54 Transactions.—Miscellaneous. 
latitude. The upper and main division would thence have found a nearly 
connected chain of islands, as the Ladrones, Carolina, and Radach Chain, 
etc., and might thus ultimately have attained even the extreme distance of 
the Sandwich Islands. In the same way those of the smaller branch, by 
the river of Canton, might have reached Luzon and the other Philippine 
Islands, and, possibly, by the same Caroline Islands have passed on to the 
more Southern as well as Easterly groups of Polynesia, such as the Fiji, 
Tonga, New Zealand, Society, and Paumatoa groups. Of course, this view 
partakes altogether of the nature of a guess; but, so far’ as we know at 
present, I do not think there is any thing in it unreasonable. 
The second main wave of emigration Eastwards, or rather South East- 
wards, I suppose to have passed from Central Asia by the lines of the great 
rivers Brahmapootra, Irawaddy and Menam, thus impinging on the ocean 
at the South-east end of the Bay of Bengal and Gulf of Siam. These 
emigrants would, thence, naturally spread themselves in the direction of 
Tenasserim, the Malay Peninsula, Sumatra, Java, Borneo, etc., thus form- 
ing the ancestors of the present Malay races, though, as it seems most 
probable to me, at a period long antecedent to the 2,000 years, to which we 
are able, historically, to trace up the existing Malays. Now, if this theory 
of two or more lines of emigration has any consistency, and it be true that 
both Malays and Polynesians did come—it matters not how many years ago 
—from one original Asiatic source, some certain forms of speech, once 
correct in their ancestral homes, would be preserved by each wave as 
portions of a common heritage. The occurrence of similar grammatical 
forms, though perhaps few in number, would prove contact, if not relation- 
ship, at some period or other, while the absence of a large vocabulary of 
similar words would prove, also, a long and entire separation. We see the 
reverse of this in cases where the vocabulary is rich, but the grammatical 
words few or none. Thus, modern Turkish and Persian are loaded with 
Arabic words, but the one has not altered its Tatar or the other its Indo- 
European Grammar ; on the other hand, France, which we know was once 
wholly Celtic, at the present moment, though still largely Celtic in race, has, 
with the exception of a few names of places, retained not one Celtic word in 
its spoken language. I am further induced to think that this view is con- 
firmed by even the little we know of the Fiji and Tonga dialects, for which 
their grammar is sufficiently cognate with those of the other islands for 
their people to be generally included under the generic head of Polynesia. 
There are a great many words not Polynesian, and other words Polynesian 
originally but now altered (like tambu for tapu) to suit their organs of 
i) Digna ee 
