^lESOZDU' TLANTS FROM CHINA. 15 



them (represented on tlie rii;lit liand side of fip;. 10) meiisurcs 

 ')"2 iiini. long, o*J mm. luoad and lOnnn. thick. They a])peai' to 

 have heen sessile, with the place of attachment about 8 mm. broad. 



The surface is nearly smooth, showing only longitudinal 

 striations in its lower part. 



These fruits have been found attached to each other by 

 their lateral faces as if tiiey were twins, so that the place of 

 contact has been quite flattened. Their cross-section, as they now 

 are, is elliptical, but this may possibly be a distortion by pressure, 

 from a shape originally round. 



It is quite singular that there is another, though much 

 smaller, specimen (fig. 11) pressed almost flat, but also consisting 

 of two jiieces. If this specimen l)elongs to the same plant as 

 the first, it is not at all impossible that the fruits were borne 

 on the stem quite close to each other. 



4. GLOSSOPTERIS ? sp. 



PI. V. Fig. 2. 



A fragment of the basal part of a frond. It shows a midrib, 

 from which the lateral veins, though indistinct, are evidently 

 divergent. It seems to belong to one of the species of (r/a^.s'- 

 opteris, such as G. inclica Feistm. already described by Zeiller 

 (Flore Fossile des Gîtes de Charbon du Tonkin, pi. LVI, 

 fig. 1) from Tai-p'in-ch'ang (■i<i^V'^\ ^ place also in Yun- 

 nan. It is much to l)e regretted that the only specimen from 

 T'ang-t'ang is so imperfect as not to allow any accurate determi- 

 nation . 



