572 J. CLIFTON WARD ON THE GRANITIC, GRANITOID, AND 



amples of such sources of error, compare figs. 1-5, in PI. XXX., "with 

 figs. 6-10 ; in the former cases the liquid-cavities, being comparatively- 

 large, have lost fluid, whilst in the latter thej have probably remained 

 closely sealed during those long ages which have elapsed since the 

 first formation of the rock containing them. There is this striking 

 difference between the two sets of cavities :. whereas figs. 1-5 contain 

 bubbles showing no free movement, the still vacuous bubbles in 

 figs. 6-10 are in a state of constant activity, moving about in the 

 liquid to different parts of the cavities. 



At first, making use of a neutral-tint glass reflector, I carefully 

 traced on paper all the best-defined cavities, making a note by the 

 side of those in which the vacuities showed a constant spontaneous 

 movement. I soon found, however, on comparison of a considerable 

 number of drawings, that the fixed bubbles were almost invariably 

 relatively larger than those which showed this free movement ; and 

 measurements with the micrometer proved the same. This led me 

 wholly to reject the cavities containing fixed bubbles for purposes of 

 measurement ; and in every case used in this investigation the micro- 

 meter-measurement has been made solely of cavities containing freely- 

 moving vacuities. In PL XXX. figures containing fixed bubbles 

 have been introduced for comparison with the others ; and all these 

 are marked with a small /. 



It will at once occur to many that, unless liquid- cavities can be 

 ascertained to be of a very uniform shape, it is impossible to gain a 

 true idea of their actual size by a measurement in one plane alone. 

 That their form is often very irregular is not to be doubted ; and, as 

 shown in figs. 1, 4, 58, 59, and 63, they often have more or less 

 tubular processes projecting from them in various directions, which 

 may or may not happen to be in the focus-plane ; if the latter, the 

 measurement of the cavity in one focus does not represent the true 

 size. A fairly reliable measurement may, however, be made in 

 those cases in which the line bounding the cavity is sharply defined 

 all round in one focus, and in which the vacuity moves freely to all 

 parts, keeping all the time in focus : this is so with most of those figured 

 in PI. XXX. Even in such cases it may be supposed that minute 

 tubular processes sometimes extend outwards from the focus-plane, 

 only that the bubble, minute as it may be, cannot find entrance 

 into them ; an example of such fine tubular processes in the focus- 

 plane is seen in fig. 58, the bubble moving freely everywhere except 

 into these tubes. While, however, we must be on our guard against 

 this possible source of error, there is little reason why the average 

 great number of good observations should not be relied upon. 



It sometimes happens that in cavities of a form similar to 

 figs. 7 and 9 the vacuity moves freely about at one end, but never 

 enters upon the field of the other half. This may be due to the 

 cavity being more shallow at one end than the other • therefore in 

 this case the measurement taken in one plane will be slightly in- 

 correct. Indeed when one reflects upon the many sources of possible 

 error in measurement, and the difficulty of observing a sufficient 



