But for all that it is not easy to give _Observations of a Naturalist in the Pacific Between 1896 and 1899, Volume 2 Plant-Dispersal

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.)

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

Column headings:

A: Indigenous.

B: By aborigines in ancient times.

C: By Europeans soon after discovery.

D: Old World.

E: New Worlds.

F: Both Worlds.

G: Large.

H: Small.

I: Float for months.

J: Sink at once or in a few days.

K: Attract frugivorous birds.

+---------------------------+---+---+---+---+---+---+---------------------+

| +---+---+---+---+---+---+-------+---------+---+

| | |duced. | | | | [6] | | |

| | +---+---+ | | +---+---+-----+---+---+

| | A | B | C | D | E | F | G | H | I | J | K |

+---------------------------+---+---+---+---+---+---+---+---+-----+---+---+

| 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 | + |...|...| + |...|...|...| + | + |...| + |

+---------------------------+---+---+---+---+---+---+---+---+-----+---+---+

Footnote 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.

Footnote 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.

+-----------------------------+-------------+-------------+-------------+

+-----------------------------+-------------+-------------+-------------+

+-----------------------------+-------------+-------------+-------------+

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.

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

Column headings:

#: Number.

%: Per cent.

A: Percentage of non-adaptive species.

B: Percentage of adaptive species.

+------------------+-----------------------------+---------------+--------+

| +-------------------+---------+ Proportion of | Total |

| +---------+---------+---------+ and adaptive | of |

| +----+----+----+----+----+----+-------+-------+ with. |

| | # | % | # | % | # | % | A | B | |

+------------------+----+----+----+----+----+----+-------+-------+--------+

|Pacific Islands | 27 | 40 | 10 | 15 | 30 | 45 | 55 | 45 | 67 |

| | | | | | | | | | |

|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._

[Illustration: _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_){O. sandwicensis, Gray, of Hawaii. } Both probably varieties of

{O. elliptica, Lab., of New Caledonia.} O. borbonica.

(_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. (1-3/5 - 2

inches). Those of P. joskei and P. thurstonii in Fiji measure

respectively 6 cm. (2-2/5 inch) and 2-1/2 cm. (1 inch).

Out of about sixteen species in the Mascarene Islands (Mauritius,

Réunion, and Rodriquez) quite half have drupes 2 - 3-1/2 cm.

(4/5 - 1-2/5 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._

[Illustration: 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.

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)

Observations of a Naturalist in the Pacific Between 1896 and 1899, Volume 2 Plant-Dispersal · 第 45 / 47 章

But for all that it is not easy to give a definite explanation even from