348 Takeda. — - Some Points in the Anatomy of 
Goniopteridis’, which becomes more or less irregular owing to the aberrant 
path of the lateral transverse branches or veins, as already described by 
de Bary ( 4 , p. 303) and Sykes ( 19 , p. 181). One point to note here is that 
the short branch given off from the transverse branches is always directed 
towards the apex, and never towards the base of the leaf! 
The cotyledon is elliptical-oblanceolate and connate at the base, and 
shows a similar kind of nervation in a looser and more or less irregular 
manner. Two bundles enter each cotyledon, and traverse the latter 
straight up to the apex, taking a parallel course. At the very base of the 
cotyledon, which is connate and forms a very short tube, each of these two 
bundles gives off a lateral bundle ; this again divides, a little further up, into 
two. Thus there are six main bundles present in the cotyledon : two central, 
two lateral, and two marginal ones. The ‘ marginal ’ bundles are much 
weaker than the rest and die away about midway of the whole length of 
the cotyledon, while the ‘ lateral’ bundles also do not really reach the apex. 
The veins are much more irregular and more oblique than in the adult leaf. 
These veins anastomose or end blindly, just as in the foliage leaf (PI. XXIX, 
Fig. 1). 
The nervation with three pairs of bundles derived from the double 
leaf-trace seems to me to be the phylogenetically primary one, and this 
design is also found in the young leaf. 
It has been stated both by Bower ( 2 , p. 19) and by Hill and de Fraine 
( 7 , p. 323) that four bundles enter the cotyledon. They seem to have not 
noticed the real base of the cotyledon and taken the connate portion of the 
cotyledon for a structure belonging to the hypocotyl. 
The nervation of the young leaf of my material shows a very interesting 
feature. In the portion about 3 cm. in distance from the apex the nerva- 
tion is practically the same as in the cotyledon, i. e. it shows six main 
bundles and very oblique veins (cf. Fig. 1). However, a difference is to be 
noticed, inasmuch as the ‘ marginal 5 and ‘ lateral ’ bundles do not directly 
fuse with the ‘ central ’ ones, but are only connected, transversely, by the 
veins. 1 Further downwards there are to be seen new bundles of different 
lengths running parallel to the long axis of the leaf. These lie between 
the three pairs of bundles above noted, as already described by Bower 
( 3 , p. 586). Thus a direct connexion between the central pair of bundles 
and the ‘ lateral ’ +‘ marginal ’ ones on either side, as a result of branching, 
such as we see in the case of the cotyledon, does not exist. This is brought 
about by the new bundles of secondary origin. The central pair of bundles, 
therefore, only form the direct continuation of the hypocotyledonary 
bundles ; all the others are connected together by the plexus of vascular 
bundles in the * crown ’, which are differentiated later (cf. 3 , p. 585 et seq.). 
1 For the section of this stage cf. Bower (3), PI. XXXII, Fig. 7 . My ‘ marginal ’ and ‘ lateral ’ 
bundles correspond to his * secondary vascular bundles 
