire, taper at the base, and have a petiole 1 inch long. The fruits are oblong, at least 3 inches in length, and are obscurely angled.

It would appear from Schimper’s description (p. 173) that the fruits of the Malayan Barringtonia excelsa possess both the hard stone-shell of the inland Fijian species and the dry air-bearing fibrous husk of the littoral species. This is of special interest, since the tree is both a coast and an inland species.

The following notes on the structure of the seeds of Barringtonia were made whilst I was drifting about in my canoe in the creeks of the Rewa delta in Fiji; and whatever may be their deficiencies they have the merit of having been written in the home of the plants.... When we cut across a seed like that of B. racemosa or B. speciosa, we observe that the different parts of the embryo are indistinguishable, being united into a homogeneous, firm, fleshy mass. But if we look closely we notice a central fusiform portion marked out from the surrounding parts by a faint line, along which a delicate membrane of vascular tissue has been developed. When “germination” begins, though, as the reader will subsequently perceive, this term is here hardly appropriate, the real nature of this singular structure becomes more apparent, as is indicated in the accompanying figure. The central fusiform portion proves to be the young plant without cotyledons and growing at either end to form the root and the stem. The delicate investing membrane becomes thicker and more apparent as germination proceeds, extending upwards and downwards with the growth of the stem and root and forming a cortical covering in either case. The investing fleshy portion of the seed, which is now separable with the fingers, remains attached to the lower part of the seedling for some time, being evidently a source of nutriment, and gives a bulbous appearance to the young plant. Young bulbous plants of B. racemosa, 1 to 2 feet high, are very common on the edge of Fijian mangrove swamps where the parent tree thrives. The seedlings of B. speciosa have the same appearance, but the outer fleshy part of the bulb is not so thick.

[To face page [574].

B. racemosa., B. speciosa.

Diagrams illustrating the structure of the growing seeds of Barringtonia (two-thirds the natural size). That of B. speciosa represents a seed removed from a fruit displaying the young plant protruding two or three inches. That of B. racemosa represents the lower end of the seedling when the plant is eighteen inches high.

• a = the exorhiza.
• b = the neorhiza invested by the medullary sheath.

This structure of the seeds of Barringtonia speciosa and of B. racemosa was for a long time meaningless to me, until one day, whilst seated on the banks of the Lower Rewa, with a number of the sected seeds and bulbous seedlings gathered around, I reflected that the fruits of the latter species that floated past me in the river-drift were nearly always germinating. This called up “vivipary” to my mind; and as I looked at the Rhizophora seedlings dangling from the branches of the mangrove-trees close by, it occurred to me that this seed-structure might be the result of a lost viviparous habit. One apparently had to deal here not with an ordinary seed containing an embryo in the midst of albumen, but with a seed in an arrested stage of germination surrounded by a body that might perhaps prove homologous with the “cotyledonary body” of Rhizophora. The process of development that goes on without a break in Rhizophora, from the fertilisation of the ovule to the detachment of the seedling from the branch, was here, as I considered, arrested after germination had begun, but before the protrusion of the seedling from the fruit. With nearly all plants, as I reflected, there is a rest-stage of varying length, which might be called the seed-stage. With the mangrove-genera, Rhizophora and Bruguiera, I had convinced myself by a long series of observations, the results of which are given in [Chapter XXX.], that this rest-stage does not exist. It occurs, I argued, in Barringtonia, but only after germination has begun, and, therefore, displaced when compared with the typical seed-stage of most plants.

In this connection it may be noted that a difference in germinating behaviour might be expected between the two shore species on account of their difference in stations, Barringtonia speciosa growing on the sandy beach, and B. racemosa in the wet ground around a mangrove-swamp. There is a strong suspicion that the rest-stage in B. racemosa is very short, though I never found germination in progress on a tree (see [Note 37]). There is no doubt, on the other hand, that the rest-stage of B. speciosa is often, as with most other plants, very long. This, then, was my lesson from the Barringtonia fruits on the banks of the Rewa, and the question arose whether this interpretation of these curious seed-structures accorded with the opinion formed of their nature by botanists.

Curious seed-structures of this kind must have their significance in the history of the plant; and on returning to England I looked a little further into the matter. To follow up this kind of inquiry, however, would carry me far beyond the limits prescribed for this note, and I have only treated it here in a tentative fashion. Different botanists of eminence have paid attention to this subject, amongst them Roxburgh, Thomson, and Miers (see Dr. T. Thomson in Journ. Linn. Soc. Bot., vol. ii., p. 47, 1858, and Mr. J. Miers in Trans. Linn. Soc. Bot., vol. i., 1880). It would appear that the seed-structure of Barringtonia is also found in Careya, a genus of the same Myrtaceous tribe, and in Garcinia and other genera of the Guttiferæ, as well as in other inland plants.

Mr. Miers, after reviewing the opinions of his predecessors, gives the results of his own investigations. The solid embryo found in Barringtonia and many other genera consists, he observes, (a) of an external portion, the “exorhiza,” which nourishes the germinating seed and then dies away; (b) of an internal portion, the “neorhiza,” which, growing at each end, forms the central portion of the stem and root; and (c) the “medullary sheath” of Mirbel, that lies between the two, and is composed of elementary vascular tissue, which ultimately gives origin to the wood, bark, and leaves of the stem and yields woody fibre to the root. The exorhizal portion in some cases, as in Barringtonia acutangula, splits into four parts during germination. Mr. Miers compares this seed-structure with that of Rhizophora, employing the same terms, “neorhiza” for the internal portion which forms the seedling, and “exorhiza” for the external portion which merely nourishes it. However, I may add that the exorhizal portion in Rhizophora, as shown in [Chapter XXX.], is now regarded as formed by the coalesced cotyledons, and is termed the “cotyledonary body”; so that by implication the corresponding part of a Barringtonia seed should be regarded from the same standpoint.

It may be apposite to notice here that Barringtonia racemosa displays one capacity which does not appear to belong to B. speciosa. The branches stuck in wet soil throw out roots and establish themselves. This capacity of vegetative reproduction is turned to account by the Fijians, who make “live-fences” of this tree in wet localities.

NOTE 51 (page [135])
On a Common Inland Species of Scævola in Vanua Levu, Fiji

This is a tall shrub, or small tree, nine or ten feet high, which corresponds with S. floribunda, Gray, as far as Seemann describes it. It has small, black, juicy drupes, well suited for dispersal by birds, having no “suberous” mesocarp as in the shore species (S. Kœnigii), and no capacity for dispersal by currents. It grows, much like the Hawaiian inland species, in exposed situations where there is plenty of light, as on mountain-peaks, at the borders of forests, in open-wooded districts, and in the plains, and is to be found at all elevations from near the sea up to the highest mountain summit (3,500 feet) when the station is suitable. I noticed it on the higher slopes and frequently on the tops of nearly all the principal mountains that I climbed. It is evident that birds carry the “stones” from one mountain-peak to another, and no doubt they explain the presence of the species in Tonga. Dr. Seemann speaks of it as a beach plant in Viti Levu. The plant familiar to me in Vanua Levu is only on very rare occasions to be seen as an intruder in the beach-flora.

NOTE 52 (page [137]).
On the Capacity for Dispersal by Currents of Colubrina oppositifolia, an Inland Hawaiian Tree

The seeds in my experiments sank within ten days; but they are not readily detached from the fruit, as in the case of the buoyant seeds of the littoral species (C. asiatica). The fruits, which may float for a week or two, break down, as Hillebrand observes, tardily and imperfectly, and could give but little assistance to dispersal by water.

NOTE 53 (page [141])
On the Genus Erythrina

We have in E. indica a widely distributed littoral species, ranging from India through Malaya to eastern Australia, and over nearly all the groups of the Pacific, reaching to Tahiti and the Marquesas, but not occurring in Hawaii. It is associated in Fiji and Tonga with another shore-species, E. ovalifolia, Roxb., found also in India and Malaya. I did not come on the second species in Fiji, and according to Seemann it is rare. It is possible that there is a genetic connection between the two; and it is noteworthy that in one case Seemann was uncertain (p. 426) whether the species was E. ovalifolia or only a variety of E. indica.

In different parts of their areas both these species may be found inland. This no doubt is to be connected with their occasional cultivation. The Polynesians who esteem E. indica for its handsome scarlet flowers and its scarlet seeds often plant it near their houses; but it is curious that if we look at the pages of Seemann, Horne, and one or two other botanical authors who have written on the Pacific, we find no reference to its littoral station, the first-named botanist merely characterising it in Fiji as occurring “wild or planted.”

However, in various localities in Fiji, as on the shores of Natewa Bay in Vanua Levu, Erythrina indica thrives as a characteristic beach tree. Dr. Reinecke speaks of it as widely spread on the Samoan coasts; and the French botanists refer to it as a tree of the Tahitian beaches. Prof. Schimper frequently mentions the two littoral species of Erythrina as amongst the components of the Malayan strand-flora. Dr. Treub, when he visited Krakatoa in 1886, three years after the eruption, noticed some young plants of Erythrina growing on the shore; whilst Prof. Penzig in 1897 found that both E. indica and E. ovalifolia had established themselves on the beach. Mr. Kurz again is quoted by Prof. Schimper (p. 170) as including E. indica amongst the “beach-jungle” of Pegu.

There is abundant evidence in support of the dispersal of the genus by currents. I have observed the seeds of Erythrina indica on the beaches of Keeling Atoll. Schimper noted Erythrina seeds amongst the stranded drift of the Java Sea. Treub remarked young plants of the genus growing on the shore of Krakatoa three years after the great eruption, and Penzig places Erythrina indica and E. ovalifolia amongst the beach-plants brought to Krakatoa through the agency of the currents. The seeds of E. indica not infrequently came under my observation stranded on the Fijian beaches and floating in the Rewa estuary; and in an experiment made in Fiji they still floated after five months in sea-water. Mr. Hemsley years ago formed the opinion, from the drift collections at Kew, that the genus was dispersed by the currents. I may here add in further illustration of this point that Erythrina seeds were found by me in South America floating in numbers in the Guayaquil estuary and stranded on the beaches of Ecuador.

It is noteworthy that, unlike some of the other shore-plants, Erythrina indica has at least three sets of names in the South Pacific. Thus it is known as Rara and Ndrala in Fiji, Ngatae in Samoa, Futuna, and Rarotonga, Atae in Tahiti, and Kenae in the Marquesas. The Samoan and Tahitian name recalls the Burmese name of Ka-thit, whilst the Marquesan word is suggestive of the Makassar name Kăne or Kanur. The Hawaiian name of E. monosperma is Wili-wili, which evidently has arisen from the screw-like movement of the open pod when thrown into the air. The same name in the form of Wiri-wiri is applied for a similar reason to Gyrocarpus Jacquini in Fiji. It is possible that the Polynesians have assisted the dispersal of the coast-species (E. indica); but the currents could have performed the distribution unaided, and the variety of aboriginal names is not in favour of human intervention.

With reference to the possible extermination by insects of Erythrina in Hawaii, it has been before remarked (p. [143]) that this would not account for the survival of an inland species, such as E. monosperma in Hawaii. However, this species since the occupation of that group by the white man is on the road to extinction. Dr. Hillebrand observes that the species was much more common formerly than in his time (1851-1871), a result evidently due to the ravages of the common tropical mealy bug, a pest of relatively modern introduction (see Koebele in Stubb’s Agricultural Report on Hawaii). It may be added here that Cordia subcordata, a littoral tree, had been almost exterminated by the ravages of a small moth even in Dr. Hillebrand’s time. During my examination of the coasts of the large island of Hawaii, in 1896-7, I was shown several places not long before occupied by this tree; and, as indicated in [Note 29], it only came under my notice in a few localities.

NOTE 54 (page [145])
On the Genus Canavalia

Of the three maritime species, C. obtusifolia, D.C., occurs on beaches all round the tropical zone. I was familiar with it on North Keeling Island in the Indian Ocean, in Fiji, and in Ecuador. C. ensiformis, D.C., is just as widely spread; but it is both inland and maritime in its station, and except when collecting it in the Solomon Islands I have had but little acquaintance with it. C. sericea (Gray) is a characteristic beach-plant in Fiji, but is infrequent. In Rarotonga, according to Cheeseman, it is a common littoral plant. It was also found in Tahiti by Banks and Solander, and is seemingly peculiar to the Pacific islands.

Besides C. ensiformis, the other two shore species may at times be found inland. Thus it is singular that the French botanists do not, as a rule, speak of C. sericea as a Tahitian beach plant; and Nadeaud only remarks, concerning its station, that it frequents the wooded slopes of the valleys of the interior. In North Keeling Island C. obtusifolia presented itself to me not only as a beach-creeper, its usual habit, but as a climber over the branches of the coast trees. In one locality in Vanua Levu I found a variety of this species growing on a hill a mile inland and about 700 feet above the sea. On one of the beaches it approached C. sericea in some of its characters, as in the form of the calyx and in the hairiness.

Although the seeds of C. obtusifolia have long been known to be dispersed by the currents, having been found in Moseley’s collections of floating drift off the New Guinea coast (Bot. Chall. Exp., IV, 291), they displayed remarkable fickleness when experimented on by me in Fiji. As a rule, however, about 10 per cent. sank at once in sea-water, 50 per cent. floated after three weeks, and 10 per cent. after twelve weeks. Of seeds that had been kept three years, 50 per cent. floated after eleven weeks. The seeds are to be found in numbers amongst the stranded drift of the Fijian and Ecuador beaches, and I noticed them also afloat in the Rewa estuary of Fiji.

I tested the floating-power of the seeds of C. sericea in Fiji, and found that half of them remained afloat after sixty days. On the seeds of C. ensiformis I have not experimented; but their buoyancy is indicated by the frequent occurrence of the plant on the Solomon Island coral islets (Guppy’s Solomon Islands, pp. 290, 292, 296), and probably the Canavalia seeds identified at Kew from my drift collections on these islets belong to this species. Schimper (p. 166) refers to the seeds of a Canavalia in Java that were still afloat after ten weeks. These littoral plants are indebted for the floating capacity of the seed to the buoyant kernel.

NOTE 55 (page [42] and [Note 20])
The Inland Extension of Scævola Kœnigii

Scævola sericea (Forst.), a hairy variety of this littoral plant, will probably prove in some localities to be the inland form of the species. Dr. Reinecke, who mentions only this variety for Samoa, says that it is found in very moist ground in river-ravines, and no other station is referred to. It would seem that both the glabrous and hairy forms occur in Hawaii. Dr. Seemann speaks of the hairy variety as littoral in Fiji.

NOTE 56 (page [149])
On the Capacity for Dispersal by Currents of Sophora tomentosa, S. chrysophylla, and S. tetraptera

(1) Sophora tomentosa, Linn.—The moniliform pods will float for few weeks, but it is to the seeds liberated by the breaking down of the pod that the wide dispersal of this beach-plant by the currents is due. When experimenting on the freshly obtained seeds in Fiji I found that four-fifths of them floated after three months in sea-water. With seeds that had been kept for three years, half floated after twelve months and retained their sound condition. The seeds owe their floating power to the buoyant kernel.

(2) Sophora chrysophylla, Seem.—The dry pods of this Hawaiian mountain species float between one and two weeks in sea-water, but being brittle they readily break down and the seeds escape. The seeds have no buoyancy even after drying for four years.

(3) Sophora tetraptera, Ait., from the coast of Chile.—After floating from ten to fourteen days in sea-water, the dry pods become sodden and begin to break up, the seeds escaping. Since, however, the pods tend to decay and break open on the tree they would not be available for dispersal by currents. Out of a number of freshly gathered seeds all floated buoyantly after a month in sea-water, when the experiment ended; and of seeds that had been kept over a year six out of ten floated after four months in sea-water, two of them germinating afterwards in soil. Like those of S. tomentosa the seeds possess buoyant kernels to which the floating power is due. On account of the hardness of the tests the seeds to ensure rapid germination require to be filed.

NOTE 57 (page [153])
On the Species of Ochrosia

Schumann distinguishes the following species:

(a) O. parviflora, Hensl., widely spread in the Pacific islands.

(b) O. compta, Schumann, confined to Hawaii and corresponding to var. B. of O. sandwicensis as given by Hillebrand.

(c) O. borbonica, Spr., synonym O. oppositifolia, Lam., from Mauritius and Madagascar to Java and Singapore.

(d)

Both probably varieties of O. borbonica.

O. sandwicensis, Gray, of Hawaii.
O. elliptica, Lab., of New Caledonia.

(e) O. parviflora, Schumann, of New Guinea, probably identical with O. mariannensis.

NOTE 58 (page [156])
On Pandanus (from Warburg)

(a) The size (length) of the drupes of endemic species in oceanic islands.—The drupes of P. reineckei of Samoa are 4-5 cm. (1³⁄₅-2 inches). Those of P. joskei and P. thurstonii in Fiji measure respectively 6 cm. (2²⁄₅ inch) and 2¹⁄₂ cm. (1 inch).

Out of about sixteen species in the Mascarene Islands (Mauritius, Réunion, and Rodriquez) quite half have drupes 2-3¹⁄₂ cm. (⁴⁄₅-1²⁄₅ inch) in size, whilst they run up to 8 or 10 cm. (3-4 inches), and may be less than a centimetre (²⁄₅ inch).

(b) The affinities of the Fijian and Samoan species.

P. odoratissimus	Wide-ranging	Section Keura.
P. joskei	Fiji	Section Lophostigma.
P. samoensis	Samoa	Section Lophostigma.
P. thurstonii	Fiji	Section Acrostigma.
P. reineckei	Samoa	Section Hombronia.

NOTE 59 (page [188])
Seeds in Petrels

Darwin, in his correspondence (1859) with Sir Joseph Hooker, refers to the occurrence of large West Indian seeds in the crops of some nestling petrels observed by Sir William Milner at St. Kilda (Life and Letters, II, 147, 148). Mr. Charles Dixon in Ibis (1885) refers to Sir W. Milner’s observation in the case of the Fulmar Petrel (Procellaria glacialis) and speaks of them as Brazilian seeds brought by the Gulf Stream, adding that he himself found a nut in the crop of one of these birds in the same locality. He supposes that the birds pick them up from the water. Mr. Hemsley very kindly wrote to Sir Joseph Hooker recently on this point with the object of obtaining some idea of the nature of the seeds; but after this lapse of time it has not been found possible to satisfy my curiosity. I live in the hope of their proving to be Cæsalpinia seeds.

NOTE 60 (page [202])
Schimper on the Halophilous Character of Littoral Leguminosæ and of Shore Plants generally

As a result of extensive microchemical investigations, this eminent German botanist arrived at the conclusion that plants living on the sea-shore, or in inland stations rich in chlorides, are able, as a rule, to store up in their tissues a large quantity of these salts, a capacity enabling them to live in localities where the subsoil is rich in these materials. This inference, as shown in his experiments, is just as applicable to the shore-plants of temperate regions, such as Aster tripolium, Crambe maritima, and Eryngium maritimum, as it is to such typical littoral plants of the tropics as Barringtonia speciosa, Ipomœa pes capræ, Scævola Koenigii, Tournefortia argentea, &c. However, with the Leguminosæ experimented upon, this capacity of storing up chlorides was often exhibited but slightly or not at all; and characteristic Pacific beach-plants, such as Canavalia turgida, Pongamia glabra, and Sophora tomentosa are especially cited as examples (Schimper’s Ind. Mal. Strand-flora, pp. 140-151; Wolff’s ash-analyses are here quoted).

NOTE 61 (page [215])
Meteorological Observations on the Summit of Mauna Loa

The summit is formed of bare rock and sand, the phanerogamic vegetation ceasing a couple of thousand feet below. Some low plant-forms doubtless occur under the moist, warm conditions near the steam-cracks, since Wilkes mentions his finding a small moss; but with this exception the surface may be described as sterile.

Dryness of the Air and Electrical Phenomena.—Wilkes refers to the association of these conditions more than once in his narrative. Whenever, as sometimes happened, the dew point could not be obtained with Pouillet’s hygrometer, electricity was easily excited, and was developed in large sparks. On taking off the clothes at night, sparks would appear. As shown in the table subjoined, electrical phenomena were noticed during the first few days of my sojourn on the summit when the relative humidity was very low. My red blanket at night crackled in my hands and emitted sparks, and a glowing line was produced by drawing the finger along. Whilst the air was in this condition I observed that the wings of dead butterflies lying on the ground stuck to my fingers tenaciously like a needle to a magnet. The adhesiveness disappeared when the excessive dryness gave place to humidity. The physiological effect on me of the associated dryness and electrical state of the air was displayed in a hot, dry, sweatless skin (cracking and chapping rapidly), severe headache and sore-throat, general lassitude, and great irritability. When the weather changed and the air became humid, these unpleasant symptoms quickly disappeared.

As a result of these dry conditions on the summit of Mauna Loa, decomposition does not occur. I found in one place on the top, on the site of an old camp, the remains of a quarter of beef, the meat fresh but dried up. From a water-bottle left behind by one of the party and subsequently restored to him, I learned that the visit had been made in the previous summer. This non-decomposition seems a little strange, since, as remarked below, flies and other insects were not infrequent on the summit. However, as Hann remarks, when speaking of mountain climates, everything dries much more quickly at great altitudes; animals that have been shot, or killed by falling, become mummies without undergoing decay (Schimper’s Plant-Geography, 697).... The scorching power of the sun in a sky usually cloudless, or nearly so, was a trying feature of my daily experiences; and I found that when I faced it with unshaded eyes during my walks I suffered from severe pain in the eyeballs at night.

Insects on the Summit.—It may seem a strange thing to relate, that in a region apparently absolutely sterile, the flies and other winged insects caused me much discomfort in my small tent when I was confined to it through illness. When lying down one morning I noticed the house-fly, the blue-bottle, and two or three other flies, small beetles not over a fifth of an inch in size, a moth, and a wasp. They were no doubt quite happy in the heat, as the temperature inside was over 80° F., and the sun’s rays felt almost scorching through the thin duck canvas. Butterflies (and occasionally large moths) were often observed flying in a drowsy condition about the summit and were easily caught. They were fond of fluttering around the steam-holes. In places, numbers were to be seen dead and dried up on the ground, the detached wings lying about. In the case of a recently dead butterfly I found its carcase already attacked by numerous small bugs. The butterflies were most frequent when there was a fresh southerly breeze, and were doubtless blown up the slopes from the forests below.

Whymper in his Travels amongst the Great Andes of the Equator gives many particulars of the occurrence of insects at great elevations. He noticed beetles, diptera, butterflies, moths, and several other insects at altitudes of 15,000 to 16,000 feet. At 16,500 feet he obtained a small bug of the genus Emesa. He quotes Humboldt and Bonpland as showing that insects are transported into the upper regions of the atmosphere 16,000 to 19,000 feet above the sea, and he remarks that the transportation of insects by ascending currents of air has occasionally been observed in operation. These facts bear directly on the dispersal of insects.

The Winds.—My tent, which was pitched near the middle of the western border of the crater, happened to be situated in the battle-ground of the northerly and southerly winds, in a region of gusty winds, fitful airs, and dead calms. The northerly winds were usually from N.-N.N.W. and the southerly winds from S.W.-S.S.W., easting in either case being rarely observed, the northerly winds rather prevailing at night. As a result of this location miniature whirlwinds were frequent in the vicinity of my tent, which carried sand into the air and more than once threatened to lift up my tent bodily and carry it off into the crater below. At the north end of the crater-border north-easterly winds prevailed, and at the south end southerly winds occasionally showing easting. When on one occasion I walked round the crater-margin, a fresh south-easterly wind prevailed at most parts of the circumference except in the vicinity of my camp, where there was a light S.S.W. wind both at 8 a.m. and 6 p.m. when I started and returned. The local character of the winds was often displayed in my walks. On one occasion, having left my camp, where a southerly wind was blowing, and walked half a mile to the north, I found a bitterly cold N.N.E. gale in my face which so impeded my progress that I returned to my camp where the same southerly breeze continued.

Commodore Wilkes was encamped on the east side of the crater, and there (December and January) he experienced strong south-west winds, on at least three days having the force of a gale. These are the prevailing winds in this season over the group; whereas in August, the time of my sojourn, south-westerly winds are quite out of season, this being in the midst of the period of the N.E. trades.

It will be gathered from the foregoing remarks that the mere record of the winds is insufficient for the purpose of obtaining any definite notion of the air-currents at this elevation (13,600 feet). It is to close observation of the clouds that we must look for data of importance.

The Clouds.—The clouds on the summit of Mauna Loa were an unending source of interest to me, and I will give briefly the results of my observations. The highest clouds were wispy cirri, often arranged as in a mackerel sky, and evidently at a great altitude. They were only observed on four or five days. (The lower clouds are indicated in the accompanying diagram.) Below them and at no great height above the mountain were to be not infrequently observed isolated woolly clouds that were carried in a few minutes across the sky and had a brief existence, often forming and melting away as one gazed at them. Next, there was a heavy bank of cumulus, which formed on the south-west slope near the top of the mountain, from which lines of cloud extended along each flank. Lowest of all was a broad belt, or rather a sea, of cumulus that was developed on both sides of the mountain about one-third way down its slopes, and during the day-time isolated the peak from the world below. It is with the last two cloud formations that we are most concerned, and I will first describe the sea of cumulus.

The sea of cumulus, as in the case of similar cloud-formations of most other isolated mountains, when viewed from above, as from the mountain-top, presents a cloud-field of dazzling whiteness, sparkling in the sun. Seen from below, as from the coast, it has the dark lowering appearance of the rain-cloud and indicates the rain-belt. Disappearing during the night, this broad belt begins to form again between 8 and 9 a.m., and by 10 or 11 a.m. the lower regions are completely hidden and the mountain’s summit, cut off from the world, rises above the level of the sea of clouds like an island in an Arctic ocean. As the day progresses the clouds become more compact and dense. The usual altitude of this broad belt of cloud is between 7,000 and 8,000 feet. This level is indicated by the burying of the Kohala mountains, which rise to a height of 5,500 feet in the distant north-west corner of the island, and by the usual emergence of the highest summit of Hualalai, which rises, still nearer, to an elevation of 8,275 feet. On some days, however, it attains a height of nearly 9,000 feet. On such occasions the highest peak of Hualalai kept reappearing and disappearing during the day, but the distant summit of Haleakala in East Maui, 10,032 feet in elevation and 80 miles away, was always visible.

Words fail to describe the magnificent aspect of this sea of cloud which shuts off the spectator from the world below. From the summit of the mountain he gazes down on its surface lit up by a sun shining in a typically cloudless sky. At one time it appears as an undulating Arctic land covered with snow of dazzling whiteness. At another time it looks like a hummocky frozen Polar sea sparkling in the sunshine. Through occasional rifts, however, one can discern a dark dismal region of mist and rain-cloud beneath. Miss Bird, who passed a night on the summit in June, 1874, well describes this sea of cloud in her book on the Sandwich Islands as “all radiance above and drizzling fog below.”

[To face page [585].

Diagram illustrating the prevailing cloud-formations of Mauna Loa during August, 1897.

The heavy bank of cumulus, that forms at noon on the south-west slope at an altitude of 10,000 to 13,000 feet above the sea, and sometimes rises above the mountain, is one of the most conspicuous of the cloud-phenomena on the summit of Mauna Loa. Apparently extending from it, but in reality moving towards it, are two lines of small cumuli that follow the same level along either flank above the sea of cumulus, as is indicated in the accompanying diagram. It was observed by Wilkes in mid-winter, 1840-41, but at a lower level. “Clouds would approach us (he writes) from the south-west when we had a strong north-east trade wind blowing, coming up with their cumulus front reaching the height of about 8,000 feet, spreading horizontally and then disappearing.” During my sojourn this bank formed a very striking feature in the landscape during the early afternoon. On two or three occasions when I visited the south side of the summit and descended for about a thousand feet I passed through this bank, being then exposed to a driving mist coming up the slopes from the south-west. Though its upper surface viewed from a distance is dazzling white, below it is dark and nimboid.

It is to an updraught of warm moist air on the south or south-west slopes of the mountain, and to the prevailing cool north-east trade that strikes the north side of the summit, that we must look for the explanation of the development and situation of this bank. Although the trade-wind is markedly stronger than the south-west updraught, some of the warm, moist, southerly air-currents find their way, as shown by the observations at my camp, along the sides of the summit, and a line of condensation is produced where they come into contact with the cool air of the north-east trade as it sweeps past the flanks of the mountain. Sometimes at my camp, when there was a light southerly breeze blowing, I have noticed the line of small cumuli moving south along the mountain side towards the bank of cumulus.... I may remark that on a few days a small bank of cumulus formed under similar conditions on the north-west side of the summit.

From my study of the clouds I arrived at the conclusion that there were three prevailing air-currents on the summit of Mauna Loa:

(1) The updraught of warm moist air on the south and south-west slopes of the mountain.

(2) The north-east trade wind, the upper limit of this air-current being probably not far above the summit.

(3) An upper air-current from the south-east (E.S.E.-S.S.E.), which, from the velocity of the clouds it carried, was often probably not over a couple of thousand feet above the summit. It may be observed that on the coast at the base of the southern slope of the mountain in the middle of September, when the wind was N.E. and carried the lower clouds with it, the upper clouds were, on several occasions, noticed travelling in the opposite direction, namely, from the south.

The volcano was quiescent during my visit and could have exercised but little influence on the air-currents.

The Shadow of the Mountain.—Every morning and evening, in clear weather, for about twenty minutes after sunrise and before sunset, the shadow of the mountain was thrown back against the sky of the opposite horizon. It seemed as if some Titanic brush, at work in the sky far away, had painted in the profile of the mountain with a very uncanny blue. At sunset the peak was the last to disappear. Commodore Wilkes, who only records it once, namely, at sunset on the 1st of January, describes it as “a beautiful appearance of the shadow of the mountain projected on the eastern sky ... as distinct as possible, its vast dome seemed to rest on the distant horizon.” This phenomenon is, of course, well known in the case of other isolated mountains. According to Murray’s Handbook of Southern Italy (1892), the correct thing for a visitor to Stromboli is to make an early ascent of the cone to observe “the very curious triangular shadow of the mountain cast by the rising sun upon the sea.” Unfortunately I neglected my opportunity when on the island. The shadow of the mountain is also one of the sights of Etna, a dark-violet, triangular shadow (Baedeker) being thrown at sunrise over the surface of West Sicily, that is, on the land. I saw the shadow but imperfectly outlined, as the weather was not favourable at the time of my ascent. When at Nicolosi, on the south slope of Etna, I noticed at sunset a faint shadow of the mountain thrown against the eastern sky. I gathered from a short conversation with Prof. Ricco, the director of the Catania Observatory, when I told him of the shadow of the Hawaiian mountain, that the interest lay in its projection against the sky. It is doubtless akin to the spectre of the Brocken and other mountain spectres.

Some Previous Meteorological Observations on Mauna Loa.—.... Mr. Douglas, the botanist, who was subsequently found dead in a cattle-pit on Mauna Kea, spent a day on the summit of Mauna Loa in the middle of January, 1834. He mentions that a little way below the top the thermometer fell at night to 19° F. The wind on the top was N.W. The air at 11.20 a.m. was 33°, the hygrometer registering 0·5. He remarks that the great dryness of the air was evident without the assistance of the hygrometer (Hawaiian Spectator, vols. I and II, 1838-9).

Commodore Wilkes, in vol. IV of his Narrative of the United States Exploring Expedition, gives the following observations on the temperature and winds on the top of Mauna Loa between Dec. 23, 1840, and Jan. 13, 1841. Those on the temperature are incomplete, but they give a fair idea of the prevailing conditions. The degrees are in Fahrenheit’s scale.

• Dec. 23, 1840: elevation, 13,190 feet; 3 p.m., 25° F.; strong S.W. gale; night temperature, 15°.
• Dec. 24, 1840: summit (13,600 feet); night minimum, 22°.
• Dec. 26, 1840: summit (13,600 feet); violent S.W. gale; night min., 17°.
• Dec. 27, 1840: summit (13,600 feet); sunrise temp., 20°; night min., 17°; wind, S.W.
• Dec. 29, 1840: summit (13,600 feet); noon temp. in shade, 47°; night min., 20°.
• Dec. 30, 1840: summit (13,600 feet); noon temp., 55°; night min., 13°.
• Dec. 31, 1840: summit (13,600 feet); night min., 17°.
• Jan. 2, 1841: summit (13,600 feet); sunrise, 20°; wind, N.E.
• Jan. 3, 1841: summit (13,600 feet); night min., 17°.
• Jan. 4, 1841: summit (13,600 feet); daylight, 20°.
• Jan. 8, 1841: summit (13,600 feet); S.W. gale.
• Jan. 10, 1841: summit (13,600 feet); night temp., 16°.
• Jan. 12, 1841: summit (13,600 feet); night temp., 17°.
• Jan. 13, 1841: summit (13,600 feet); strong S.W. wind.

The usual variation of temperature in the twenty-four hours is given as 17°-50°. The south-west was evidently regarded as the prevailing wind, and the clouds are spoken of as sometimes moving from opposite directions towards the same centre.

When Miss Bird spent a night on the summit of Mauna Loa during the eruption of June, 1874, the cold was described as intense, eleven degrees of frost (21° F.).

Observations on the Summit of Mauna Kea.—.... When Prof. Alexander with a party of scientists ascended this mountain (in the summer of 1892), the thermometer at night fell to 13° F., and the trade-wind was found to be blowing as strongly on the summit as down below (Whitney’s Tourist Guide to Hawaii). It is to be inferred that the party camped by the small lake which is a few hundred feet below the actual summit (13,800 feet). This lake, which I visited on May 20, 1897, is about 120 yards across, and evidently shallow, probably not more than three or four fathoms deep. A carpet of algæ covered the bottom. At noon, by the lake, the air in the shade was 53° F., whilst the temperature of the surface-water was 51°. The lower clouds were moving from S.S.E. This lake is said to be permanently frozen over in the winter, and to have been visited by skaters.

Permanent Water Supply on the Summit of Mauna Loa.—In this barren rocky region water derived from the winter-snow is to be found all the year through at the bottom of the deep cracks or fissures in the lava-rock. Such fissures are from two to four feet wide, and in the case of that near my tent the bucket had to be lowered to a depth of seventeen or eighteen feet to reach the water, or rather the ice, since it was often necessary to break the surface ice. In these deep, narrow fissures, which the sun scarcely penetrates, the water would probably be frozen over all through the seasons; but in those of less depth it would remain liquid in summer.

Register of Observations on Wind, Relative Humidity, Cloud, Rain, and Temperature, made by H. B. Guppy on the Summit of Mauna Loa at an Elevation of 13,500 Feet above the Sea, August 9th to 31st, 1897. (Camp about Middle of West Side of Crater Margin)

Date.	Observation.	12-4 A.M.	4-8 A.M.	8-12 A.M.					12-4 P.M.					4-8 P.M.					8-12 P.M.	Air in shade.					Remarks.
																				Min.			Max.	Range.
9	Wind	S.S.W-S.W. 2	Variable	...					W.S.W.-W.N.W. 3					N.N.W.-N. 1					N.N.W.-N. 3	F.			F.	F.	A beautifully coloured lunar halo at 1 A.M. Electrical condition of the atmosphere (see text).
	Rel. hum.	...	...	...					...					...					...	27·5			61·2	33·7
	Cloud	0	0	0					0					0					0
	Rain	0	0	0					0					0					0
10	Wind	...	N.N.W. 2	S.S.W.-W.S.W. 3					S.S.W.-W.S.W. 2					Calm					...						Electrical condition of the atmosphere. Faint lunar halo at 8 P.M.
	Rel. hum.	...	...	34					42·5					46					...	33·5			59·7	26·2
	Cloud	...	0	1					2					0					...
	Rain	0	0	0					0					0					0
11	Wind	Calm	W.S.W. 1	...					...					S.S.W.-S.W. 1					...						Electrical condition of the atmosphere.
	Rel. hum.	...	...			28·5			...					...					...			22·5	61·2	38·7
						21
	Cloud	0	0	0					1					0					...
	Rain	0	0	0					0					0					0
12	Wind	Calm	Calm	N.N.W.-N. 2-3					N.N.W. 1					N.N.W.-N. 2-4					N. 2-4, S.W.-W.S.W. 3-5						Earth tremors. Total rain, 10⁄100. At sunset, wind N.W., wet canvas of tent froze hard. At 10 P.M., strong southerly wind, canvas thawed, rain with strong gusty wind until 4 A.M., when wind less.
	Rel. hum.	...	...	45					79					89					...	23			54·7	31·7
	Cloud	0	0	5					10					10					5
	Rain	0	0	0					Rain					Rain					Rain
13	Wind	S.W. 4-6	S.W. 3-5	S.W.-S.S.E. 3-4					S.W. 3					Calm; N.W. 1					...						Earth tremors. Total rain 10⁄100.
	Rel. hum.	...	...	86					86					78					...			33·6	48·7	15·1
	Cloud	...	10	7					10					9					...
	Rain	Rain	Rain	0					Rain					Rain					0
14	Wind	N.N.W.-N. 1	Northerly, 3	N.N.W. 3					N.N.W. 2					N.N.W. 1					...						Butterflies flying about in a semi-torpid state, and easily caught with the hand.
	Rel. hum.	...	61·5	47					42					45·5					...			32·5	52·2	19·7
	Cloud	3	0	0					0-2					0					...
	Rain	0	0	0					0					0					0
15	Wind	Northerly, 1	W.S.W. 1, N.N.W. 1	S.W.-W. 2					Calm; Southerly, 1					Calm; N.N.W. 1					N.N.E.-1-2						Wind fitful during day; north-westerly and south-westerly airs with calms.
	Rel. hum.	...	...	38					44·5							62			...			28	54·7	26·7
																52·5
	Cloud	0	0	0					7					7-0					0
	Rain	0	0	0					0					0					0
16	Wind	Calm	N.N.W. 2	N. 3					...					Southerly, 1					...						Carefully observed the shadow of the mountain which, at sunrise and sunset, is projected against the opposite horizon.
	Rel. hum.	...	...			32			...					61					...			26	53.2	27.2
						28
	Cloud	...	0	0					7					0					...
	Rain	0	0	0					0					0					0
17	Wind	...	N.N.W. 1	N.N.W. 2, Southerly, 2					N.-N.N.E 3, Southerly, 2					N.N.W. 1					...						Fitful northerly and southerly winds causing miniature whirlwinds that carried dust and paper up into the air.
	Rel. hum.	...	...	32					32·5					...					...			20·5	58·7	38·2
	Cloud	...	0	0					1					0					...
	Rain	0	0	0					0					0					0
18	Wind	Calm	S.W. 1	Variable					...					S.W. 2					...						—
	Rel. hum.	...	...	26					...					47					...			23	58	35
	Cloud	0	0	0					1					0					...
	Rain	0	0	0					0					0					0
19	Wind	...	...	N.N.W. 1, W.S.W. 1					Northerly, 3, S.S.W. 3					Northerly, 2, Southerly, 2					Southerly, 3						Through the day, fitful northerly and southerly breezes.
	Rel. hum.	...	...	20					23							24			...			22	58·7	36·7
																35·5
	Cloud	...	...	0					0					0					0
	Rain	0	0	0					0					0					0
20	Wind	Southerly, 2	Northerly, 1, Southerly, 1	Northerly, 1, Southerly, 1					S.S.W.-S.W. 4					Northerly, 2, Southerly, 1					Southerly, 3						Fitful northerly and southerly airs, often reversing several times in a few minutes.
	Rel. hum.	...	...	26·5					25					29·5					...			22	57·2	35·2
	Cloud	0	0	0					0					0					0
	Rain	0	0	0					0					0					0
21	Wind	Southerly, 4	Southerly, 3-4	S.S.W. 3-4					S.S.W. 4-5					S.W. 4-5					S.W. 4						At camp, strong southerly winds all day. At 7 A.M., walked half-mile north and found a bitterly cold N.N.E. gale blowing there, which forced me to return to camp where the south wind still blew freshly.
	Rel. hum.	...	...	35·5					31					47					...			26·5	53·7	27·2
	Cloud	...	0	0					0					0					...
	Rain	0	0	0					0					0					0
22	Wind	...	S.S.W. 1	...					...					Calm, Southerly, 2					...						Walked round the crater from 8 A.M. to 6 P.M.
	Rel. hum.	...	...	58					...					...					...			20·5	46·7	26·2
	Cloud	...	0	0					1					0					...
	Rain	0	0	0					0					0					0
23	Wind	...	S.S.W. 2	...					N.N.W.-N.E. 1-3, Southerly, 3					Southerly, 1					Calm						A few drops of rain at 2 P.M.
	Rel. hum.	...	...			60			52·5					70					...			24	50·7	26·7
						54·5			64
	Cloud	...	0	3					3-8					3-6					0
	Rain	0	0	0					Rain					0					0
24	Wind	...	Northerly, 3	S.W. 1					N.N.W.-N. 2-3					N.N.W.-N. b. E. 2-3					Northerly 2
	Rel. hum.	...	...	54·5					...					53					...			20	52·7	32·7
	Cloud	...	0	0-1					4					0					...
	Rain	0	0	0					0					0					0
25	Wind	Calm	...	Calms with N.W. and S.W. airs					Calms with N.W. and S.W. airs					Northerly and Southerly airs					Northerly and Southerly airs						A few drops of rain at 3 P.M.
	Rel. hum.	...	...			42·5			73·5					61					...			17	52·2	35·2
						37·5			61·5
	Cloud	0	0	0-1					2-9					0					0
	Rain	0	0	0					Rain					0					0
26	Wind	Northerly, 2	N.N.W. 1	...					N.N.W.-N. 3					N.N.W. 2					...						Descended through the bank of cumulus on S.W. slope and found driving mist coming up the slope from S.W.
	Rel. hum.	...	...	49·5					...					64					...			19·5	53·7	34·2
	Cloud	0	...	3					3					0					0
	Rain	0	0	0					0					0					0
27	Wind	Calms with light airs	Calms with light airs	S.W.-W.S.W. 2					W.S.W.-W. 1					Southerly					Calm						Rain not measurable. Rain-clouds poured into and filled the huge crater.
	Rel. hum.	...	...	35					73·5					...					...			18·5	50·7	32·2
	Cloud	0	0	4					10					4					0
	Rain	0	0	0					Rain					0					0
28	Wind	Calms with northerly airs	Calms with variable airs	N.N.W.-N. 2-3					S.S.W. 3					Northerly, 3					Calms with variable airs						9 A.M., high stationary cirrus; at noon, solar halo; in afternoon, nimbus partly filling crater causing a rainbow there; a few drops of rain at 4 P.M.
	Rel. hum.	...	...			46·5			64·5					...					...			15	49·7	34·7
						59·0			73·0
	Cloud	0	0	5					7					0					0
	Rain	0	0	0					Rain					0					0
29	Wind	N.N.W.-N. 3	N.N.W.-N. 2	N.-N. b. E. 3					N.N.W. 3, S.S.W. 2					Calms with variable airs					Northerly, 1
	Rel. hum.	...	...	45					44					60·5					...			21·5	48·7	27·2
	Cloud	0	0	0					0					0					0
	Rain	0	0	0					0					0					0
30	Wind	Calms with northerly airs	Southerly, 1	N.N.W. 2-3					N.N.W. 3					Calms with variable airs					...
	Rel. hum.	...	...	32·5							41·0			...					...			18	50·7	32·7
											54·5
	Cloud	0	0	0					0					0					0
	Rain	0	0	0					0					0					0
31	...	...	...	...					...					...					...	18·5			50·0	31·5

Date.	Observation.	12-4 A.M.	4-8 A.M.	8-12 A.M.
9	Wind	S.S.W-S.W. 2	Variable	...
	Rel. hum.	...	...	...
	Cloud	0	0	0
	Rain	0	0	0
10	Wind	...	N.N.W. 2	S.S.W.-W.S.W. 3
	Rel. hum.	...	...	34
	Cloud	...	0	1
	Rain	0	0	0
11	Wind	Calm	W.S.W. 1	...
	Rel. hum.	...	...			28·5
						21
	Cloud	0	0	0
	Rain	0	0	0
12	Wind	Calm	Calm	N.N.W.-N. 2-3
	Rel. hum.	...	...	45
	Cloud	0	0	5
	Rain	0	0	0
13	Wind	S.W. 4-6	S.W. 3-5	S.W.-S.S.E. 3-4
	Rel. hum.	...	...	86
	Cloud	...	10	7
	Rain	Rain	Rain	0
14	Wind	N.N.W.-N. 1	Northerly, 3	N.N.W. 3
	Rel. hum.	...	61·5	47
	Cloud	3	0	0
	Rain	0	0	0
15	Wind	Northerly, 1	W.S.W. 1, N.N.W. 1	S.W.-W. 2
	Rel. hum.	...	...	38
	Cloud	0	0	0
	Rain	0	0	0
16	Wind	Calm	N.N.W. 2	N. 3
	Rel. hum.	...	...			32
						28
	Cloud	...	0	0
	Rain	0	0	0
17	Wind	...	N.N.W. 1	N.N.W. 2, Southerly, 2
	Rel. hum.	...	...	32
	Cloud	...	0	0
	Rain	0	0	0
18	Wind	Calm	S.W. 1	Variable
	Rel. hum.	...	...	26
	Cloud	0	0	0
	Rain	0	0	0
19	Wind	...	...	N.N.W. 1, W.S.W. 1
	Rel. hum.	...	...	20
	Cloud	...	...	0
	Rain	0	0	0
20	Wind	Southerly, 2	Northerly, 1, Southerly, 1	Northerly, 1, Southerly, 1
	Rel. hum.	...	...	26·5
	Cloud	0	0	0
	Rain	0	0	0
21	Wind	Southerly, 4	Southerly, 3-4	S.S.W. 3-4
	Rel. hum.	...	...	35·5
	Cloud	...	0	0
	Rain	0	0	0
22	Wind	...	S.S.W. 1	...
	Rel. hum.	...	...	58
	Cloud	...	0	0
	Rain	0	0	0
23	Wind	...	S.S.W. 2	...
	Rel. hum.	...	...			60
						54·5
	Cloud	...	0	3
	Rain	0	0	0
24	Wind	...	Northerly, 3	S.W. 1
	Rel. hum.	...	...	54·5
	Cloud	...	0	0-1
	Rain	0	0	0
25	Wind	Calm	...	Calms with N.W. and S.W. airs
	Rel. hum.	...	...			42·5
						37·5
	Cloud	0	0	0-1
	Rain	0	0	0
26	Wind	Northerly, 2	N.N.W. 1	...
	Rel. hum.	...	...	49·5
	Cloud	0	...	3
	Rain	0	0	0
27	Wind	Calms with light airs	Calms with light airs	S.W.-W.S.W. 2
	Rel. hum.	...	...	35
	Cloud	0	0	4
	Rain	0	0	0
28	Wind	Calms with northerly airs	Calms with variable airs	N.N.W.-N. 2-3
	Rel. hum.	...	...			46·5
						59·0
	Cloud	0	0	5
	Rain	0	0	0
29	Wind	N.N.W.-N. 3	N.N.W.-N. 2	N.-N. b. E. 3
	Rel. hum.	...	...	45
	Cloud	0	0	0
	Rain	0	0	0
30	Wind	Calms with northerly airs	Southerly, 1	N.N.W. 2-3
	Rel. hum.	...	...	32·5
	Cloud	0	0	0
	Rain	0	0	0
31	...	...	...	...

Date.	Observation.	12-4 P.M.					4-8 P.M.					8-12 P.M.
9	Wind	W.S.W.-W.N.W. 3					N.N.W.-N. 1					N.N.W.-N. 3
	Rel. hum.	...					...					...
	Cloud	0					0					0
	Rain	0					0					0
10	Wind	S.S.W.-W.S.W. 2					Calm					...
	Rel. hum.	42·5					46					...
	Cloud	2					0					...
	Rain	0					0					0
11	Wind	...					S.S.W.-S.W. 1					...
	Rel. hum.	...					...					...
	Cloud	1					0					...
	Rain	0					0					0
12	Wind	N.N.W. 1					N.N.W.-N. 2-4					N. 2-4, S.W.-W.S.W. 3-5
	Rel. hum.	79					89					...
	Cloud	10					10					5
	Rain	Rain					Rain					Rain
13	Wind	S.W. 3					Calm; N.W. 1					...
	Rel. hum.	86					78					...
	Cloud	10					9					...
	Rain	Rain					Rain					0
14	Wind	N.N.W. 2					N.N.W. 1					...
	Rel. hum.	42					45·5					...
	Cloud	0-2					0					...
	Rain	0					0					0
15	Wind	Calm; Southerly, 1					Calm; N.N.W. 1					N.N.E.-1-2
	Rel. hum.	44·5							62			...
									52·5
	Cloud	7					7-0					0
	Rain	0					0					0
16	Wind	...					Southerly, 1					...
	Rel. hum.	...					61					...
	Cloud	7					0					...
	Rain	0					0					0
17	Wind	N.-N.N.E 3, Southerly, 2					N.N.W. 1					...
	Rel. hum.	32·5					...					...
	Cloud	1					0					...
	Rain	0					0					0
18	Wind	...					S.W. 2					...
	Rel. hum.	...					47					...
	Cloud	1					0					...
	Rain	0					0					0
19	Wind	Northerly, 3, S.S.W. 3					Northerly, 2, Southerly, 2					Southerly, 3
	Rel. hum.	23							24			...
									35·5
	Cloud	0					0					0
	Rain	0					0					0
20	Wind	S.S.W.-S.W. 4					Northerly, 2, Southerly, 1					Southerly, 3
	Rel. hum.	25					29·5					...
	Cloud	0					0					0
	Rain	0					0					0
21	Wind	S.S.W. 4-5					S.W. 4-5					S.W. 4
	Rel. hum.	31					47					...
	Cloud	0					0					...
	Rain	0					0					0
22	Wind	...					Calm, Southerly, 2					...
	Rel. hum.	...					...					...
	Cloud	1					0					...
	Rain	0					0					0
23	Wind	N.N.W.-N.E. 1-3, Southerly, 3					Southerly, 1					Calm
	Rel. hum.	52·5					70					...
		64
	Cloud	3-8					3-6					0
	Rain	Rain					0					0
24	Wind	N.N.W.-N. 2-3					N.N.W.-N. b. E. 2-3					Northerly 2
	Rel. hum.	...					53					...
	Cloud	4					0					...
	Rain	0					0					0
25	Wind	Calms with N.W. and S.W. airs					Northerly and Southerly airs					Northerly and Southerly airs
	Rel. hum.	73·5					61					...
		61·5
	Cloud	2-9					0					0
	Rain	Rain					0					0
26	Wind	N.N.W.-N. 3					N.N.W. 2					...
	Rel. hum.	...					64					...
	Cloud	3					0					0
	Rain	0					0					0
27	Wind	W.S.W.-W. 1					Southerly					Calm
	Rel. hum.	73·5					...					...
	Cloud	10					4					0
	Rain	Rain					0					0
28	Wind	S.S.W. 3					Northerly, 3					Calms with variable airs
	Rel. hum.	64·5					...					...
		73·0
	Cloud	7					0					0
	Rain	Rain					0					0
29	Wind	N.N.W. 3, S.S.W. 2					Calms with variable airs					Northerly, 1
	Rel. hum.	44					60·5					...
	Cloud	0					0					0
	Rain	0					0					0
30	Wind	N.N.W. 3					Calms with variable airs					...
	Rel. hum.			41·0			...					...
				54·5
	Cloud	0					0					0
	Rain	0					0					0
31	...	...					...					...

Date.	Air in shade.			Remarks.
	Min.	Max.	Range.
	F.	F.	F.
9	27·5	61·2	33·7	A beautifully coloured lunar halo at 1 A.M. Electrical condition of the atmosphere (see text).
10	33·5	59·7	26·2	Electrical condition of the atmosphere. Faint lunar halo at 8 P.M.
11	22·5	61·2	38·7	Electrical condition of the atmosphere.
12	23	54·7	31·7	Earth tremors. Total rain, 10⁄100. At sunset, wind N.W., wet canvas of tent froze hard. At 10 P.M., strong southerly wind, canvas thawed, rain with strong gusty wind until 4 A.M., when wind less.
13	33·6	48·7	15·1	Earth tremors. Total rain 10⁄100.
14	32·5	52·2	19·7	Butterflies flying about in a semi-torpid state, and easily caught with the hand.
15	28	54·7	26·7	Wind fitful during day; north-westerly and south-westerly airs with calms.
16	26	53.2	27.2	Carefully observed the shadow of the mountain which, at sunrise and sunset, is projected against the opposite horizon.
17	20·5	58·7	38·2	Fitful northerly and southerly winds causing miniature whirlwinds that carried dust and paper up into the air.
18	23	58	35	—
19	22	58·7	36·7	Through the day, fitful northerly and southerly breezes.
20	22	57·2	35·2	Fitful northerly and southerly airs, often reversing several times in a few minutes.
21	26·5	53·7	27·2	At camp, strong southerly winds all day. At 7 A.M., walked half-mile north and found a bitterly cold N.N.E. gale blowing there, which forced me to return to camp where the south wind still blew freshly.
22	20·5	46·7	26·2	Walked round the crater from 8 A.M. to 6 P.M.
23	24	50·7	26·7	A few drops of rain at 2 P.M.
24	20	52·7	32·7
25	17	52·2	35·2	A few drops of rain at 3 P.M.
26	19·5	53·7	34·2	Descended through the bank of cumulus on S.W. slope and found driving mist coming up the slope from S.W.
27	18·5	50·7	32·2	Rain not measurable. Rain-clouds poured into and filled the huge crater.
28	15	49·7	34·7	9 A.M., high stationary cirrus; at noon, solar halo; in afternoon, nimbus partly filling crater causing a rainbow there; a few drops of rain at 4 P.M.
29	21·5	48·7	27·2
30	18	50·7	32·7
31	18·5	50·0	31·5

Method of Observation employed by the Author on the Summit of Mauna Loa.—My camp was placed near the middle of the west margin of the crater about 13,500 feet above the sea. The instruments employed were a Sixe’s maximum and minimum thermometer made by Negretti and Zambra, several unmounted thermometers, and a reference thermometer (with a Kew certificate) by the above-named makers, which was used as a standard. The freezing point was also tested for all the instruments on the summit in melting powdered ice. The maximum air observations and those on the relative humidity were taken in a small cave with a hole in the roof, through which there was a steady flow of air. One day was occupied in comparing the cave-observations with those obtained under a temporary screen rigged up outside my tent, the only difference shown being as a rule less than a degree. The minimum observations taken in my tent, where there was no artificial heat, were usually only 1·5° higher than those given by a thermometer outside the tent.

Results of the Observations on the Top of Mauna Loa, Aug. 9-31, 1897

Mean minimum temperature of air in shade	23·2° F.
Mean maximum temperature of air in shade	53·8°
Mean daily range of temperature	30·6°
Lowest reading	15·0°
Highest reading	61·2°
Mean temperature for the period	38·5°

Mean relative humidity, 8-9 a.m., 44·5 %

Mean relative humidity, noon ... 43 %

Mean relative humidity, 5-6 p.m., 56 %

Many observations included which are not given in the register.

On Aug. 11th, at 10 a.m., wet bulb, 33·2°; dry bulb, 52°; difference, 18·8°.

On Aug. 19th, at 11 a.m., wet bulb, 35·7°; dry bulb, 56°; difference, 20·3°.

Owing to the varying winds at my camp, the relative humidity fluctuated greatly in a short time. Thus, on Aug. 12 it was 46% at noon, and 79% at 2 p.m.

Average Cloudiness (10 indicating a Sky completely Overcast)

12-4 A.M.	0	Cloudless during 12 out of 13 days
4-8 A.M.	0	Cloudless during 19 out of 20 days
8-12 A.M.	1·3	Cloudless during 13 out of 22 days
12-4 P.M.	3·5	Cloudless during 6 out of 22 days
4-8 P.M.	1·5	Cloudless during 17 out of 22 days
8-12 P.M.	0	Cloudless during 11 out of 12 days

The winds at the camp were extremely variable and local from north and south, usually light, with force 1-3: see under Winds and Clouds in the text.

Rain fell on six days, total ³⁰⁄₁₀₀ of an inch: but on four of the days it was not measurable.

NOTE 62 (page [222])
On the Relative Proportion of Vascular Cryptogams in Fiji

According to Seemann’s work, where about 617 indigenous flowering plants and about 195 ferns and lycopods are enumerated, the vascular cryptogams would form about 24 per cent. of the whole flora. (All weeds and cultivated plants are here excluded.) The vascular cryptogams, however, seem to figure too prominently in Seemann’s collections. From Horne’s data, who says that he added 363 flowering plants to the flora, the flowering plants would amount to about 980; and since Baker implies, in Trimen’s Journal of Botany, 1879, that Horne added 42 species of ferns and lycopods to the flora, this would increase the vascular cryptogams to 237, which enables us to estimate the relative proportion of vascular cryptogams in Fiji as about 20 per cent. of the whole flora of vascular plants. This is probably near the truth.

NOTE 63 (page [222])
On the Table of Vascular Cryptogams of Tahiti, Hawaii, and Fiji

In the case of Tahiti, I have gone carefully through the list given by Drake del Castillo, comparing it with other Polynesian lists given by Seemann, Horne, Hillebrand, Hemsley, &c., and have reduced his endemic species from 19 to 13. The same thing has been done with Hillebrand’s list for Hawaii, some of his species having been found in other parts of Polynesia, thus reducing the endemic species from 75 to 70. The data relating to Fiji are referred to in [Note 62].

NOTE 64 (page [223])
On the Distribution of the Tahitian Ferns and Lycopods

I have arranged them as follows, according to the distributions given by Drake del Castillo:—Cosmopolitan, 5; Tropics of Old and New Worlds, 33; Tropics of Old World, mainly Indo-Malaya, 58; “Océanie,” including Australia, 17; Polynesia, 26; South America, 2; peculiar to Tahiti, 13: total, 154.

Out of 141 non-endemic Tahitian species, 107 at least have been recorded from the Fijian area comprising Samoa and Tonga, and 42 from Hawaii. Of the last, all but four occur also in Fiji. There is thus a very small element peculiar to Hawaii and Tahiti alone. Some of them will no doubt be found in the Fijian area; whilst two of them, Acrostichum squamosum and Lycopodium venustulum, are high-mountain forms in Hawaii and Tahiti, which have evidently failed to find a suitable elevation in Fiji.

NOTE 65 (page [225])
Distribution of some of the Mountain Ferns of Hawaii that are not found either in Fiji or Tahiti (mainly from Hillebrand)

Species.	Altitude of station in feet.	General distribution.
Schizæa robusta, Bak.	3,000-6,000	Perhaps a form of S. australis, Gaud., from the Falkland and Auckland Islands.
Polypodium serrulatum, Mett.	3,000-6,000	Generally diffused in the tropics and subtropics.
Aspidium caryotideum, Wall.	In the forests	Himalayas, South Africa, &c.
Aspidium filix mas, Sw.	In the highlands	Over four continents, from the arctic circle to the higher levels of tropical mountains.
Asplenium trichomanes, L.	5,000-8,000	Temperate zones and the higher levels of tropical mountains.
Asplenium monanthemum, L.	3,000-6,000	American Andes, Madeira, Tristan d’Acunha, Azores, Abyssinia, &c.
Asplenium fragile, Presl.	4,000-6,000	Andes.
Asplenium contiguum, K.	2,000-5,000	Lord Howe Island, Ceylon, Neilgherry Hills.
Asplenium adiantum nigrum, L.	4,000-10,000	Europe, Asia, Africa, Atlantic Islands.
Asplenium aspidioides, Sch.	1,000-6,000	Andes, Africa, India.

NOTE 66 (page [226])
Endemic Genera of Ferns in Hawaii

Hillebrand gives two genera of ferns peculiar to Hawaii, one, Sadleria of Kaulfuss, “scarcely distinct from Blechnum,” and containing four species; the other, Schizostege, constituted by himself, and represented by a single species found in only one or two of the islands.

NOTE 67 (page [241])
On the Dispersal of Compositæ by Birds

The goldfinch’s habit of pecking at the heads of thistles, and pulling out the achenes in bundles, is well known. Gätke mentions two suggestive instances of birds feeding on the fruits of a Composite plant. According to this observer, the Scarlet Grosbeak (Pyrrhula erythrina), when it alights on Heligoland, always feeds on the achenes of Sonchus oleraceus, which it picks off the plant; whilst the Parrot Crossbill (Loxia sp.), feeds in Heligoland on burrs and thistles (Heligoland as an Ornithological Observatory, pp. 407, 409). See [Note 91].

NOTE 68 (page [264])
On some of the Hawaiian Endemic Genera, excluding those of the Compositæ and Lobeliaceæ

Haplostachys, Phyllostegia, and Stenogyne, all Labiate Genera.—Phyllostegia is not strictly peculiar to Hawaii, since out of the 17 species enumerated in the Index Kewensis, 15 are Hawaiian, 1 Tahitian, and 1 is accredited to Unalaska (one of the Aleutian Islands). The last locality appears to be an error. The species in question is P. microphylla, Benth.; and on looking up the original authority in Linnæa (vi. 570, 1831), I find the locality is thus given: “insula coralligena Romanzoffii,” which is either one of the atolls of the Paumotu Islands in about lat. 15° S. and long. 144° W., or a coral island of the Marshall Group, most probably the former.... I paid some attention to the suitability of the fruits of these three Labiate genera for dispersal by frugivorous birds, for which the fleshy nucules in the cases of Phyllostegia and Stenogyne apparently fit them. Out of the fruits of five species of Phyllostegia examined by me, the seed-coverings in three species, after the removal of the fleshy covering of the nucule, were too soft for the protection of the seed in a bird’s stomach. Hillebrand also observes (p. 347) that the nucules when dried are wrinkled, and absorb moisture easily, a quality which, if true of all the species, would make the distribution of the genus by birds impossible. However, in two species I found the seed-coverings somewhat harder. It would seem that since birds have largely ceased to disperse these plants, the soft-skinned nucules would in the absence of their selective agency more frequently characterise the genus. It is possible that the dry nucules of Haplostachys, which according to Hillebrand are not affected by drying, represent the original condition of those of Phyllostegia, and that the fleshy character has been acquired in this archipelago. It will be seen in the list on page [263], that Haplostachys is regarded by Gray as a section of Phyllostegia. The remarks under Phyllostegia, regarding the softness of the seed-coverings beneath the fleshy coat of the nucule, also apply to Stenogyne; and Hillebrand, in contrasting its fleshy nucules with the dry nucules of Haplostachys, implies that they absorb water, which, I may remark, would render them quite unfit for dispersal by frugivorous birds.

Touchardia (Urticaceæ).—According to Hillebrand, the solitary species is by no means common in the group now. In 1897 I found it growing abundantly some miles up the Waipio gorge, Hawaii.

Cheirodendron (Araliaceæ).—C. Gaudichaudii, the well-known “Olapa” tree, is common in the forests of all the Hawaiian Islands between 2,000 and 5,000 feet; but I noticed it occasionally at greater elevations, as on the south-east slopes of Mauna Kea, where it extends to 7,000 feet. As described on page [343], the “Olapa” often grows in close contact with the Lehua (Metrosideros polymorpha), the two trunks appearing as one. The drupes would attract frugivorous birds and the pyrenes are well adapted for this mode of dispersal. Mr. Perkins states that the drupes are much sought after by the various species of Phæornis, a genus of birds peculiar to Hawaii.

Deterioration of Fruits for Purposes of Dispersal.—Among fruits or endemic genera that have evidently deteriorated in the Hawaiian Group as far as fitness for dispersal is concerned, may be mentioned, in addition to those of Phyllostegia and Stenogyne above noticed, those of the Araliaceous genera, Pterotropia and Triplasandra, and the Amarantaceous Nototrichium. The pyrenes of the first two genera on account of their thin covering, and the seed of the last-named genus on account of its thin testa, seem ill-fitted now for transport in a bird’s stomach, yet we cannot doubt that their ancestors originally arrived in this fashion. The same principle is also illustrated by some Hawaiian non-endemic genera of later eras that possess peculiar species, such, for instance, as in the case of Elæocarpus discussed in [Chapter XXVI.]

NOTE 69 (page [366])
On the Germination of Cuscuta

My observations were made on the Hawaiian endemic species (C. sandwichiana) and on a Fijian introduced species. Germination occurs readily in fresh water, the floating seedling growing rapidly. When the germinating seed is placed on wet soil in the shade, the seedling grows at the rate of ³⁄₄ inch (19 mm.) a day. The store of nutriment contained in the swollen radicular end will support the seedling for a couple of days, and if it has not then found a host it withers and dies. At first lying prone the seedling then lifts its upper end into the air, and it was almost pathetic to notice it moving round and round, endeavouring vainly to find some object near. The seedlings make no effort to strike into the soil, and when they are allowed to attach themselves to a plant they ascend rapidly, growing at the upper end and dying at the lower end.

NOTE 70 (pages [477], [480-1])
On Beach-Temperature

My data are rather scanty; but, judging from observations made in Hawaii, in South America, and in the south of England, the following scale would probably be true of typical beaches where the sand is found relatively cool and moist at a depth of four or five inches. This moisture seems to arise entirely from subsoil drainage seaward. When a beach fronts an arid, rainless region, few if any plants grow on it; the sand is loose, hot, and dry at the depth indicated; and the temperature of the surface half-inch rises to between 130° and 140° F., whilst four inches down it is 95° to 100°. Salt-marshes situated behind a beach even in a desert-region change its thermal behaviour, and it would then be more like a beach skirting a vegetated sea-border in the same latitude. The method of observation was as follows:—An unmounted thermometer of the size of a clinical thermometer, but graduated higher, was placed horizontally in the sand half an inch below the surface and a reading taken. It was then pushed vertically into the sand until the bulb was four inches deep and another reading taken. Provided that the sand is moist beneath, the colour does not seem to make much difference, except perhaps in very dark sands, none of which were tested.

Ordinary Beach-Temperatures with an Unclouded Sky in the Hot Season

during the Early Afternoon.

	Surface half-inch.	Four inches deep.
Temperate latitudes about 50-55°	100-105° F.	77° F.
Sub-tropical latitudes about 30-35	105-110	80
Tropical latitudes about 10-20	110-120	85

This illustrates only the average condition. On a calm day in the case of a beach facing south in the South of England, I have obtained exactly the same readings in July as at Valparaiso in January, 112° at surface, 80° four inches deep.

NOTE 71 (page [479])
On the Buoyancy of the Seeds or Seed-vessels of some Chilian Shore Plants

(1) Nolana, probably paradoxa. Common on the beaches of Southern Chile. The ripe drupes have a somewhat fleshy outer covering which they lose when lying on the sand, and present themselves then as dark-brown angular “stones,” often five to six millimetres across. Inside the outer hard covering of the stone is a layer of spongy tissue which gives it buoyancy; but since these coverings are wanting at the scars marking the basal insertion of the drupe, the embryo seems insufficiently protected against injury during flotation in sea-water; and the seed-vessel at first appears to be only fitted for conveyance by the currents over a limited tract of sea. However, in a preliminary experiment on seed-vessels that had been kept a few weeks, I found that 30 per cent. floated after three weeks in sea-water. Subsequently, after drying for a year, the seed-vessels were again tested in sea-water, nearly all of them floating after three months’ immersion. Two of them, removed after six weeks’ flotation, germinated healthily. These fruits are common in beach-drift between Corral and Valparaiso.

(2) Raphanus, near R. maritimus. Growing near beaches in South Chile, and not infrequently represented in the stranded beach-drift by the pods, which in my experiments floated seven to ten days in sea-water, after drying some weeks.

(3) Franseria. A species common on the beaches of Valparaiso and Talcahuano. Its prickly fruits, after being kept six weeks, floated only two to four days. They are well suited for transport in birds’ plumage.

NOTE 72 (page [483])
The Southern Limit of the Mangrove Formation in Ecuador.

... The southern limit of the mangrove formation on the west coast of South America is usually placed at 4° S. lat.; but it is probable that the vicinity of Tumbez in lat. 3° 30ʹ S. would be more correct. Baron von Eggers would place it rather further to the north-east, near the frontier of Ecuador and Peru in lat. 3° 20ʹ S. I spent eight days in the locality last named and saw no evidence of the beginning of the mangrove-formation.

NOTE 73 (page [495])
Additional Note on the Temperature of the Dry Coast of Ecuador between Puna Island and the Equator.

... Baron von Eggers gives the mean annual temperature for El Recreo, about half a degree south of the equator, at 75° F., which is near that of Rio de Janeiro in lat 23° S. on the east coast of the continent. Mr. F. P. Walker has kindly given me the results of temperature-observations covering a period of ten years, taken in the room for testing cables at Santa Elena Point (2° 10ʹ S.), usually about 6·30 a.m. The range of the monthly means was 71° F. (August) to 79·1° (March), and the mean for the year was 74·8°. In that locality a typical daily range would be 65° to 80°; and Mr. Walker believes that a minimum of 59° has been recorded.

NOTE 74 (page [495])
Observations on the Temperature of the Humboldt Current from Antofagasta Northward, between January and March, 1904 (Fahrenheit scale)

The observations were usually taken at the anchorages, but in some places, as at Ancon and Puerto Bolivar, they were taken from a boat outside the roadstead.

If we wish to ascertain how the Humboldt Current retains its cool temperature as it advances through the tropics to the equator, a glance at the following table will show that the surface-temperatures can aid us but slightly, since they do not vary in accordance with the latitude, a subject further discussed below. We can, however, obtain some valuable indications from the deeper temperatures. Let us take for instance the plane of 60°. Whilst south of Ancon (lat. 11° 45ʹ S.) it was rarely deeper than four fathoms, north of this latitude it descends rapidly, being probably about ten fathoms down at Salaverri and Eten and about twenty fathoms deep at Payta, in latitude 5° S., where the Humboldt Current leaves the coast. Within the Gulf of Guayaquil it is probable that the plane of 60° would descend to nearer thirty fathoms, the region being outside the influence of the current.

Some interesting facts are also elicited from the variation of the surface-temperatures. When we were coasting along at a distance of five or six miles from shore the readings were fairly constant from hour to hour varying only a degree or so. But nearer the land, for instance, about two or three miles away, the variation from hour to hour amounted to two or three degrees, whilst within the limits of the anchorages, a mile and less from the coast, the change from hour to hour amounted to three or four degrees. Nor was there any uniformity at the same hour over the surface of a roadstead. The temperature would often rise or fall a degree every few boat-lengths. Sometimes the inshore water was the coolest and sometimes it was the warmest. Thus at Iquique the inshore water was three degrees warmer than the water half a mile out, whilst at Mollendo, when the temperature one-third of a mile off the shore was 70°, it was 63° close to the rocky coast. The same thing was exhibited at Pisagua, where the surface-water two miles out at sea was 61°, whilst close inshore at the anchorage it was 58°. It was evident that there was a considerable intermingling of the warmer surface and the colder, deeper waters on the coasts of Chile and Peru. This was particularly noticeable on a rocky, steep-to coast, or where there was an uneven bottom. At some places, indeed, the warm upper layer did not exist, the cold water welling up all along the coast. This was especially the case between the 22nd and 19th parallels of latitude, a tract of coast in which lie Tocopilla, Iquique, and Pisagua, and probably the coolest part of the sea-border at this season of the year.

During a fortnight spent at Ancon (11° 45ʹ S.), between January 27 and February 10, I paid considerable attention to the local climatic conditions, and especially to the temperature of the inshore water. The daily range of the air-temperature was only five or six degrees, the average minimum and maximum being 71° and 75·9°, and the mean for the period 73·5°. The mean temperature of the surface-water at the head of the pier, from observations taken at about 7 a.m. and 4 p.m., was 68·6°, or five degrees cooler than the air, the mean temperature in the morning being 69·1° and in the afternoon 68°.

Observations on the Temperature of the Humboldt or Peruvian Current

(Made by H. B. Guppy, January to March, 1904.

Those at Panama are added for the sake of comparison)

Locality.	Depth (fathoms).	Distance from shore (miles).	S. lat.			Date.			Hour.	Depths in fathoms: temperature in Fahrenheit degrees.
										Surface.	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.	12.	13.	14.	15.	16.	17.	18.	19.	20.	21.	22.
Antofagasta	22	2⁄3	23°40ʹ					Jan. 12	5 p.m.	71°	70°	...	...	...	...	...	61°	...	...	57°	...	...	...	...	...	56°	...	...	...	...	...	...
								Jan. 13	6 a.m.	70	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	55	...	...
Tocopilla	18	1⁄2	22 0			Jan. 14			8 a.m.	57	...	...	...	...	...	...	...	...	...	56	...	...	...	...	...	55	...	...	...	...	...	...
Iquique	9	2⁄3	20 15			Jan. 15			5.30a.m.	59	...	...	...	...	58	...	...	...	55	...	...	...	...	...	...	...	...	...	...	...	...	...
Pisagua	10	1⁄2	19 30			Jan. 16			8 a.m.	57·5	...	...	...	...	...	...	...	...	...	56	...	...	...	...	...	...	...	...	...	...	...	...
Arica	...	1⁄4	18 25			Jan. 16			7 p.m.	66	...	...	...	...	...	...	...	57	...	...	...	...	...	...	...	...	...	...	...	...	...	...
Mollendo	22	1⁄3			17 0	Jan. 17			8 a.m.	61·5	...	...	...	...	...	...	...	57	...	...	...	...	...	...	...	...	...	...	...	...	...	...
					17 0	Jan. 17			6 p.m.	65	...	...	...	...	...	...	...	...	...	59	...	...	...	...	...	...	...	...	...	...	...	...
					17 0	Jan. 18			9.30a.m.	62	...	...	...	...	57·5	...	...	...	...	57	...	...	...	...	56	...	...	...	...	...	...	55
Callao	5	1	12 3			Jan. 20			6 a.m.	60·5	...	...	...	...	58·5	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
Ancon	17	13⁄4	11 45			Jan. 29			11 a.m.	66	...	...	6·15	...	58·5	...	...	57	...	57	...	...	...	...	...	...	56·5	...	...	...	...	...
Salaverri	6	3⁄4	8 15			Febr. 21			6.30p.m.	65	...	...	...	...	...	62	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
Eten	7	1⁄2	7 0			Febr. 22			6 a.m.	66·5	...	...	...	...	...	...	63	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
Payta	16	11⁄2	5 0			Febr. 23			9 a.m.	70	...	...	...	...	...	...	67·5	...	...	...	...	...	...	...	62·5	...	...	...	...	...	...	...
Puerto Bolivar	10	1	3 10			Mar. 7			noon	78	...	...	...	...	71·5	...	...	...	...	70	...	...	...	...	...	...	...	...	...	...	...	...
Panama	8	3			(8 50)	Mar. 23			4 p.m.	82	...	...	...	...	...	...	79·5	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
					( N.)	Mar. 24			7 a.m.	79·5	...	...	...	...	...	...	...	79·5	...	...	...	...	...	...	...	...	...	...	...	...	...	...

The Ancon climate at this period is full of oddities and abnormalities, and in this way typifies much of the coast of Peru. Thus, since the heat of the day is tempered by the cool south-westerly winds which die away in the evening and give place usually to warm, light, northerly and north-westerly breezes, there is, as above remarked, but a small difference between day and night temperatures. The coldest time of the twenty-four hours is not in the early morning but at sunset. The sea off the beach is, on the average, much cooler than the air, which is not a normal state of things; and again, the water is often two or three degrees colder in the evening than it is in the morning, which is very unusual. Though the sea-border is practically a desert for the greater part of the year and has no rain, it is frequently enveloped in drizzling fogs or “garuas.” Judged from a European standard, things go by contraries on the coast of Peru; and this is entirely the effect of the Humboldt Current.

The temperature of the inshore waters of Ancon Bay varied considerably during the twenty-four hours. During the day, with the prevailing southerly wind, the cool waters of the current had free access to the bay, and swept around its border in their course north; but in the night, when northerly breezes occurred, the cold waters of the current were pushed off the coast and their place taken by the warmer inshore waters from the north; and this sometimes continued for a day or two. When the current again got mastery and its clean, cool waters filled the bay, the temperature of the water dropped suddenly five or six degrees, and the bay was filled with fish. At such times men in boats leave the beach, and in a few minutes, with hand-nets and baskets, they obtain thousands of the small fry. Other men, fishing with lines from the pier-head, seem ill-contented unless they can catch fish of the size of small mackerel at the rate of one a minute.

There can be little doubt that on the coasts of Chile and Peru the instincts of fish often lead them astray, on account of the sudden changes of temperature arising from the conflict between the warmer waters of the open sea and the cooler waters of the current. From the preceding remarks it will be inferred that sometimes the current is pushed off the coast for a while and its place taken by the warm waters from the north. At other times it dives down, so to speak, and flows at a deeper level, and warmer waters prevail both out at sea and inshore. At other times again, and this must be most disconcerting to the fish, the cold current suddenly appearing at the coast predominates at the surface for days together, and we have stretches of coast which, although lying within tropical latitudes, are washed by waters having the temperature of the temperate zone. It is to such causes that we must attribute the reckless habits of fish on these coasts. They are known to throw themselves on the beaches in thousands, where by their decay they taint the air long afterwards. Mr. Anderson Smith in his recent book on Temperate Chile vividly describes what goes on on such occasions at the port of Valdivia. At times the scene must be indeed a strange one, since huge octopi are rolled up on the beaches in numbers, and are regarded by the indigenes as deliberately seeking their death. Whether they commit suicide or not, “their beaks that blacken the edge of the sea-wash in places” afford a melancholy proof that their instinct has blundered.

The Mode of Observation.—A thermometer made on the Sixe pattern which I used several years ago for taking the bottom-temperatures of rivers, was employed for the deeper temperatures, and at critical depths the observations were always repeated. This instrument was compared after each set of observations with an ordinary thermometer graduated on the stem, which was compared with my standard thermometer provided with a Kew certificate.... The observations in the Panama Roadstead have been added for the sake of contrast.

NOTE 75 (page [496])
On the Stranded Massive Corals apparently of the Genus Porites found on the Coast of Peru and North Chile, at Arica (18° 25ʹ S.), Callao (12° 3ʹ S.), and Ancon (11° 45ʹ S.)

At Arica they occurred on the beach only. At Callao they also extended inland on the low spit at Punta for about 100 yards. At Ancon they were found not only on the beach but also twenty or thirty paces inland on the low adjoining plains. Their size varied from three inches to three feet. They were all more or less rounded by wave action, and were extensively burrowed by boring molluscs. Whilst some on the beach still displayed the dried-up soft parts of the boring mollusc, others inland were falling to pieces and undergoing chemical change. There was nothing to indicate that the corals were recently alive; and at Ancon they appeared to have been torn off a rocky spit of andesite that had become exposed on the beach during a recent movement of emergence, of which there is other evidence on this coast. Further particulars are given on page [496].

NOTE 76 (page [429])
Stranded Pumice on English and Scandinavian Beaches

Sernander, in his description of the Atlantic drift of the Scandinavian coast, refers to the occurrence of a small amount of true pumice. I have found solitary fragments of acid pumice well rounded by wave-action at Croyde Bay on the north coast of Devonshire, at the mouth of Salcombe Harbour on the south coast of the same county, and at Maenporth, near Falmouth, in Cornwall. Steamer slag, in some cases rudely simulating pumice, is common on all the South of England beaches I have examined. It is also common on the Scandinavian coasts, though seemingly regarded by Helge Bäckström, who is quoted by Sernander, as derived from the factories on the east coast of England. (See on these subjects a paper by Helge Bäckström, “Über angeschwemmte Bimsteine und Schlacken der nordeuropäischen Küsten”; Bihang till K. Sv. V. A. Handl. Bd. 16. Afd. 3, 1890; also a letter in Nature, about 1886, by H. B. Guppy.)

NOTE 77 (page [21])
On the Mode of Dispersal of Kleinhovia hospita

This small tree has a very wide distribution in the tropics, ranging from East Africa and the Mascarene Islands through India, South-eastern Asia, Malaya, New Guinea, and the Solomon Islands to Fiji and Tahiti. It is a plant that grows in inland open woods as well as amongst the littoral trees on the beach; and it is always doubtful (in Malaya, Fiji, and Samoa) whether to regard it as a shore plant or as an inland plant, different authors varying on this point. In Vanua Levu I formed the opinion that it is only an intruder amongst the littoral vegetation. In accounting for its distribution we have to choose between man, the bird, and the current. Though it may sometimes be noticed in native plantations, as I observed in the Solomon Islands, the tree has no special use; and the Solomon Island natives themselves indicated to me that the parrots that fed on the fruits of the tree aided in distributing the plant. The buoyant behaviour of the seeds, which are freed by the dehiscence of the bladder-capsules on the tree, is not constant. Whilst in the case of the seeds of littoral trees in Fiji I found that 30 per cent. floated after ten weeks, Prof. Schimper ascertained in the case (seemingly) of Malayan seeds that they sank at once. The seed-structure connected with the buoyancy is, as shown on page [105], accidental in character, and reference is made on page [20] to other plants of doubtful littoral reputation, in which the buoyant qualities are variable. The occasional buoyancy of its seeds will only, as I think, explain its occasional station at the coast; and I agree with Prof. Schimper (p. 156) when he attributes its wide distribution to birds, the seeds being hard, crustaceous, and about three millimetres across.

NOTE 78 (page [436])
On the “Sea”: an Unidentified Wild Fruit-tree in Fiji

This is a fair-sized forest tree common in places in the lower forests. I have never been able to identify it; but a “putamen” which was sent to the Kew Museum was named Spondias with a query. It is to be hoped its true botanical name will be discovered by one of my successors. Seemann places it amongst the “desiderata” concerning which further information is needed. The fruit is a drupe 2 to 2¹⁄₂ inches long possessing a pleasant fruity odour and inclosing a hard two-celled stone about 1²⁄₃ inch long, one cell containing a large fleshy seed covered with tawny hair, the other filled with the hair only and containing no seed. The Fijians say that these fruits, large as they are, are swallowed by the fruit-pigeons, the stones being found in their gullet. The leaves are distichous, alternate, lanceolate, eight or nine inches long, glabrous and dark green above, and covered below with a whitish woolly matted tomentum. The empty stones are not uncommon in the stranded beach-drift.

NOTE 79 (page [395])
On Willow-leaved River-side Plants

A number of observers, beginning with Humboldt, in his Ansichten der Nature, and including Seemann, L. H. Grindon, Ridley, Beccari, and others, have referred to what is called “stenophyllism” in plants. These willow-leaved river-side plants are found all over the globe, such plants usually growing close to the water’s edge in situations where they are liable to be more or less submerged when the river is in flood. Seemann, Beccari, and Ridley mention more than two dozen genera belonging to a great variety of orders, and including Acalypha, Antidesma, Calophyllum, Eulalia, Eugenia, Fagræa, Ficus, Garcinia, Ixora, Lindenia, Melastoma, Podocarpus, Psychotria, &c., all tropical, and represented either in Fiji, Borneo, or in the Malay Peninsula; whilst my readers will recall amongst temperate floras river-side plants of the genera Epilobium, Lythrum, Salix, &c., possessing the same form of leaf and the same station. The genus Eugenia comes under this category in Fiji, Borneo, and the Malay Peninsula, with reference to one or more of the species. In Fiji, species belonging to the genera Lindenia and Dolicholobium especially attracted my attention in this respect. It is noteworthy that several of the Bornean plants and some of the Fijian plants here concerned are endemic. Just as I have remarked in the question of the buoyancy of seeds and fruits, that not all water-side plants have buoyant seeds or fruits, but that nearly all plants thus endowed are found at the water-side, so we may say of the willow-leaved plants, that not all river-side plants have the willow-form of leaf, but that plants thus characterised gather at the river-side. Beccari and Ridley regard this willow-form of leaf as the result of adaptation. Seemann remarks that we have here the old question whether the webbed feet of a duck are the cause or the effect of the bird’s swimming; and I take the same position. (See Seemann’s Flora Vitiensis; Ridley in Trans. Linn. Soc. Bot., vol. iii. 1888-94; and Beccari’s Nelle Foreste di Borneo, 1902, or the English edition of 1904.)

NOTE 80 (pages [255], [504])
Mr. Perkins on the Hawaiian Lobeliaceæ (Fauna hawaiiensis, vol. I.)

My view, that the early Hawaiian Lobeliaceæ acquired the monstrous form of their flowers in the humid forests of a later age, is supported by the observations of Mr. Perkins on the connection between the highly-specialised nectar-eating Drepanids of Hawaii and the highly-specialised flowers of the Tree-Lobelias, a subject further discussed in [Chapter XXXIII.] This naturalist ascertained, in the case of one of the trees, that fertilisation could only be effected by these birds. So close is the biological connection between the Drepanid and the Tree-Lobelia, that Mr. Perkins finds here in part the cause of the development of the most remarkable forms of the birds. The botanist, also, would not dissociate the plants from this conclusion. There would be every reason to look for abnormal growth in birds and plants when the bird depends on the flower for its food, and the flower is dependent on the bird for its pollenisation. It is through such guises that the zoologist and the botanist have to penetrate when establishing the systematic affinity.

NOTE 81
On the Vertical Range of some of the most Typical and most Conspicuous of the Plants in the Forests on the Hamakua Slopes of Mauna Kea, Hawaii

During a descent of this mountain on its north side to near Ookala, the conditions were unusually favourable for recording the range of altitude for some of the plants easily recognisable.

Acacia koa began at 6,700 feet, and extended down to 2,300 feet.

Rubus (“akala”) began at 6,500 feet, and extended down to 2,500 feet.

Cheirodendron (“olapa”) began at 6,400 feet, and extended down to 2,200 feet.

Cyanea, a lobeliad growing on trunks of tree-ferns, began at 4,000 feet, and extended down to 2,300 feet.

Freycinetia began at 3,850 feet, and extended down to 2,000 feet.

Asplenium nidus began at 2,800 feet, and extended down to 2,200 feet.

Aleurites moluccana began at 1,800 feet, and extended down to 50 feet.

Metrosideros polymorpha, ranging through all the zones.

NOTE 82 (page [416])
Aboriginal Weeds[^[7]]

(Found by Captain Cook’s Botanists, Banks, Solander, the Forsters, Nelson, &c.,

in the Pacific Islands, 1768-80)

	Locality given by Cook’s botanists.			General distribution.
Cardamine sarmentosa	Tahiti			Polynesia. Introduced into Peru.
Sida microphylla	Tonga, New Hebrides			Old World tropics.
Sida rhombifolia	New Hebrides. H.			Tropics of Old and New World.
Urena lobata	Tahiti			Tropics of Old and New World.
Waltheria americana			Tahiti. H.			Tropics of Old and New World.
			New Caledonia
Oxalis corniculata	Tahiti			Old and New World.
Cardiospermum halicacabum	Tahiti. H.			Tropics of Old and New World.
Desmodium polycarpum	Tahiti			Old World.
Phaseolus truxillensis	Tahiti. H.			Tropics of Old and New World.
Lablab vulgaris	Tahiti			Old World tropics.
Abrus precatorius	Tahiti			Tropics of Old and New World.
Cassia sophora	Tonga			Tropics of Old and New World.
Hydrocotyle asiatica	New Caledonia			Tropics of Old and New World.
Oldenlandia tenuifolia	New Hebrides			Old World?
Oldenlandia paniculata	Tonga			Old World tropics.
Geophila reniformis	Tahiti			Tropics of Old and New World.
Ageratum conyzoides	New Hebrides. H.			Tropics of Old and New World.
Adenostemma viscosum	Tahiti. H.			Tropics of Old and New World.
Eclipta alba	New Hebrides			Tropics of Old and New World.
Siegesbeckia orientalis	Tahiti			Tropics of Old and New World.
Bidens pilosa	Tonga			Tropics of Old and New World.
Dichrocephala latifolia	Tahiti, Tonga, New Hebrides			Old World tropics.
Sonchus asper	Tonga, New Zealand			Old World.
Ipomœa insularis	New Hebrides, Tonga, Hawaii			Australia and Polynesia.
Ipomœa bona-nox	New Hebrides, Tonga, Tahiti. H.			Tropics of Old and New World.
Solanum nigrum, var. oleraceum.	Tahiti. H.			Old and New World.
Physalis angulata	Tahiti			Tropics of Old and New World.
Vandellia crustacea	Tahiti			Tropics of Old and New World.
Leucas decemdentata	Tahiti			Old World tropics.
Teucrium inflatum	Tonga			New World tropics.
Amarantus melancholicus, var. tricolor	New Hebrides, Tahiti			Old World.
Euxolus caudatus	Tonga, Tahiti			Old World tropics.
Achyranthes aspera	Tahiti			Old and New World.
Cyathula prostrata	Tahiti			Old World tropics.
Fleurya interrupta	Tahiti, Tonga			Old World tropics.
Commelina pacifica	Tonga, New Caledonia. H.			Tropics of Old and New World.
Eleusine indica	Tahiti. H.			Tropics of Old and New World.

[7]. Seemann is the principal authority, the results of his examination of the old collections being given in his Flora Vitiensis. Species regarded by Hillebrand as indigenous in Hawaii or as existing in that group at the time of its discovery by Cook are indicated by H in the second column.

NOTES 83-89 omitted

NOTE 90 (page [29])
On the Buoyancy of the Seeds of Euphorbia amygdaloides and E. segetalis

The seeds of both species have no proper buoyancy, and display no structure in their testas suggesting it; though, through the shrinking of the nucleus, a temporary floating power may be acquired with less mature or imperfect seeds. They support the general principle indicated for the British species on page [29].

NOTE 91
Mr. E. Kay Robinson on the Dispersal of Aster tripolium

According to this naturalist, the seeds of this plant are eaten in winter by snow-buntings on the English east coast. In reply to my query he tells me that the “draggled fluff still containing seeds” might easily adhere to birds (The Country-Side, Sept. 30, 1905).