Salisbury . — Methods of Pcdaeobotanical Reconstruction. 275 
of wires. The intervals between the sections are thus represented by spaces, 
so that the alterations both in these intervals and in the planes of section 
can be allowed for. Further, the magnification of the model is in no way 
limited either by the size or thickness of the wax sheets or other material 
used. 
(c) The glass method. 
The most useful, and at the same time the most ingenious, of all the 
methods of this class is that invented by Professor Graham Kerr, 1 which, 
though primarily intended for the study of microtome sections, is, with 
certain modifications, of extreme value to the palaeobotanist. Here ground- 
glass sheets are employed ; the successive sections are drawn upon them 
in pencil and the various structures differentiated by means of water-colour 
paints. A few drops of clove oil are placed upon each sheet and the next 
then added, so that the whole block thus formed appears transparent, whilst 
the structure itself stands out as if reconstructed. Graham Kerr’s method 
is similar to that formerly employed by Vosmaer, but differs in the use of 
ground glass and an interposed fluid. 
For the purpose of the palaeobotanist the sheets of glass should be 
attached by clips to three upright supports arranged in a triangular manner 
and permitting of movement in the vertical direction. In this way the 
variation in the interspaces can be allowed for and differences of angle 
obtained. 
II. Reconstruction of Non-serial Sections. 
Where serial sections are not available, or the structures under investi- 
gation are so short as to preclude series of more than 3 or 4 sections, the 
methods described above are of very little value. In all such cases it is 
necessary to assume, unless of course there be good evidence to the 
contrary, that all the sections at one's disposal have been cut from structures 
of approximately equal relative dimensions ; and in the case of seeds 
where this method has been utilized such an assumption does not appear to 
have been unwarranted. 2 
At the outset a primary examination of all the sections is made 
in order to obtain all the dimensional data possible. 
In any oblique section there is one plane in which the dimensions of 
the structure are not exaggerated by the direction in which it has been cut. 
By noticing these true values in all the sections at one’s disposal, the real 
dimensions of several structures, such as the thickness of a testa or the 
width of a seed, are obtained. 
The angle of obliquity of some of the remaining sections can then be 
1 Q. J. M. S., No. 177, p. t, 1902. 
2 Oliver and Salisbury, Ann. Bot., vol. xxv, 1911, p. 4, 
