the river Djemnil. An arm of the sea more accurately describes this
stream, which is (or was at the time of which I write) over three
hundred yards across. Here we had some difficulty with the Khivan,
who was for encamping till morning. I, however, strongly objected to
sleeping _a la belle étoile_, especially as the sky had now clouded
over, and it was beginning to snow. Partly by conciliation, partly
by threats, we at last persuaded him to make the attempt, following
closely in his wake. It was nasty work. Twice our horses were carried
off their feet by the strong current running out to sea (we were
only a quarter of a mile from the mouth); and once we, or rather the
horses, had to swim for it; but we reached the opposite shore in under
half an hour, wet and numbed to the waist, but safe. At seven we were
snugly housed for the night at Katvesera, a so-called village of three
or four mud hovels, selecting the best (outwardly) for our night's
lodging. We were badly received by the natives. Neither money nor
threats would induce them to produce provisions of any kind, so we
fell back on sticks of chocolate and Valentine's meat-juice. The
latter I never travel without--it is invaluable in uncivilized and
desert countries.

The inhabitants of Katvesera are under a score in number, and live
chiefly on fish, though I noticed in the morning that a considerable
quantity of land was under cultivation--apparently rice and barley.
They were a sullen, sulky lot, and we had almost to take the hut
by force. The Khivan, Gerôme, and myself took it in turns to watch
through the night. It was near here that the Italian was assassinated.

A start was made at daybreak. The weather had now changed. A cutting
north-easter was blowing, accompanied with snow and sleet. We forded,
about 11 a.m., the Kokajeri river, a mountain stream about thirty