276 Salisbury - — Methods of Palaeobotanical Reconstruction . 
deduced from these known values, and since the calculation of these angles 
involves considerable time it is usually much simpler to obtain them 
graphically. For example, the thickness of a testa is represented on the 
magnified scale by a horizontal line, and from one end of this a second 
line is drawn vertically upwards. Using the other end of the first line as 
a centre, an arc is then described having as a radius the magnified dimension 
of the testa in the oblique section. On joining up the centre to the point of 
intersection with the vertical, the angle of obliquity is obtained and can be 
read off by means of a protractor. 
In this way from known values the plane of several of the sections can 
be determined, and by reversing the above graphical method the dimensions 
of other structures to those already known can be deduced from their 
oblique values, the latter being plotted at the known angles of section and 
projected on to the horizontal. 
It will be seen that in this way quite a considerable number of 
dimensional values can be arrived at, and on the basis of these a preliminary 
ideal section of the structure as a whole can be drawn. 
Each section is then measured along the line corresponding to the 
plane of intersection with the ideal section already constructed. The 
relative positions of the various parts and the boundaries of tissues along 
this imaginary line are plotted off along the edge of a strip of paper, so 
that for each section we have in convenient form the relative positions of 
its parts as they will appear when appropriately plotted. 
If the original ideal section be approximately correct we should be able 
to fit in these lineal representations, so that the limits coincide at the angles 
which have already been determined. 
In actual practice we shall find that our ideal section will undergo 
considerable modifications, after which we can proceed to reproduce it in 
the form of a model, using for the purpose some plastic and easily cut 
material such as plasticine. 
It is obvious that for all the foregoing methods, and especially the last 
described, where the absence of serial sections greatly enhances the possi- 
bility of error, some means of checking the results must be adopted, and 
such a proof of one’s results can be obtained by cutting the model at the 
angles and in the planes indicated by the plottings. If the model be 
a correct reconstruction the results will be magnified representations of the 
corresponding sections. Failure to obtain this will necessitate a revision 
of the model until any section can be reproduced upon it. 
Adequately to cut such models in a perfectly flat plane necessitates the 
use of special apparatus, and a mechanism designed for this purpose will 
now be described by means of which the planes of obliquity of other 
sections, which do not admit of interpretation by the ordinary methods, can 
usually be elucidated (see Figure). 
