498 Transactions.—G cology. 
with beds of agglomerates, ashes, tufas, and laterites can be traced to the 
very summit. Still clearer sections are open to our inspection if we follow 
the barranco or entrance into the harbour, forming sometimes vertical 
cliffs of considerable altitude, and where the whole series of beds can easily 
be followed. However, the most interesting and complete insight was 
obtained in the railway tunnel passing through the caldera wall, and of 
which, as the work gradually advanced, I prepared a careful section. The 
succession and dip of the lava streams and the intervening beds can also be 
made out by following the slopes of the ridges between the deep valleys 
washed out on the outer side of the crater wall, where it will be found that 
the lava streams forming the lip of the crater have generally a slighter 
inclination than those lower down, the dip of the upper ones being only nine 
degrees on the average. In the tunnel the dip is greater, an inclination of 
twenty degrees not being uncommon. It is evident that the building up of 
such a huge system during numerous eruptions, often of great magnitude, 
could not be accomplished without a great destruction of portions of the 
beds previously formed taking place, the point of eruption in the crater 
shifting continuously about the centre. If, at the same time, we examine 
the lava streams and the interstratified agglomerate and ash beds along the 
water’s edge, we have to come to the conclusion that all the eruptions by 
which the caldera wall was formed from summit to bottom, occurred under 
the same physical conditions. 
Examining into the formation of the Lyttelton caldera, and beginning 
our observations in the harbour, we find that many lava streams have been 
preserved which have cooled in their ascent; others lie horizontal for a 
short distance, and are then seen to descend, conforming to the gradient of 
the underlying lava streams or agglomerate beds. In many instances we 
have also clear evidence that considerable destruction of the beds previously 
formed had taken place before new streams flowed over the lip of the crater, 
or before beds of ashes, scorie and lapilli, were deposited anew. The 
tunnel section in this respect is also very instructive. Thus, in course of 
time, the great crater wall was formed, rising to an altitude of nearly 2,000 
feet, and having a diameter of more than five miles at its crest. It is clear 
that close to the vent, from which scorie and ashes were thrown out in 
large quantities, the greatest thickness of the agglomerate beds ought to be 
formed, and this, in fact, is the case, as the largest beds, having sometimes 
a thickness of several hundred feet, are situated within the inner side of the 
caldera wall. The lava-streams here between these agglomerates are irre- 
gular in their direction, and mostly of small dimensions. The more we 
advance towards the outer slopes of the caldera wall, the less frequent 
become these agglomeratic or tufaceous layers, whilst the lava-streams, 
