GERMIVATIOX AND EMBRYOGENY OF GN'ETl'.M GNEMON. 279 



and has t\vo small cotyledons, &c. The latter author, how- 

 ever, gives drawings of seeds with a cavity in the endosperm 

 but no embryo. Finally, in the published account of the 

 botanical results of the voyage'of the Coquille (1822 — 1825), 

 seeds of GneUim Gnemon are figured with an embryo. 



Following a reference in Griffith's paper, above cited, I 

 find that the embryo of Gnetwn scandens had already been 

 drawn (but not published) by Roxburgh, and on examining 

 the original drawing in the Kew Herbarium, I find the 

 embryo represented as having two unequal cotyledons of 

 large size, a short radicle, and a plumule as long as the 

 cotyledons, bearing at its apex two very small leaves ; the 

 suspensor is omitted. It will be subsequently seen that 

 this does not tally with my observations on Gnetum Gnemoti. 



In other species [G. Gnemoii) there is usually found a 

 cavity near the apex of the endosperm of the mature seed, 

 but no suspensors are apparent till sections are examined 

 under the microscope, when it is found that tubular cells 

 permeate the apical part of the endosperm. This fact has 

 already been recorded by Sir J. Hooker (1. c), who also 

 noted that these tubes occasionally branch (cf. infra). 

 In specimens of Gnetum Gnemon from Java, which appa- 

 rently included all stages of development of the flowers, 

 Strasburger (' Angiosp. und Gymnosp.,' p. 101) found no 

 such bodies. 



Ripe Seed. 



Early in 1881 I received from .Java, through the kind- 

 ness of Dr. Treub, two parcels of ripe seeds of Gnetum 

 Gnemon. On examining longitudinal sections of the endo- 

 sperm of these seeds, it is seen that there is no embryo 

 already developed ; the main body of the section consists of 

 ordinary cells of the endosperm, loosely aggregated, with inter- 

 cellular spaces between them; near its apex tliere is usually 

 a cavity (but in specimens of G. Gnemon from the Kew 

 Museum this is not always the case). The cavity is evi- 

 dently due to rupture of the tissue, and its occurrence may 

 depend upon the manner of ripening of the seed. Among 

 the cells of the endosperm may also be found numerous long 

 tubular cells, with walls which stain blue with solution of 

 I in KI; they have rather transparent protoplasm, and a 

 nucleus (I have never observed more than one nucleus) ; 

 transverse septa occur in rare cases, but there does not 

 appear to be any definite terminal cell cut off before germi- 

 nation. The course of the tubes through the endosperm is 

 sinuous, and for the most part longitudinal ; where ihey 



