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

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

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

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

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

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

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

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

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

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

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

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

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