GNBTALES 123 



spike functions, Lotsy ^^ observed that it possesses tlie three en- 

 velops of the complete ovulate flowers. Karsten ^°' ^^' ^^ found 

 that the middle envelop appears during the development of the 

 incomplete ovulate flowers, but gradually disappears. 



Strasburger ^ regards all three envelops as integuments, the 

 middle and innermost corresponding to the inner envelop in 

 Ephedra. Beccari * flrst described the organogeny of the 

 flower, discovering that the three envelops appear in acropetal 

 succession. This seemed to militate against the theory of ovular 

 integuments, which appear in basipetal succession. Beccari's 

 interpretation, however, seems somewhat fanciful, since he re- 

 gards the middle envelop as the homologue of the staminate 

 cycle in Tumhoa, the outermost envelop being a perianth. 

 Lotsy regards the two outer envelops as two whorls of bracts, 

 which he calls the perianth, and claims that there is only a single 

 integument. He lays stress upon the fact that in the two outer 

 envelops there is vascular tissue, claiming that this indicates 

 their foliar nature. 



If Celakovsky's general claim be accepted, that there are 

 two integuments in all Gymnosperms, the so-called perianth of 

 Gnetales becomes an outer integument, and the middle envelop 

 of the ovulate flowers of Gnetum may well be only an outer layer 

 of the inner integument, as Strasburger has suggested. In 

 any event, there seems to be no question as to the tegumental 

 nature of the innermost envelop, and the acropetal succession 

 of the envelops would indicate that the one or two outer ones 

 are modified bracts. It must be remembered that this does not 

 necessarily militate against Celakovsky's view, since he regards 

 integuments as of foliar origin. 



The development of the megasporangium has been described 

 in detail by Strasburger ^ for Gnetum Gnemon, and coincides 

 with the general sequence in Gymnosperms. The archesporium 

 consists of a group of hypodermal cells which cut off wall cells 

 by periclinal division. By repeated periclinal divisions the 

 wall cells form the usual thick mass of sterile tissue above the 

 sporogenous region (Fig. 91, F). Before pollination the cells at 

 ■ the tip of the nucellus disorganize, an irregular pollen chamber 

 being the result, much as in Pinus. The several primary spo- 

 rogenous cells divide more or less and organize the deep-seated 

 group of mother cells. 



