280 F. O. BOWER. 



traverse the cavity of the endosperm their course is more 

 direct, and here cases of branching may occasionally be 

 observed (this is described also by Sir J. Hooker, 1. c). The 

 tubes may be traced backwards towards the apex of the 

 endosperm and up to certain shrivelled bodies, which corre- 

 spond in position and form (as far as could be judged in 

 their disorganised condition) to the corpuscula of Ephedra. 

 Whether this be their real nature remains to be decided by 

 comparison with younger stages of development. We may 

 for the present assume that they are the corpuscula, and 

 apply that term to them, while we call the tubular cells 

 suspensors. 



Seeds of an unnamed species of Gnetum from Chittagong, 

 preserved in the museum at Kew, were also examined ; here 

 each one of the suspensors does not, as in G. Gnemon, pursue 

 a separate course, but they together form a coiled bundle 

 which lies in the cavity of the endosperm, and can easily be 

 removed with a needle (fig. 1). If the suspensors be teazed 

 apart, it is seen that each consists of a long tubular cell 

 having the characters above described ; the protoplasm 

 increases in density towards the end of the tube, at some 

 distance from which is embedded a single nucleus (fig. 2). 



This species would doubtless be better fitted for the study 

 of the development of the embryos than G. Gnemon, as 

 they could thus be easily removed from the endosperm, 

 whereas in G. Gnemon the only method possible for obtaining 

 preparations of the youngest embryos is to cut sections of 

 the endosperm, and search for the embryos, which are often 

 found to be fixed in positions unsuitable for observation. 

 Hence the study of the early stages of the embryo of G. 

 Gnemon is by no means easy, and this will explain the 

 incompleteness of my series of figures. 



Germination. 



' The two parcels of seeds sent from Java were sown at 

 Kew respectively in March and June, 1881. The first 

 seedling of the first parcel appeared above ground early in 

 November, eight months after sowing. The period of germi- 

 nation is, however, variable in diff'erent individuals. Owing 

 partly to this fact, and partly to the difficulty in obtaining 

 suitable preparations, I have not succeeded in observing the 

 first changes at the apex of the suspensor. The young 

 embryos are usually found near the axis of the endosperm, 

 but at a very variable distance from its apex. Though the 

 arrangement of cells in the youngest embryos which I have 

 observed points to an origin from a single cell cut off from 



