ON THE LEA VALLEY. 15
tions, changes such as have been so amply demonstrated there
come to be looked upon as mere abstract possibilities which may
safely be disregarded in practice.
APPENDIX I.
Lake-Dwellings in the Lea Marshes.
I have mentioned that a close association between the
existence of dug-out canoes and that of primitive lake-dwellings
has been observed. Of course there is no necessary connection
between them, and in all probability there was never anything
that could be called a lake in the Lea Valley from Broxbourne
southward. But there were probably on the Lea marshes, in
primitive times, houses of the kind mentioned by Munro in The
Lake Dwellings of Europe, p. 553. He says :—
"A reference to Lake Dwellings occurs in a passage by
Hippocrates (De Eribus, &c. XXXVII):—'Concerning the
people of the Phasis, that region is marshy and hot, and full of
water and woody; . . . . The inhabitants live in the
marshes and have houses of timber and of reeds constructed in
the midst of the waters; and they seldom go out to the city or
the market, but sail up and down in boats made out of a solid
tree trunk, for there are numerous canals in that region' [This
locality is east of the Black Sea.]"
The sections displayed in the reservoir excavations show
frequent changes in the main channel of the Lea, and suggest the
existence of many backwaters, or partly deserted channels, in
which water-loving plants grew, leaving beds more or less peaty
for our inspection to-day. The Lea Marshes would therefore
afford many positions either wholly surrounded by water or easily
made insular, as sites for pile dwellings. As when (for example)
we traverse the hilly parts of the Chalk, and seldom, if ever, find
a naturally strong site unoccupied by an ancient camp of refuge,
so we may fairly conclude that eligible sites on marshes for pile-
dwellings would not, in primitive times, remain unused. To
come to a later period than that of dug-out canoes, we find the
Lea shown, in Norden's Map of Essex (1594) as having two or
more channels of apparently equal importance all the way from
Broxbourne to Blackwall, the other rivers marked on the map
having but one channel, except that the Thames, as at the
present day, has many around Canvey Island.
It seems therefore highly probable that the pre-Roman users
of dug-out canoes on the Lea lived in pile-dwellings on the