G4 BIXNIE: MESOZOIC rocks of the north-east coast of IRELAND. 
prisms being continuous, so that when viewed from a distance, the 
pillar appears to consist of but one portion. It has been noticed 
that when there is a joint, tlie cup or the hemisphere does not extend 
right across the section at right angles to the prism, but that there is 
a small annulus surrounding the cup or hemisphere, which extends as 
a plane to the outside of the prism, Professor James Thompson, of 
Belfast has explained this in the following way : — When the basalt 
cools, it expands at the moment of sohdification. The outside 
expands first and throws the interior into a state of tension ; the part 
thus thrown into tension gives way and forms the ball and cup. The 
outside of the pillar begins to weather until the basalt is of such 
a consistancy that on any slight movement the narrow rim of material 
still connecting the two joints snaps, and thus the annulus is formed. 
There seems to be no definite rule determining the occurrence of the 
cup and ball of the joints. 
One portion included between two joints may have a cup at 
each end, a ball at each end, or a cup at one end and a ball at the 
other. In the pillars there seems to be no definite law connecting 
the relative position of these. We can have the cup of one portion 
uppermost, and a ball of the next one uppermost and so on. This 
follows from what has been said about the joints. Neither does there 
seem to be any law producing similar jointing in aggregates of 
adjoining pillars. The only generalisation which seems possible to 
be drawn is that, on examination of the separate flows, we find that 
the columnar jointing stops short some feet below the top of the 
floor, not taking into account the amount of basalt which may have 
been removed before the next outburst. The columnar formation has 
been supposed to be due to the fact that, in cooling, the basalt splits 
up into globules, each of which is surrounded and pressed upon by a 
certain number of others, thus giving it the shape of a many-sided 
prism. This theory does not account for the singular coincidence of 
so many joints one on top of the other. 
(2) We find the basalt occurring in a form, which, on weathering, 
presents the appearance of an aggregation of spheroids cemented 
together by an ochreous-looking mud. These spheroids, when split 
open, present a concretionary appearance, as if they were formed of 
