underwent substantial contraction. In nature, the shrinkage 
probably would not have been so extensive, as the environ- 
ment would have been more protective than that in the labora- 
tory. Changes in parameters suchas rate of water loss, temper- 
ature fluctuations, etc. would have been moderated. Also, 
further input of silica would probably have entered the void 
spaces created in contraction, maintaining the total occupation 
of the lumen. Perhaps this may be how the agate-like concen- 
tric bands of silica, so commonly found in the lumina of cells of 
naturally silicified woods, arise. 
Since the wood is serving as an active template for silica 
deposition during petrifaction, the condition of the wood prior 
to mineralization, i.e., the nature and extent of deformation 
and deterioration of wood structure at the commencement of 
impregnation, will be reflected in the quality of the retention of 
histological detail achieved in the petrifaction. The silicified 
Ginkgo woods from the John Day Basin of north central Ore- 
gon, with their near perfect preservation of structural detail, 
must have been essentially undegraded when initially silicified 
some 40 million years ago. 
As the silicification process proceeds to more advanced 
stages, the woody template, which initially sets the pattern and 
form of silicaemplacement, begins to deteriorite, creating open 
space in which subsequent silica influx can deposit. In this 
manner, the entire cell wall may be filled with silica. 
The time sequence of silica deposition with respect to the 
subdivisions of the cell wall probably follows the same se- 
quence as that observed by Barghoorn (1952) for cell degrada- 
tion. In other words, after penetration of the middle lamella 
region, deposition next occurs largely within the S2 layer of the 
secondary wall. The other structural units of the wall, which 
are volumetrically much smaller but more resistant to decay 
due to lower polysaccharide to lignin ratios, remain physically 
intact at the perimeters of both sides of the wall during silica 
infilling of the S2 layer. In this manner, they contribute to the 
maintenance of wall shape and form during build-up of silica 
within the cell walls. Hence, these more resistant structural 
units provide for the retention of the physical integrity of 
structure at the anatomical level of morphology. These factors 
collectively permit direct comparison of silicified woods with 
their living counterparts in botanical identification of the 
woods. Hence, silicification is not a single stage event, but a 
19 
