508 Arber and Parkin.— S Indies on the 
which but few descendants survive, were once dominant constituents in the 
floras of the past. In many cases we have discovered relics of their former 
greatness in the Palaeozoic or Mesozoic rocks. It may therefore be asked 
how is it that if, as we suppose, the Gnetales were a fairly abundant 
element in a former flora, we can produce no evidence of them in the fossil 
state ? The answer is to be found in the fact that our present knowledge 
of the real affinities of Cretaceous and Tertiary plants is extremely weak, 
and far less satisfactory than that of the Palaeozoic or even of the Mesozoic 
floras. Practically no petrifactions are known, with the exception of much 
silicified wood of Angiospermous or Gymnospermous affinities, which does 
not contribute to the question either one way or the other. Further, even 
impressions of fructifications have hitherto proved to be extremely rare, 
and very unsatisfactory from this point of view. The few flowers preserved 
in the Oligocene amber stand quite alone as affording reliable evidence in 
this direction. By far the larger number of specimens, which have been 
described from the Cretaceous and Tertiary rocks, consist solely of im- 
pressions of leaves, in the great majority of cases quite detached. The 
only criteria which they present, as indicative of their affinities, are the shape 
of the leaf and the details of the nervation, both, in our opinion, extremely 
unsatisfactory indices of systematic position, where they have to be relied 
upon alone. 
Let us imagine for a moment that it was known, from the evidence 
of fructifications preserved in amber, that the Gnetales really did exist 
in the early Tertiary period. What would then be our attitude towards 
the Cretaceous and Tertiary leaf impressions ? One of the three living 
genera is practically leafless. Another, W elwitschia , is so extremely 
specialized, that its leaves give us no certain clue as to what was the type 
of the foliage characteristic of the group as a whole. There remains 
Gnetum , with its opposite leaves, which in form and nervation recall exactly 
those of certain Dicotyledons. Thus, so far as the living members of the 
group are concerned, the only hint they give us, as to what was possibly a 
typical Gnetalean leaf, is one which coincides exactly with that of a 
Dicotyledon. Were the fact known from fructifications preserved in amber, 
that the Gnetales did exist throughout the Tertiary period, we should 
then be led inevitably to the conclusion that some of the reticulately- 
nerved leaves of the Tertiary rocks were, in all probability, not the leaves 
of Dicotyledons, but of the Gnetales. 
That such is the case we believe to be highly probable, although 
we have no undoubted evidence of the group in the fossil state. The 
flowers preserved in amber, and those known as impressions, which are 
sufficiently well preserved to be more or less recognizable, are, after all, 
extremely few in comparison with the number which have perished com- 
pletely, and the negative evidence in this case is even more untrustworthy 
