LYCOPODIALES— LEPIDODENDREyB— LEPIDOPHLOIOS. 209 



point of attachment. None of the isolated bolsters are so preserved as to 

 show a dorsal snrface of any considerable size. Such as are preserved so 

 as to sliow the carbonized cortex concave seem to re^^resent merely an inner 

 surface conforming- to the wall shown convex in the otlier instances, and 

 without trace of the leaf scar, though the ventral cicatrix is distinct. A 

 conspicuous character of all the bolsters from Avhich the cortical residue 

 has been removed is the large and deep pit, often vertically elongated, 

 corresponding to the ventral trace. The ventral trace is prominent also on 

 the rmder surface, as may be seen in Fig. 5, PL LVI, or Figs. 3 and 6. A 

 not uncommon wrinkling in a generall}^ longitudinal direction of the 

 interior (?) casts of the leaf cushions, which is probably due to the flatten- 

 ing of the bolster, is best seen in PI. LVI, Fig. 4, in which the interior casts 

 of the lateral cicatricules or transpiratory tracts are seen to be very large 

 and prominent. In Fig. 7 the wrinkling is A^ery slight. 



Several other detached bolsters deserve illustration. One of these, 

 PL LVI, Fig. 8, from Pitcher's coal mine, represents the impression or 

 mold of a large detached bolster, 47 mm. broad and over 25 mm. in 

 altitude. In this specimen the outline of the foliar side and of the widely 

 rounded lateral angles is well shown, as is also the inflation in the side 

 opposite the leaf scar. This inflation, which corresponds to that seen in the 

 specimens from Gilkerson's Ford, has here caused the wall to turn up nearly 

 vertical, producing wrinkling, the precise extent of which can onlv be esti- 

 mated from the comparison of the other bolsters. The margins on either 

 side of the leaf scar are regular and natural, though slighth' abraded along 

 the central portion of the foliar scar. The position of the ventral cicatrix, 

 relatively close to the leaf scar, is ver}^ clear. 



In arranging the figures on the plates and in my taxonomic references 

 I have conformed to the generally accepted conclusion that the flatter or 

 more emarginate border <:)f the leaf scar is its upi^er margin, the small 

 deltoid or subtriangular pit or trace being thiis ventral. Nevertheless, the 

 apparent outlines of the detached bolsters, tlie marginal position of the leaf 

 scars, together with the general form of the side opposite the scars, seem 

 entirely incompatible with a pront>unced l^ulbil or stalklike form or habit 

 in these bolsters, as in L. erassicaulis Gold., or L. scoticus Kidst. It is 

 evident that either the leaf scar was at the summit of a relatively thin, 

 shell-like or scale-like cushion or leaf base, attached along a portion of the 



MON XXXVII 14 



