630 DR. TEMPEST ANDERSON ON THE [XoY. I 9 IO, 
Dana 1 to be even less, sometimes not even more than 1°, an obser¬ 
vation which I can personally confirm, in the case of the lavas of 
Kilauea. It is to be remarked that at Matavanu the lava-stream 
is highest in the centre over the line of the tunnel, and tends to 
become more so owing to frequent small flows of lava which, 
whenever the flow beneath is obstructed from any cause, rise to 
the surface and then spread out and solidify there. This class of 
flow is more common as the sea is approached, so that the surface 
nearer the sea is often higher than that inland, and the direction of 
the flows in the lower part is thus often away from the sea, and in 
the reverse direction to that of the flow underneath by which they 
are supplied. That the surface of the lava over the tunnels is 
higher than elsewhere is a matter of observation, and that the 
difference tends to increase is vouched for by careful observers like 
Mr. Williams and Capt. Allen. They think that this increase of 
height is also largely due to lava intruding from the tunnel into 
the surrounding lava and forming sills and dykes; and, although 
this is mainly a matter of inference, I see no reason to question 
its accuracy. On my first attempt to cross the lava and visit the 
outflow into the sea, our way was stopped by a fairly large flow of 
lava only a day or two old. It had originated in an opening of 
the lava-crust- near the sea, and spread out inland as mentioned 
above. My men tried to cross it, and pronounced it impracticable ; a 
dog tried it, and retreated howling. I found a hummock of cool lava 
near it, on which I established myself and took photographs of the 
steam-explosions and the intervening lava (PI. L, fig. 2). This led 
me to notice particularly a crack in the hummock where the surface 
had risen apparently by the lateral compression of the surface-crust. 
When I made another attempt two days later, I succeeded in getting 
to the same spot, but a new lava-flow had covered a great part 
of the previous one and advanced even nearer the hummock. 
I noticed that the surface on which the new flow rested was pressed 
down and lowered for a distance of some feet in front of its edge; 
while the hummock itself was pressed up, and the crack materially 
widened. This shows that the lava below the surface was stilt 
plastic and capable of being squeezed into new positions, and the 
acuteness of the deformation or bending of the crust showed its 
small thickness. It is clear that a considerable amount of intru¬ 
sive action might take place in such a mass, without leaving much 
subsequently recognizable alteration. On the lower part of the 
lava-fields there are also several considerable mounds (or perhaps 
they might even be called plateaux) formed of large angular blocks 
of broken compact lava. The mounds were flat-topped, and the 
blocks presented no trace of igneous action since they had been 
broken. I could not at the time think of any satisfactory ex¬ 
planation of their formation, but now believe that they have been 
probably formed by intrusion of lava under the previously con¬ 
solidated crust, which was broken into pieces as it was pushed up 
(PI. XLVIII, fig. 2 ). 
1 J. D. Dana, ‘ Characteristics of Volcanoes’ 1890, p. 148. 
