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.