274 Salisbury . — Methods of Palaeobotanical Reconstruction. 
former be employed great care must be exercised to ensure that there is no 
distortion, and that all the drawings are perfectly comparable ; the other 
alternative is to project an image of each section on to the plate-glass screen 
of a photo-micrographic apparatus over which has been stretched a sheet of 
tracing paper, the section is then drawn and a fresh piece of tracing paper 
substituted. By the latter method all distortion is avoided and the same 
magnification can be always obtained without difficulty. 
By means of the drawings each section is then cut out of a wax sheet, 
connexions being left where isolated portions occur. The wax sections thus 
formed are placed in order and joined together, either by pricking with a hot 
needle and applying pressure, or by running melted wax around the edges . 1 
It is obvious that the thickness of the wax sheets must have the same 
relation to the actual interval between the successive sections as the 
magnified representations bear to the real sections, and the accuracy of the 
model depends upon the assumption that the sections are equidistant and 
parallel. 
The method is particularly useful in the reconstruction of the stems or 
other structures with an elongated axis ; its chief defects, however, come 
from the fact that the successive sections of a series are frequently far from 
parallel, and the interval between them not a constant one. Where, as in 
stems, roots, and petioles, there is seldom any very rapid change in direction 
of either the structure as a whole or of its internal organization, and the 
series is, moreover, usually a long one, these objections are not of great 
importance. But in seeds and similar structures where rapid changes occur 
such considerations necessarily render it useless to the palaeobotanist. 
If permanent models are required which shall be unsusceptible to 
extremes of temperature, the sheets out of which the sections were cut can 
themselves be built up and employed as a mould from which a plaster 
of Paris model can be made. 
Professor and Miss Sollas 2 used this method with considerable 
success. They found that the plaster of Paris adhered readily to that 
which had freshly set, and by taking advantage of this fact they were able 
to add the sections one by one, so that all projections could be filled in 
with the plaster, and by means of a special apparatus each was planed down 
before adding the next in order. 
(b) The cardboard method. 
A method which has been adopted by the present writer, and which, 
though similar to the above, offers considerable advantages, is to paste 
drawings of the sections obtained by projection on to pieces of cardboard, 
which are then cut out and fixed in their appropriate positions by means 
1 See W. J. Sollas, Phil. Trans. Roy. Soc., Ser. B, vol. cxcvi, pp. 259-65. 
2 Phil. Trans. Roy. Soc., Ser. B, vol. ccii, pp. 231-2. 
