NOVEMBER 27, 1913] 
almost behind them, whereas man could only see the 
reflection when flying more or less towards it. 
In following these routes birds may also be able to 
travel with less exertion. When the sun is shining, 
land is warmer than water; the reverse is the case at 
night. This difference of temperature causes a down- 
ward air current over water on sunny days, and in 
calm weather this is most markedly felt even when 
pe over as small a piece of water as the Fleet 
ond near Farnborough. This pondis about 1000 yards 
long and about 7oo yards wide at its widest point, 
and is very shallow. Mr. O’Gorman tells me that 
on a sunny day a balloon drifted slowly over the pond 
and at once began to fall with considerable rapidity 
through a distance of perhaps 2000 feet, and ballast 
had to be thrown out to preyént it reaching the water. 
Aéroplanes are sensitive to’ the down current over 
quite small ponds on sunny days, and drop in passing 
over them. We thus have direct evidence of a down- 
ward current over a small sheet of water on a sunny 
day, and this must mean an updraught over the land 
near the water. 
Coast-lines are often marked out by cumulus clouds 
during the daytime. Dr. Shaw tells me that this is 
an indication of local rising air currents, and that 
the bases of clouds of this type are assigned to a 
height of from 4000 to sooo ft. 
If birds make use of these upcurrents they should 
‘fly over the land near the water in the daytime, and 
if there is a wind they should fly on the windward 
side of a river. Observations on this point would be 
of great interest. In windy weather the upcurrent 
would be much reduced, and perhaps would be in- 
appreciable. 
_ At night we should expect the opposite effect to be 
produced, but I know of no evidence on this point, 
and the upcurrent over water may be inappreciable. 
To take advantage of it, if it exists, birds should fly 
over water at night, or if there is a wind on the lee 
side of a river. 
' An on-shore wind striking against the cliffs pro- 
“duces an upcurrent, and this also birds would find 
advantageous. : 
There may be other advantages in valley routes, 
‘such as perhaps better conditions with regard to 
wind, and Dr. Gadow has pointed out to me that 
_ many of the birds that follow coast-lines and rivers 
‘are aquatic or semi-aquatic, and that even the more 
terrestrial birds will find Letter stores of food in river 
valleys than along the bordering hill ranges. 
' The foregoing throws very little light on the difficult 
ems involved in the migration of birds. It is 
oped, however, that other and more important obser- 
vations will be made from aéroplanes and airships, 
and that these will enable us to understand a little 
more about the mystery of the migration of birds. 
‘ Horace Darwin. 
November, 1913. 
tr The Elephant Trench at Dewlish. 
In the hope of finding an explanation as to the 
origin of the so-called elephant trench at Dewlish, 
Mr. Clement Reid (Nature, vol. xcii., p. 96), asks 
if. under desert conditions, there is any tendency for 
winds to cut trenches with rounded blind ends in 
soft limestone deposits. Having travelled in the 
Egyptian and other deserts, and having camped for 
some months on soft Tertiary limestones in the stormy 
region at the mouth of the Gulf of Suez, a few 
_remarks from my pen may be of interest in this 
connection. 
The only desert locality where I have seen trenches 
at all resembling that at Dewlish is in the Jemsa area 
near the mouth of the Gulf of Suez. On the low 
NO. 2300, VOL. 92] 
NATURE 
371 
flat isthmus which joins the headland of Ras Jemsa to 
the mainland there are cracks or openings in the soft 
“Raised Beach” deposits which cover the Tertiary 
Gypseous limestone formation of this area. The 
cracks are usually directed N.E.-S.W., and are 
parallel to the slip-planes which have disturbed the 
underlying deposits. The prevailing wind, which is 
strong and persistent, blows from the N.W. off the 
plain behind the Gebel Zeit range. Part of the sand 
and dust which it carries is dropped when its velocity 
decreases, namely in any hollow or wind-shadow that 
may occur. The cracks above-mentioned, which may 
be likened to crevasses in glaciers, form one of the 
receptacles for this wind-borne material, and being 
thus partially filled and obliterated, are not easily 
observed, and men and camels have been known to 
flounder into them. The fact that they are only 
partially filled shows that they are in process of 
formation now, and their origin would seem to be 
due to the solution of both series of deposits along 
such lines of weakness as joints or slip-planes. 
The infilling of hollows is typical of the desert and 
we know that an artificially excavated hole is not 
deepened, but, on the contrary, tends to be obliterated 
by wind-driven sand. 
If a rock of uniform texture but containing hard 
nodules, such as flints, is abraded by sand, the sur- 
face is fluted, and small hollows are scoured round 
the nodule. Mr. Reid’s letter does not suggest the 
presence of such hollows, and their absence must be 
regarded as another point against the wind-erosion 
theory. 
From the description of the trench, I gather that 
it occurs, not on the edge of a plateau, but on the 
surface of the open downs, and thus resembles, as 
Mr. Reid suggests, the well-known swallow-holes of 
the Great Scar Limestone, which frequently engulf 
sheep and other denizens of the plateaux. Mr. Reid 
writes in the singular number, as if only one end 
of the trench was rounded. This again is a feature 
common to swallow-holes, where the detritus carried 
by the disappearing stream abrades only the upstream 
end of the hole. Is this rounded end so situated with 
regard to the surrounding chal'k topography that it 
would be possible for a stream of water to have 
entered the trench from that end? Mr. Reid tells us 
that open trenches in the chalk are unknown else- 
where; are not the deep and narrow holes in the 
chalk, now filled with red clay, which are to be seen 
in the railway cuttings between Cambridge and Lon- 
don, supposed to have the same origin as the sinks 
of the Yorkshire wolds? 
As cracks and joints in the chalk are conspicuous 
by their absence, perhaps the point or line of weak- 
ness, which originally determined the position of the 
trench, has been eroded away completely. The ques- 
tion arises—Is there any relation between the direction 
of the trench and the direction of the joints in the 
country rock? 
The nature of the bottom of the trench and the 
relation of the trench to the surrounding topography 
should tell us something definite as to its possible 
origin, but in the meantime I think we may regard 
the wind-erosion hypothesis as untenable. 
H. T. Ferrar. 
Survey Department of Egypt, November 5. 
On a Habitat of a Marine Ameba. 
As our knowledge of marine Amcebe is very scanty 
it is worth while recording what appears to be a 
common habitat of one of these animals. 
At various times from May to October this year 
Amcebz were observed casually in the water obtained 
: by squeezing out the contents of the gastral cavity 
